research article efficient priority access to the shared...

Research ArticleEfficient Priority Access to the Shared Commercial Radio withOffloading for Public Safety in LTE Heterogeneous Networks

Chafika Tata and Michel Kadoch

Department of Electrical Engineering Ecole de Technologie Superieure 1100 Rue Notre-Dame OuestMontreal QC Canada H3C 1K3

Correspondence should be addressed to Chafika Tata chafikatata1ensetsmtlca

Received 4 June 2014 Accepted 30 August 2014 Published 12 October 2014

Academic Editor Lixin Gao

Copyright copy 2014 C Tata and M Kadoch This is an open access article distributed under the Creative Commons AttributionLicense which permits unrestricted use distribution and reproduction in any medium provided the original work is properlycited

This paper presents the algorithm Courteous Priority Access to the shared commercial radio with offloading (CPAwO) for publicsafety network (PSN) over LTE heterogeneous networks (HetNets) We propose a solution for prioritization of PSN users withaccess to the commercial radio network resources Our model offers additional radio resources to the PSN Furthermore it ensuresa certain priority for commercial users by assigning quantities of additional radio resources through the courteous scheme Thisallows delaying preemption and blocking bearers when the radio resources are limitedThe other part of CPAwOmodel is to applythe principle of offloading in order to reduce the impact of the macrocell congestion This technique is to switch the new bearersarriving at the LTEmacrocells to small cells in order to decrease the number of the blocked and interrupted bearers in the networkThe results of the simulation showed that the allocation of radio resources via the courteous mechanism with offloading of newbearers to small cells reduces the rate of blocking bearers and delays the preemption of active bearers present in the LTE HetNetsIt also reduces the cost of the end-to-end communications thanks to the reallocation of free frequencies

1 Introduction

Wireless operators are facing a growing demand for com-munication broadband services On the one hand the sat-isfaction of the users is very important for the commu-nication companies while on the other hand the lack ofbandwidth and radio resources prevents wireless operatorsto meet all clientsrsquo requests Long term evolution (LTE)over heterogeneous networks (HetNets) [1 2] is a pertinentsolution to consider The LTE resources have been used toimprove the performance of public safety network (PSN) [3]Indeed the frequency band of 700MHz was dedicated to itin North America [4] However the increasing demand interms of radio resources reduces the available PSN resourcesto ensure a good QoS especially during times of crisesTherefore sharing the commercial frequency band with thePSN network became an important solution addressed byseveral authors [4ndash7]

In this study a new approach for Courteous PriorityAccess to the shared commercial radio with offloading

(CPAwO) for the PSN over LTE HetNets networks is pre-sented The main objective of this solution is to ensureprioritization of PS users to access the shared commercialradio network resources Our system as well as offeringsupplementary radio resources for the PS network ensuresa certain priority for commercial users by assigning addi-tional resources quantities through the courteous processThis is possible only if the QoS of the PS traffic is at anacceptable level This approach enables delaying preemptionand blocking of bearers during congestion However cour-teous algorithm may not be applied when the resources areexhausted in the LTE macrocell For this purpose anotheraspect of our solution is modeled to use unlicensed networkssuch as WiFi WMN or ad hoc and to offload the LTE users(EU) of the congested macrocell This has to be entirelytransparent to the end Hence the EU uses the unlicensednetwork to exchange data This is Device-to-Device (D2D)communications [8 9] Our approach delays the preemptionof bearers in the LTE HetNets network and reduces packetloss In fact new bearers will be redirected to the WMN

Hindawi Publishing CorporationJournal of Computer Networks and CommunicationsVolume 2014 Article ID 597425 15 pageshttpdxdoiorg1011552014597425

2 Journal of Computer Networks and Communications

instead of preempting the active bearers to release the neededradio resources Offloading scheme is applied as long asthe unlicensed network is not congested Otherwise thepreemption must be applied on the lower priority activebearers

Different radio resource allocation constraintsmodels arebuilt formanaging the frequency resources quantity allocatedto each traffic type in order to guarantee a good qualityof service for all bearers in the network Courteous band-width allocation constraints model for frequencies (CAMF)is used to design a new frequency allocation model withconstraints in terms of radio resources based on the courteousbandwidth allocation constraints (CAM) model [10] CAMis a bandwidth allocation constraints model built especiallybut not exclusively for the MPLS network It focuses onimproving the allocation of resources for the classes of lowerpriority traffics The CAM enhances the traffic allocationin the network compared to the two approaches maxi-mum bandwidth constraints allocation model (MAM) [11]and Russian dolls bandwidth constraints allocation model(RDM) [12] In this work we adapt the CAM model forthe frequencies allocation scheme Finally three models havebeen developed namely the basic model CAMF (CAM forfrequencies) the generalized CAMF model G-CAMF andRUS-CAMF model (radio usage situation based courteousallocation constraints model for frequencies) These modelsare detailed below

Note that RUS-CAMF and G-CAMF may be used byCPAwO solution besides CAMF is used by our previoussolution CPA algorithm [5] CPA (Courteous Priority Accessalgorithm) is one of our contributions published in [5] CPAaims at enhancing the radio resources allocation for theCommercial and public safety networks The main objectiveof CPA is to allow the PS network to access the sharedcommercial radio over LTE with priority without penalizingcommercial users The PS network may give an amountof its resources if the CN one suffers from a high rate ofpacket loss When comparing with the solution given by [4]the simulation results show that CPA algorithm reduces thenumber of the commercial blocked bearers delays the CNblocking mechanism increases the number of commercialactive bearers and keeps an acceptable level of PS blockingbearers Note that when resources are limited for PS networkit may not yield any part of its dedicated resources At thistime no more courteous allocation is done and a preemptionwill be invoked

This paper is organized as follows Section 2 presents therelated work In Section 3 the CPAwO algorithm is modeledSection 4 illustrates the different radio resource allocationconstraints models Section 5 summarizes the most impor-tant simulation results and their interpretation Section 6concludes this paper Finally Section 7 presents the futureworks relevant to this study

2 Related Work

Several studies in the literature deal with the radio resourcesallocation issue in LTE network [13ndash15] The objective is to

provide efficient schemes to manage the radio resources inthe LTE and LTE HetNets networks Article [13] offers a newsolution to the radio resources allocation through a novelradio admission control (RAC)mechanism RACproposes todivide bearers into three groups namely the first groupwhichincludes the bearers vulnerable to preemption the secondgroup including the nonpreempting bearers which can inter-rupt bearers from the first group and the third one whichcontains the nonpreempting bearers that cannot interruptothers bearers Furthermore bearers in the same group mayshare the same amount of radio resourcesTherefore once thesecond group resources are limited its bearers can interruptothers bearers from the first group Authors in [14] proposea control admission mechanism combined with a congestioncontrol scheme as a solution to reduce the blocking and theloss calls probability in the system In [15] high prioritybearers interrupt lower priority active bearers until gettingthe necessary resources to reach the minimum required QoSlevel This may decrease the number of bearerrsquos preemptionin the network Some other studies address the problem ofpublic safety userrsquos access to the shared commercial radioover the LTE and the LTE HetNets The two papers [4 5]develop solutions to provide a priority access to the sharedcommercial radio for the PS clients over the LTE networkIn the first paper PS users may use only a part of the sharedradio even if no commercial user requests the radio resourcesBesides commercial users can use all the radio bands butcan be interrupted if PS bands are requested by PS clientsTherefore commercial bearers may suffer from high levelof preemption The second paper offers additional resourcesto the commercial users from the PS radio while the QoSlevel of the PS traffic remains acceptable which delays thepreemption of commercial bearers and reduces call loss

Managing radio resources is efficient to enhance the radioresources allocation in the LTE network when resourcesare available But once all resources are allocated and nopreemptable bearer is active in the system the level ofQoS of different traffics will decrease To address this issuethe bearers offloading mechanism from the macrocell tothe small cells can be applied The study [16] discussesthe offloading of bearers from an LTE macrocell to WiFinetwork This approach considers the case where the smallcells position discovery is performed by the cellular networkin order to optimise the number of D2D connections Theauthors demonstrate that their model can provide significantimprovement in capacity and energy consumption Othersolutions are presented in to deal with the enhancement ofD2D communications [17ndash19]

3 Courteous Priority Access to the SharedCommercial Radio with MacrocellOffloading for PSN over LTE HetNets

In this section we detail the PS network Courteous PriorityAccess with offloading approach to the commercial radioover the LTE HetNets namely CPAwO We offer a solutionfor Commercial Radio Access with prioritization for thePSN Our scheme gives additional radio resources to the PS

Journal of Computer Networks and Communications 3

network Moreover it keeps a certain priority for commercialusers by assigning an amount of residual PS radio resourcesthrough the process of courtesy This is possible only if theQoS of the PS traffic is acceptable This approach delayspreemption and blocking bearerrsquos process when the radioresources are limited for the commercial network The otherstep of the CPAwO approach is offloading the macrocells inorder to reduce the congestion within them This techniqueconsists of switching the newbearers arriving at the congestedLTE macrocells to small cells such as WiFi WMN and adhoc cells a totally transparent operation to users which is ableto reduce the number of blocked bearers and interrupted onesin the whole network

Following is the detailed CPAwO scheme including thebearers classification which is a priority calculation model

31 Bearers Classification and Priority Calculation As wementioned before all bearersrsquo types may share the availableradio resources in the LTE HetNets However all bearersdo not have the same priorities in terms of access to radioresources and not all bearers can access these resources whenthey are limited Therefore it will be relevant to classify thebearers according to their priority level In this section wediscuss the classification and priority calculation of differentbearers in the LTE HetNets Indeed in this study we classifybearers of commercial network together with bearers ofPS network Moreover the bearers belonging to the samenetwork PS or CN are classified too

The classification of bearers according to their priority isvery important for theARPmechanismduring the admissionof new bearers or the modification of active bearers in theLTEHetNets and also during the preemption of some bearersto release some radio resources in favour of higher prioritybearers

The priority calculation is performed when a new bearerestablishment is requested Three elements affect the cal-culation of the priority of a bearer PS or CN namely theusage status represented by 120572 which specify if the bearerestablishment is relevant to an emergency or nonemergencysituation the ARP value related to this bearer noted 120599(Br)and the QCI value of the bearer noted QCI(Br) Authors in[4] give more details about these three elements

Thus we propose the following formula for the calcula-tion of the priority coefficient of a Br bearer

Φ (Br) = 120572 lowast 120599 (Br) lowast (120573 +QCI (Br)) (1)

where Φ(Br) is the bearer priority coefficient 120572 is the usagestatus = 01 if emergency 3 otherwise 120599(Br) is ARP valuerelevant to the bearer andQCI(Br) is QoS class identifier 120573 =max(120599(Br)) = 8 if bearer type = CN 0 Otherwise

Note that as the value of Φ(Br) is small the priority ofthe bearer increases Furthermore in our work we considerthat all emergency PS packets have higher priority thanemergency CN ones to access the PS dedicated spectrumband belonging to the shared commercial radio By the sameway the whole nonemergent PS traffic is prioritized overall nonemergent CN ones on access to the shared spectrumband This is made possible by the value of 120573 120573 ensures that

the lower priority bearer of PSwill have lower value forΦ thanthe high priority CN bearer Moreover emergency calls PS orCN must have higher priorities than the nonemergency oneso lowest coefficientsΦ To address this issue the usage statusnamely 120572 is considered for the computation of the prioritycoefficient of the bearer

Remember thatQCI values vary between 1 and 9 inclusiveand ARP varies from 1 to 15 inclusive [4] The priority calcu-lationrsquos numerical modeling is given below The objective isto show that our approach allows prioritizing higher trafficpriority over lower priority one and PS traffic is compared tothe CN traffic

Formulas (2)ndash(10) show the coefficients of priority trafficon two different usage situations including emergencies anda classic situation The Min(119894) and Max(119894) notations meanrespectively minimal value and maximal one for a variable119894 with 119894 isin R Also we take BrPS as a PS bearer and BrCN asan CN bearer

Min (Φstatus (Br)) le Φstatus (Br) le Max (Φstatus (Br)) (2)

If status = emergency then 120572 = 001 and

Min (Φemergency (Br))

= Min120572 lowastMin (120599 (BrPS)) lowastMin (QCI (Brps))

= 001 lowast 1 lowast 1 = 001

(3)

Max (Φemergency (Br))

= Max (Φemergency (BrCN))

= 120572Max (120599 (BrCN)) lowast (120573 +Max (QCI (BrCN)))

= 001 lowast 15 lowast (8 + 9) = 255

(4)

001 le Φemergency (Br) le 255 (5)

Also when we apply formula (1) and consider the ARP andQCI values of bearers we get

001 le Φemergency (BrPS) le 072 (6)

081 le Φemergency (BrCN) le 255 (7)

If status = nonemergency then 120572 = 3 and

Min (Φnonemergency (Br))

= Min (Φnonemergency (BrPS))

= 120572 lowastMin (120599 (BrPS)) lowastMin (QCI (BrPS))


Max (Φnonemergency (Br))

= Max (Φnonemergency (BrCN))

= 120572 lowastMax (120599 (BrCN)) lowast (120573 +Max (QCI (BrCN)))

= 3 lowast 15 lowast (8 + 9) = 765

3 le Φnon-emergency (Br119894) le 765 119894 = PS or CN

(8)


Also

3 le Φnonemergency (BrPS) le 216 (9)

243 le Φnonemergency (BrCN) le 765 (10)

The numerical model proposed above for the calculation ofbearersrsquo coefficients of priority ensures that during transmis-sions belonging to the same communication status namelyan emergency situation or nonemergency one the PS traffichas priority over the CN traffic see formulas (6) and (7)and formulas (9) and (10) Moreover the model guaranteesthe prioritization of the emergency CN traffic over thenonemergency PS traffic see formulas (7) and (9) Note thatthis model is based on the ARP and QCI values defined inthe LTE standard as used in [4] An eventual QCI and ARPvalues change do not affect ourmechanism Indeed a suitablechoice of the values of 120572 and 120573 is possible so that the currentformulas (1) and (2) are still valid

Our classification model allows four principal types ofbearers namely PS emergency bearers which deal withthe PS communication during emergency situations the PSnonemergency bearers that depend on the nonemergency PStransmissions the CN emergency and the CNnonemergencybearers which are relevant to the urgent commercial commu-nications and the nonurgent commercial ones respectively

Authors in [4] give a classification solution by consideringthe PS emergency as the higher priority bearers groupfollowed by the PS nonemergency one The third priority isgiven to the CN emergency and the last one is given to theCN nonemergency one As we work on a shared commercialradio bands we believe that only the PS emergency traffic hasto be more prioritized than the CN emergency one becauseon one hand the PS traffic has already its dedicated radioband and on the other hand during situation of crisis thetransmission of the CN emergency traffic is very importanttoo Therefore our proposed classification scheme offers thehighest priority to the PS emergency traffic followed by theCN emergency traffic after that we give the third prioritylevel to the PS nonemergency traffic and finally the CNnonemergency traffic

32 Implementation and Description of CPAwO AlgorithmCPAwO algorithm is a generalization of CPA scheme [5]Thefirst contribution of CPAwO regarding CPA is the scalabilityof our solution The generalization of CPA enhances theradio resources allocation with constraints for 119899 frequencygroup classes with isin N CPAwO scheme is a flexible solutionsince it allows a multiple bearers gathering to construct afrequency class group (FCG) This is possible as the prioritycoefficients Φ of these bearers are within the same rangevalues [119886 119887] 119886 and 119887 isin R Moreover this approach may beused theoretically to an infinite number of FCG groups

The CPAwO algorithm is a radio resource allocationmechanism with constraints on the resources reserved foreach group of bearers FCG The quantity of the allocatedfrequency bands depends on the value of the priority of theFCG groupThis principle is applied to ensure the availabilityof a portion of resources to high priority bearers as well asto guarantee the service to lower priority bearers However

CPAwO is not limited to the management of the radioresource allocation constraints but is also interested in theredistribution of unused resources by courteous schemegiven to the disadvantaged classes This is applicable undercertain conditions called courteous conditions namely

(i) the resources assigned by courteous scheme mustbelong to the group of bearers of the adjacent highpriority Recipients of resources by courteous mech-anism must belong to adjacent groups of lowerpriority

(ii) the rate of packet loss of high priority class mustbe acceptable and below the tolerable packet lossthreshold

(iii) the rate of packet loss of the lower priority classmust be critical therefore exceeds the threshold oftolerable packet loss in order to benefit from thecourtesy

Certainly underutilized resource reallocation improves thequality of service of bearers belonging to disadvantagedfrequency group classes since it can delay the preemption ofthe lower priority active bearers in the network and succeedeven in delaying blocking bearers Besides more bearerswill be admitted in the network thanks to the resourcesavailability

Algorithms 1 2 3 and 4 detail the CPAwO algorithm andthe functions associated with it

In the initial state all the resources are available and therate of the allocated radio is nullThese resources are allocatedas far as a new establishment or modification of bearersrequest is formulated For each request the required resourcesrelated to the establishment of the bearer is estimated andthe group FCG of the bearer and the status of its use namelycrisis management or simple communication is determinedThis information is mandatory for computing the coefficientof priority of the bearer Φ (formula (1))

The classification of bearers within FCG groups is per-formed regarding the value of Φ Remember that bearershaving coefficient values of the priority coefficientΦ includedin the same interval [119886 119887] where 119886 119887 isin R will be put in thesame FCG Defining FCG group membership is relevant forthe allocation type constraints applied to the bearer of the G-CAMF which will be detailed later in this paper (formula (15)and Figure 2 generalized CAMF)

If resources are limited then upper adjacent group FCG119894+1

will be requested for additional resources Therefore thealgorithm will run the courteous scheme (Figure 3) If theconditions of courteous mechanism cited above are checkedand unused resources can meet the demand of a new bearerthe unused or extra resources assigned to the group FCG

119894will

be allocated to the FCG119894+1

groupThiswill be done temporarysince the QoS of the FCG

119894is still acceptable Otherwise the

courteous resources allocation request will be rejectedCertainly courteous resources allocation request rejec-

tion causes a deterioration of the QoS of the related FCGAbove all the demand of additional resources by courteousscheme is performed only during congestion time To addressthis issue the CAPwO presents an appropriate solution to


Algorithm CPAoWSet C = number of active bearers groupFOR EACH i = 1 to C minus 1

Set RR(FCG119894) = 0

Set RR1015840(FCG119894) = 0

ENDSet RR(FCGC) = 0WHILE Arrival(NB)

Get (RRreq 120572 Piority)Φ = 120572 lowast 120599 (NB) lowast (120573 +QCI (NB))ResoucesAllocation (RRreq FCGNB)

ENDWHILE

Algorithm 1 CPAwO algorithm

Function ResourcesAllocation(RRreq FCGi)IF RRreq lt 120593(FCGi) minus (RR(FCGi) + RR

1015840

(FCGiminus1))

New Bearer AcceptedRR (FCG

119894) = RR (FCG

119894) + RRreq

ELSE IF RRreq lt 120593(FCGi) minus 1205931015840

(FCGi+1)

CourteousAllocation(RRreq FCGi)

Else if IsOffloadingHOW

ElsePre-emption(RRreq FCGi)

ENDIF

Algorithm 2 Resources allocation function

improve the performance of disadvantaged traffic namelythe use of the unlicensed frequencies in order to acquiescethe new bearerrsquos requests when the whole radio resourcesare allocated at the macrocell A contribution is added toour mechanism for managing the radio resources allocationwithin the LTE HetNets

120574119895is the number of admitted bearers in the small cell

119895 of the macrocell LTE Therefore we note Ω as the gainin number of extra bearers admitted within the whole LTEHetNet network Following is the calculation ofΩ

Ω =

119899

sum

119895=1

120574119895 (11)

Small cells are subject to saturation After the exhaustionof available resources within the unlicensed radio no smallcell will accept any additional LTE bearer There we talkabout a lack of radio resources In such a case the CPAwOinterrupts low priority preemptable active bearers in thenetwork Preemption is done according to rules establishedby the LTE standard and applied in [4]

Indeed during congestion time and when resourcesare not sufficient to satisfy the bearerrsquos requests becauseof the saturation of the macrocells in the LTE networkthe CPAwO algorithm executes the handoff to the WLAN

Function 1 Pre-emption(RRreq FCGj)119894 = 1

LP = lowest priority in FCGjBS(119894) = ActifBearers(LP)Γ = RadioQ(BS(119894))WHILE Γ lt RRreq ampamp LP lt PriorityNBIF exist ActifBearers(LP) ampamp VP(ActifBearers) = +1Γ = RadioQ(BS(119894))119894 = 119894 + 1

BS(119894) = ActifBearers(LP)LP = Next lowest priority in FCG

119895

ENDENDWHILEIF Γ ge RRreq

RRgained = 0FOR 119894 lt Length(BS) ampamp RRgained lt RRreqPreempt BS(119894)RRgained = RRgained + RRBS(119894)ENDFORNew Bearer is Accepted

ELSENew Bearer Blocked

ENDIF

Algorithm 3 Preemption function

(HOW) algorithm HOW scheme can offer an additionalradio resources to the PSN and CN networks Furthermoreit may reduce the congestion degree by offloading the bearersfrom the macrocell to small cells inside the LTE HetNetsnetworkThough instead of rejecting bearers or interruptingothers belonging to the lower priorities ldquoHOWrdquo mechanismperforms switching of new bearers to WLAN network suchas WMN ad hoc or WiFi Note that HOW chooses toredirect a new bearer of high priority to small cells insteadof preempting a lower priority active bearer This approachensures high level of stability in the LTE HetNets network

The proposed solution focuses on two factors relevant tothe improvement of the performance of QoS in LTE HetNetsnetworks namely delaying the preemption of active bearersin the network by redirecting new bearers to small cellsand increasing the number of active bearers in the wholenetwork Remember that the overall number of active bearersthroughout the network is equal to the sum of active bearersin the macrocells and the small cells

33 Data Offloading to Small Cells Algorithm As mentionedbelow the courteous algorithm may not be applied whenresources are limited for the PS network In such casespreemption is inevitable Hence the other aspect of oursolution is to use unlicensed networks such as WiFi WMNand ad hoc networks to offload the LTEHetNets users (EUs)This should be entirely transparent to the EUs Thereforethe EU uses the unlicensed network to exchange data ratherthan LTE frequency bands This is the Device-to-Device(D2D) communicationThis approach delays the preemptionof bearers in the LTE HetNets network and reduces packetloss even more In fact new bearers will be redirected to


Function Courteous Allocation(RRreq FCGi)

Set 120596i the number of blocked bearer belonging to FCGi classeSet 120595i the FCGi classe threshold tolerated blocked bearerSet RRcourteous(FCGi+1) the maximum Radio quantity can begiven to FCGi classe by FCGi+1 classe by courteous scheme

IF 120596119894ge 120595119894ampamp RR1015840(FCG

119894+1) lt RRcourteous(FCG119894+1)

120591 = RRcourteous(FCG119894+1)minus RR1015840(FCG119894+1)

IF 120591 ge RRreqNew Bearer AcceptedRR1015840 (FCG

119894+1) = RR (FCG

119894+1) + RRreq

ENDIFENDIF

Algorithm 4 Courteous for frequencies function

Algorithm HOWΩ = 0119895 = 1

Set 119899 = number of small CellsWhile 119895 lt= 119899 ampampΩ lt RRreq120574j = 0if RRavailable (j) gt 0

if TimeLive (RRavailable (j)) gt TimeLive(NB)120574j = RRavailable (j)

EndEndΩ = Ω + 120574j119895 = 119895 + 1

End

Algorithm 5 Algorithm handoff to WLAN (HOW)

the WMN instead of preempting the active bearers to releasethe needed radio resources Offloading scheme is applied aslong as the unlicensed network is not congested Otherwisethe preemption must be applied on the lower priority activebearers

The principle of the HOW algorithm scheme is to switchthe new bearers to the small cells to themacrocell LTEHetNetnetwork during congestion times A new bearer is redirectedto small cells even if it has a higher priority On the otherhand the low priority bearers remain active in the networkand are still served by the macrocells despite the arrivalof higher priority bearers The objective is to minimize thenumber of handoff in the network However this strategyshould not affect the QoS of the high priority bearers whichwill be admitted within the public frequencies Note thatoffloading operation occurs only if the following conditionsapply

(1) resources must be exhausted within macros cellsThis is represented by formula (23) This formulainterprets the case where each reserved radio by FCG

119894

must reach the maximum amount of reservable radio

resources where 119894 is the number of group classes ofactive frequencies in the network

forall119894 isin 119862 RR (FCG119894) = 120593 (FCG

119894) (12)

(2) Radio resources have to be available in at leastone small cell and should be sufficient enough tosatisfy the request for the establishment of the arrivedbearers (Algorithm 5)

(3) The duration time of the available resources in the cellmust be greater than or equal to the duration time ofthe bearer to be established (Algorithm 5)

4 Radio Resource Allocation Models

CAMF is a new frequency allocation model with constraintsin terms of radio resources based on the CAM model [10]CAM is a bandwidth allocation constraints model built espe-cially but not exclusively for the MPLS network It focuseson improving the allocation of resources for the classes oflower priority trafficTheCAMenhances the traffic allocationin the network with respect to the two approaches MAM[11] and RDM [12] In this work we adapt the CAM for thefrequencies allocation Three models have been developednamely the basic model CAMF (CAM for frequencies) thegeneralized model G-CAMF (generalized CAMF) and RUS-CAMF model (radio usage status based courteous allocationconstraints model for frequencies)Thesemodels are detailedbelow ldquoDefined of Used Variablesrdquo Section summarizes thedifferent variables used in the two allocation models forfrequencies namely the classical approach and the courteousone

41 Classical Approach to Priority Access to the Shared Com-mercial Radio Authors in [4] describe the access solutionto the shared commercial frequency band for the PSN usersThis approach which we call classical approach in this paperillustrates the sharing of commercial radio resources betweenthe PS bearers and CN ones

In this solution CN bearers may use the whole sharedradio bands when no PS bearer is requesting resourcesAgainst PS bearers may only use part of the shared bands


even if no CN bearers request the resources On the otherhand PS dedicated frequencies allocated to CN bearers arereleased as soon as a PS bearer requests them

Many conditions apply for preemption of an active bearer(AB) in the network by a new bearer (NB) Indeed a lowerpriority bearer has no ability to interrupt another higherpriority In addition a new bearer needs to have the attributecapacity to preemption (CP) set to yes which we representin this work by the positive value +1 A bearer with a CPvalue set to no which is equal to minus1 will not be able tointerrupt the active bearers In addition the bearer whichcan be interrupted must be vulnerable to preemption Thisvalue expressed by VP must be equal to yes In our solutionwe attribute the positive value +1 to VP if the correspondingbearer is vulnerable to preemption minus1 otherwise

Based on the preemption process described in [4] wedefine the following formulas

Condition1= Pri (NB) gt Pri (BAvp)

CP (NB) = +1

VP (BAvp) = +1

(13)

Condition2= Pri (NB) le Pri (BAvp) or

CP (NB) = minus1 or

VP (BAvp) = minus1

(14)

Preemption (NBBAvp) = True if Condition

1

False if Condition2

(15)

Figure 1 shows the model of radio resource allocation basedon the classical approach [4] (Figure 1(a)) and the adaptationof the CAM [10] for frequencies (CAMF) (Figure 1(b)) Notethat CAMF is the model used by CPA as a constraintsallocation scheme

In the classical approach [4] the allocation of radioresources with constraints model can be represented by thethree following joint forms

(1) For 119894 = CN and 119895 = PS

RR119904 (119894) + RR

119904(119895) le 120593

119904 (119894) (16)

(2) with the constraint

sum

119909=CNPSRR119889 (119909) + RR

119904 (119909) le 120593Max (17)

(3) Finally

sum

119909=CNPS120593119889 (119909) + 120593119904 (

119909) ge 120593Max (18)

The classical approach of frequency allocation is modeled toshare resources between PS and CN radio networks so thata portion of the resources will be dedicated to the CN andanother will be dedicated to the PS In addition a frequencyband will be shared between the two networks To do this thefollowing conditions apply

120593d(PS)120593d(CN)

120593s(PS)

120593s(CN)

120593max

RRd(PS) RRs(PS) RRd(CN)

RRs(PS) + RRs(CN)

(a) Conventional ARP approach

120593d(CN)120593d(PS)

120593s(CN)

120593max

RRd(PS) RRd(CN)

RRs(PS) + RRs(CN)

120593998400s(PS) 120593998400s(CN)

RR998400s(CN)RR998400

s(PS)

(b) CAMF approach

Figure 1 Courteous allocation model for frequencies

(1) The shared frequency band may be allocated by thePS and CN bearers

(2) PS bearers may allocate an amount of resource radiowithin the shared frequency band But the sharedradio quantity allocated must not exceed the value120593119904(PS) This represents the maximum threshold of PS

shared radio allocation

(3) The bearers of type CN may use a portion of theshared radioThis portion can allocate the total sharedresources if there is no PS bearer which requests forthe establishment of bearer CN users release sharedradio resources dedicated to PS as soon as they areclaimed

42 Courteous Allocation Model for Frequencies For CAMFthe allocation of radio resources with courtesy is given asfollows

(1) For 119894 = CN and 119895 = PS

RR119904 (119894) + RR

119904(119895) le 120593

119904 (119894) (19)


sum

119909=CNPSRR119889 (119909) + RR

119904 (119909) le 120593Max (20)


1205932(FCG2)

1205939984002(FCG2)

120593998400i (FCGi)

120593i(FCGi)

1205939984001(FCG1)

120593iminus1(FCGiminus1)

120593nminus1(FCGnminus1)

120593n(FCGn)

120593998400n(FCGn)

120593998400nminus1(FCGnminus1)120593998400nminus2(FCGnminus2)


RRs119899(FCGn)

RR998400 s 119899minus1(F

CGnminus1)

RR998400 s 119899minus2(F

CGnminus2)

RR998400 s iminus1(F

CGiminus1)

RRs119899(FCGn) + RRs119899minus1

(FCGnminus1)

RRs119899(FCGn) + middot middot middot + RRs119894

(FCGi)


(FCGnminus1) + middot middot middot + RRs1(FCG1)



RR998400 s 1

(FCG

1)

Figure 2 Generalized CAMF

(3) For 119894 = CN and 119895 = PS

RR119904(119895) = RR1015840

119904(119895) + RR1015840

119904(119894)

RR1015840119904(119895) le 119877th119904 (119895)

(21)

(4) Finally

sum

119909=CNPS120593119889 (119909) + 120593119904 (

119909) ge 120593Max (22)

The CAMF model differs from the traditional approach bythe way the shared commercial frequency band is allocatedIndeed the portion of resource radio allocated to CN anddedicated to PS will not be released as soon as the newPS bearer call request The PS network could waive itsrequest of resources where the packet loss rate is considerable

within the CN network This is possible only when the CNpacket loss rate is greater than a threshold THpklost(CN)Furthermore the level of quality of service of the PS networkis acceptable This is interpreted by a rate of allocationof resources by PS radio below the threshold of tolerance(formula (3) of the CAMF model) [20]

43 Generalized Courteous Constraints Allocation Model forFrequencies Figure 2 shows the generalization of the CAMFmodel detailed below in this paper for FCG

119899 with 119899 isin

N considering the allocation constraints for each frequencyclass group (FCG

119896) 119896 isin 1 119862119862 is the number of active FCG

in the network Note that all bearers belonging to the sameFCG hold a coefficient of priority between cp

1and cp

2 where

cp1is the coefficient of priority corresponding to the higher


120593s(CN)

120593se(PS) + 120593se(CN) + 120593sne(PS)

120593998400se(PS) + 120593998400se(CN) + 120593998400sne(PS)

120593se(PS) + 120593se(CN)

120593998400se(PS) + 120593998400se(CN)

120593se(PS)

120593998400se(PS)

120593998400sne(CN)

120593998400sne (PS)

120593998400se (CN)

RRs119890(PS)

RR998400 s 119890(P

S)

RR998400 s 119899119890(

CN)

RR998400 s 119899119890(

CN)

RRs119890(PS) + RRs119890

(CN)

RRs119890(PS) + RRs119890

(CN) + RRs119899119890(PS)

RRs119890(PS) + RRs119890

(CN) + RRs119899119890(PS) + RRs119899119890

(CN)

Figure 3 Frequency situation usage based CAMF

priority bearer and cp2is the coefficient priority value of the

lower priority bearer both belonging to the same FCGThe main objective of modeling a courteous frequency

allocation constraints scheme is to guarantee a better QoS forthe bearers of high priority and improve the performance oflower priority bearers

In this model all active bearers belonging to the sameFCG group share the same radio resources according to thefollowing conditions

(i) All active bearers belonging to FCG119899may not use

more than 120593119899(FCG

119899) radio resources

(ii) All active bearers belonging to FCG119899and FCG

119899minus1

cannot use more than 120593119899minus1(FCG

119899minus1) + 1205931015840

119899minus1(FCG

119899minus1)

resources

(iii) And so forth

(iv) All active bearers belonging to FCG119899+ FCG

119899minus1+

sdot sdot sdot + FCG1have to use no more than 120593

119899minus1(FCG

119899minus1) +

1205931015840

119899minus1(FCG

119899minus1) + 120593119899minus2(FCG

119899minus2) + 1205931015840

119899minus2(FCG

119899minus2) + sdot sdot sdot +

1205931(FCG

1) + 1205931015840

1(FCG

1)

(v) The sum of all frequency allocated resources shouldbe less than the maximum quantity of radio resourcesthat can be allocated denoted by119872

(vi) Each bearer within FCG119896 119896 isin 1 119862 minus 1 119862 is

the number of active FCG in the network mayallocate RR

119904119896

from 120593119896(FCG

119896) resources and RR1015840

119904119896+1

from 120593119896+1(FCG

119896+1)

(vii) Higher priority bearers in FCG119899cannot allocate

added resources by courteous scheme

(viii) However the sum of 120593119896(FCG

119896) for 119896 isin 0 119862 minus 1 can

go beyond the threshold119872

Given the conditions mentioned above the mathematicalmodeling of G-CAMF model is given by the set of formulas(23)

The following joint forms represent the allocation con-straints model for frequencies relevant to the adapted CAMFfor 119899 FCG

For each bearer in FCG119896 119896 isin 1 119862 119862 is the number of

FCG groups 119862 = 119899

(1)

119888

sum

119894=119896

RR119904119894

le

119888

sum

119894=119896

120593119894le 119872 (23)



RR119904119862

+

119888minus1

sum

119894=0

(RR119904119894

+ RR1015840119904119894

) le 119872 (24)

(3) For each bearer in FCG119896 119894 isin 1 119862 minus 1 119862 = 119899

RR1015840119904119894

= RR119904119894

+ RR1015840119904119894+1

(25)

(4) For 119862 = 119899

RR1015840119904119899

= RR119904119899

(26)

(5) Finally for 119862 = 119899

119888

sum

119894=0

120593119904119894ge 119872 (27)

44 Radio Usage Situation Based Courteous Constraints Allo-cation Model for Frequencies Application of the G-CAMFfor Four Frequencies Remember that the CAMF (courteousallocationmodel for frequencies)model is based on the CAM(courteous bandwidth constraints allocation model) [10]CAMF as it has been developed in this paper manages accesspriority to the shared commercial frequency bands for PSand CN users without taking into account the urgency factorof the traffic Although the PS traffic has a higher prioritythan CN traffic but users within the same network PS or CNare considered having the same priority level Consequentlyemergency calls will be treated in the same way than the lessurgent calls In order to effectively manage crisis situationsthe information exchanged during times of disasters must besent as a prior traffic Tomeet this objective CAMF is adaptedto provide high priority traffic for traffic corresponding tocrisis management

Therefore the aim of this new approach is to give ahigher priority to the urgent traffic compared to anotherespecially prioritize the urgent traffic CN compared to non-emergency traffic PS On the other hand the courteousprocess is redesigned so that the resources released for the PSurgent traffic type will be offered for urgent traffic CN trafficonlyThis will reduce the rate of loss of PS packets during thecrisis management However nonemergency traffic of typeCN will benefit only from nonemergency traffic PS radioresources offered by courtesy

Figure 3 illustrates CAMF adaptation for shared com-mercial frequency allocation with constraints model for theemergency and nonemergency PS and CN traffic

In this work the mathematical formulation of the G-CAMF is applied to the case of the allocation of radioresources to the PS and CN networks with supporting ofthe frequency situation usages namely emergency usage andnonemergency usage This model is given as follows

(i) frequency class groups (FCG) is set of bearers belong-ing to the same network PS or CN and havingpriority coefficient belonging to the same values cp

1

and cp2

(ii) bearers are mapped to FCG as follows

(1) nonemergency CN bearers are mapped toFCG1

(2) nonemergency PS bearers are mapped to FCG2

(3) emergency CN bearers are mapped to FCG3

(4) emergency PS bearers are mapped to FCG4

Radio allocation and frequencies depend on the bearernetwork belonging PS or CN and the radio usage situationemergency or nonemergency Therefore

(1) CN nonemergency situation is represented by thevalue 1

(2) PS nonemergency situation is represented by thevalue 2

(3) CN emergency situation is represented by the value 3

(4) PS emergency situation is represented by the value 4

The following joint forms represent the allocation constraintsmodel for frequencies relevant to the adapted CAMF foremergency or nonemergency situation usages

(i) For each bearer in FCG119896 119896 isin 1 119862 119862 is the number

of FCG groups 119862 = 4

(1)

119888

sum

119894=119896

RR119904119894

le

119888

sum

119894=119896

120593119894le 120593119904 (CN) (28)


RR119904119862

+

119888minus1

sum

119894=0

(RR119904119894

+ RR1015840119904119894

) le 120593119904 (CN) (29)


RR1015840119904119894

= RR119904119894

+ RR1015840119904119894+1

(30)

(4) For 119862 = 119899

RR1015840119904119899

= RR119904119899

(31)

(5) Finally for 119862 = 119899

119888

sum

119894=0

120593119904119894ge 120593119904 (CN) (32)

Note that CAMF G-CAMF and RUS-CAMF are the fre-quencies allocation constraints model adapted by CPA andCPAwO for the allocation of the radio resources Note thatthe constraints are included in CPA and CPAwO algorithmsto assure that the constraints are applied


5 Simulation and Results

In this section the most important results related to thesimulation of CPAwO scheme are presented and commentedon The simulation is performed in Matlab The differentbearer type arrivals and finish times are generated randomlyPS traffic arrives 120 seconds after the arrival of the firstcommercial bearers The four types of generated trafficsare PS emergency PS nonemergency CN emergency andCN nonemergency traffics We simulate 2500 bearer arrivalsfor PS emergency the same number for PS nonemergency8000 calls as CN emergency and 3000 new bearers forCN nonemergency The simulation duration is about 600seconds

Note that for all simulation graphs we consider threeradio allocation schemesThe classical one [4] the courteousmodel that represents our previous contribution namelyCPA model [4] and the ldquocourteous + offloadrdquo model whichrepresents the CPAwO algorithmmodeled in this paper Notethat CPA model is an enhancement of the classical one byadapting the courteous scheme it is based on CAMF modelBeside CPAwO is based on RUS-CAMF and the classicalmodel uses the constraints adapted in the paper [4]

Otherwise remember that the CPAwO algorithm allo-cates resources within the LTEmacrocell and LTE small cellsIn this paper we use one macrocell and four small cellsThe available radio resources quantities within the small cellsare initiated randomly in the beginning of the simulationThe arrivals of bearer establishment requests to these smallcells occur with respect to the offloading macrocells bearersand the bearers finish times within these cells are generatedrandomly

Besides we suppose that the LTE HetNet is used inthis simulation in order to allow performing the D2Dcommunications although we consider that CPAwO maynot offload bearers if no resources are available in the smallcells In this case CPAwO will have the same behaviour asCPA algorithm Furthermore both CPA and CPAwO cannotapply courteous scheme if the PSN resources are limited inthe shared commercial radio Finally we suppose that the CNtraffic is higher than the PSN traffic in this case the CN trafficmay have a critical level of QoS

Figures 4 to 7 illustrate the number of active bearers in thenetwork for the four bearer types simulated within this paperThe active bearers represent the bearers that are admitted inthe network and still established within it Remember thateach bearer needs a quantity of radio resources to be admittedand active in the network As number of active bearers for akind of traffic increases it will be allocated highest quantityof resources

For the CPA results shown in Figures 4 and 5 dependingon the PS traffic the number of active bearers in the networkis decreasing when courteous model is applied comparingto the classical method this is due to the courteous functionapplied by CPA algorithmThe aim of the courteous functionis to release some PS radio resources from the higher prioritytraffic for the lower priority class traffic Indeed the numberof commercial active bearers is more important when we

00

50

100

100

150

200

200

250

300

300

350

400

400 500 600

CPA

Classical

CPAwO

Classical

Simulation time (s)

CourteousCourteous + offloading

Num

ber o

f act

ive b

eare

rs

Number of PS emergency active bearers versus simulation time

Figure 4 Number of PS emergency active bearers

apply courteous scheme (CPA) than when we apply theclassical method (Figures 6 and 7)This is possible only whenthe PS emergency traffic has a good QoS On the other handwe note an enhancement in the number of active bearersin the whole network when applying the CPAwO modelwhere the number of bearers which are active increasedfor both traffic types PS and CN and for the two situationusages emergency and nonemergency Thus the use of thepublic frequencies improves the QoS of all traffics types byincreasing the number of the admitted bearers in the network

Figures 4 to 7 show also that the three algorithms havethe same behaviour before 300 seconds This is due tocongested or noncongested network Hence when resourcesare available before 300 seconds all bearersrsquo requests aresatisfied but once the resources are limited after 300 secondsthe network become congested and the lower priority bearerswill get less resources comparing to the higher prioritybearers Note that the appearing of the congestion at 300seconds depends especially on this simulation The reasonsof this are related to the number of arrivals and departures ofbearers which are generated randomly in these experiments

On the other hand the results show that CPA algorithmrepresented by the courteous scheme in the graph in Figure 8enhances the QoS of the less priorities traffic by reducing thenumber of commercial blocking bearers Furthermore theCPAwO gets best results when Figure 8 shows that no bearersare blocked in the LTE HetNets network This performanceis reached thanks to the use of the small cells The newarrived bearers of CN types are redirected to the small cellsso all bearersrsquo establishment requests are accepted Rememberthat the macrocell offloading is done transparently fromthe clients so the use of CPAwO enhances the QoS of thecommercial users without disturbing the clients

In the same way we see in Figures 9 and 10 that thenumber of blocked bearers with PS type is increased when


00

100

100

200

200

300

300

500

600

400

400 500 600

CPA

Classical

CPAwO

Classical

Simulation time (s)


Num

ber o

f act

ive b

eare

rs

Number of PS nonemergency active bearers versus simulation time

Figure 5 Number of PS nonemergency active bearers

00

50

100

100

150

200

200

250

300

300

350

400 500 600

CPA

Classical

CPAwO

Classical

Simulation time (s)


Num

ber o

f act

ive b

eare

rs

Number of CN emergency active bearers versus simulation time

Figure 6 Number of CN emergency active bearers

applying CPA algorithm This is caused by the courteousmechanismwhich can give some radio resources to the lowerpriorities bearers as the number of blocked bearer does notreach the blocked bearer threshold In the contrary the resultsgiven by applying CPAwO illustrate that no PS bearer isblocked in the system for both emergency and nonemergencystatus usage So despite the use of the courteous mechanismin the CPAwO the courteous bearers do not suffer fromblocking phenomenon Once again this performance has

Number of CN nonemergency active bearers versus simulation time

00

2000

400

100

6000

800

200

1000

1200

300 400 500 600

CPA

Classical

Classical

Simulation time (s)


Num

ber o

f act

ive b

eare

rs

CPAwO

Figure 7 Number of CN nonemergency active bearers

Number of CN nonemergency blocked bearers versus simulation time

00

50

100

100

150

200

200

250

300 400 500 600

CPA

Classical

Classical

Simulation time (s)


Num

ber o

f blo

cked

bea

rers

Figure 8 Number of CN nonemergency blocked bearers

been reached thanks to the use of the small cells whichprovide additional radio resources for the whole network

Figure 11 represents the number of interrupted bearers inthe network Remember that CPAwO has active bearers inboth macro- and small cells so the preemption may occur inboth cells if resources are limited Note that for this simula-tion we have preempted bearers only for the nonemergencycommercial type bearers The graphs in Figure 11 show thatthe preemption rate is reduced by the courteous mechanismrepresented by the CPA algorithm This happens thanks tothe reallocation of the unused radio resources reserved to


Number of PS emergency blocked bearers versus simulation time

00

50

100

100

150

200

200

250

300

300 400 500 600

Classical

Simulation time (s)


Num

ber o

f blo

cked

bea

rers

CPA

Classical

CPAwO = null graph

Figure 9 Number of PS emergency blocked bearers

CPA

Classical

00

50

100

100

150

200

200

250

300 400 500 600

Classical

Simulation time (s)


Num

ber o

f blo

cked

bea

rers

Number of PS nonemergency blocked bearers versus simulation time

CPAwO = null graph

Figure 10 Number of PS nonemergency blocked bearers

the higher priorities traffic in order to be used by the lowpriorities traffic In this way the allocationmodel is enhancedby avoiding the underuse of frequencies bands Better yetthe CPAwO eliminates outright the preemption within theLTE HetNets network Remember that for CPAwO when anew bearer arrive to the macrocell and it cannot be admittedbecause of the lack of resources it will not interrupt anotherbearer even if it is more prioritized than the active one It willrather be redirected to the small cellThepreemption does notoccur until the resources are limited in the whole network

Number of CN nonemergency bearers versus simulation time

0 100 200 300 400 500 6000

100

200

300

400

500

600

CPA

Classical

Classical

Simulation time (s)


Num

ber o

f pre

empt

ed b

eare

rs

CPAwO = null graph

Figure 11 Number of CN nonemergency preempted bearers

This technique makes the systemmore stable by reducing thenumber of interrupted bearers

6 Conclusion

This paper investigates the priority access to the sharedcommercial radio for the public safety over the LTE hetero-geneous networks A new approach for Courteous PriorityAccess to the shared commercial radio with offloading(CPAwO) for the PSN over LTE HetNets networks is pre-sented This scheme manages the radio resources in an LTEHetNets network for assuring and enhancing the PS userrsquosaccess to the commercial shared radio over this networkThecongestion issue occurring in the LT macrocell is addressedby the offloading mechanism ldquoHOWrdquo provided by CPAwOprotocol On the other hand three frequency allocationconstraints models are modeled in this paper by namelythe CAMF the RUS-CAMF and the G-CAMF models CPAand CPAwO use the frequency allocation constraints modelsto manage the resources CAMF detailed the frequencyallocation scheme for two traffic types especially the PStraffic and the CN one As CAMFModel RUS-CAMFmodelconsiders two types of traffics namely Public Safety trafficand Commercial traffic In addition RUS-CAMF considersthe two situations usages of traffic emergency and nonemer-gency The third model G-CAMF is a generalization of theCAMF for n traffic classes and situations

The simulation results of the CPA and CPAwO mech-anisms show that while the QoS of the higher prioritytraffic is acceptable CPA algorithm enhances the QoS of thelower priorities traffic by admitting more bearers belongingto theme On the other hand CPAwO scheme correctsshortcomings of CPA while providing the same benefits andeven better for the all types of traffics in any usages situation


7 Future Works

To improve the performance of D2D communications wehave developed several algorithms Two algorithms of sig-nalling have been modeled namely RBC (reliable butterflyconstruction) algorithm [21] and LBS-AOMDV (load bal-ancing AOMDV) algorithm [22] Both of them define a setof routes connecting a source to a destination RBC buildsseveral butterfly effects in aWMN and ad hoc networksThisallows using network coding mechanism for D2D transmis-sions RBC allows also the routes backup after a butterflyfailure If RBC constructs no butterfly effect then the LBS-AOMDV algorithm will be applied to get a multipath withintheWMN Constructing several paths is very pertinent to theD2D communication since it ensures the dispatch operationsan essential mechanism for the first responders to exchangedata during situations of crisis The security aspect has notbeen neglected in our study The G-SNCDS (generalizedsecure network coding based data splitting) algorithm anextension of our SNCDS algorithm [23] will be developedfor the transmission of data based on the splitting techniquein order to address the confidentiality the integrity and theavailability attacks in the WLAN networks

Furthermore some studies are dealing with the differentcells cooperation to enhance the resources management thepower allocation and the QoS in the HetNets networks [24ndash26] Therefore to improve the resources management forthe PSN network base stations of different transmissions(macrocells microcells picocells and small cells) will beconsidered in our future studies to enhance the offloadingmechanism New schemes will be developed in order to delaythe preemption of PSN bearers during the situation of crisiswhen the resources are limited in the WLAN network

Finally a set of experiments will be carried out in orderto show the performance of CPA and CPAwO in terms ofend-to-end QoS (latency packet loss and throughput) andcongestions

Definition of Used Variables

120593119904(PS) PS shared radio resources constraint

120593119904(CN) CN shared radio resources constraint

120593119904119890(PS) PS shared radio resources constraint for

emergency situations (ES)120593119904119890(CN) CN shared radio resources constraint

for ES120593119904119899119890

(PS) PS shared radio resources constraint fornonemergency situations (NES)

120593119904119899119890

(CN) CN shared radio resources constraintfor NES

1205931015840


for ES got by courteous scheme1205931015840


ES given by courteous to CN1205931015840

119904ne(CN) CN shared radio resources constraintfor NES got by courteous scheme

1205931015840

119904ne(PS) PS shared radio resources constraint fornon-NES given by courteous to CN

RR119904(PS) PS allocated radio resources

RR119904(CN) CN allocated radio resources

RR119904119890(PS) PS allocated radio for ES

RR119904119890(CN) CN allocated radio for ES

RR119904ne(PS) PS allocated radio for NES

RR119904ne(CN) CN allocated radio for NES

RR1015840119904119890(CN) CN allocated radio for ES among the PS

shared frequencies after the courteousprocess

RR1015840119904ne(PS) PS allocated radio for NES among the PS

and CN shared frequencies for ES aftercourteous process

RR1015840119904ne(CN) CN allocated radio for NES among the

other shared frequencies after thecourteous process

119872 Maximum reservable radio resources119872 = 120593

119904(CN)

120593Max Maximum amount of radio resources120593119889(PS) PS dedicated radio resources constraint


120593119889(CN) CN dedicated radio resources


1205931015840

119904(CN) CN shared radio resources constraint got

by courteous scheme1205931015840

119904(PS) PS shared radio resources constraint given

by courteous to CNRR119889(PS) PS dedicated radio resource

RR119889(CN) PS allocated radio among the PS dedicated

frequenciesRR119904(PS) PS allocated radio among the PS shared

frequenciesRR119904(CN) CN allocated radio among the PS

dedicated frequenciesRR1015840119904(PS) PS allocated radio among the PS shared

frequencies after courteous processRR1015840119904(CN) PS allocated radio among the PS shared

frequencies after the courteous processRth119904(PS) Courteous threshold

Conflict of Interests

The authors declare that there is no conflict of interestsregarding the publication of this paper

References

[1] D Astely E Dahlman G Fodor S Parkvall and J Sachs ldquoLTErelease 12 and beyondrdquo IEEECommunicationsMagazine vol 51no 7 pp 154ndash160 2013

[2] R Q Hu and Y Qian Heterogeneous Cellular Networks JohnWiley amp Sons New York NY USA 2013

[3] T Doumi M F Dolan S Tatesh et al ldquoLTE for public safetynetworksrdquo IEEE Communications Magazine vol 51 no 2 pp106ndash112 2013

[4] S Borkar D Roberson and K Zdunek ldquoPriority Accessfor public safety on shared commercial LTE networksrdquo inProceedings of the Technical Symposium at ITU Telecom World(ITUWT rsquo11) pp 105ndash110 Geneva Switzerland October 2011

[5] C Tata and M Kadoch ldquoCourteous Allocation Model forFrequencies for Public Safety Priority Access to the Shared


Commercial Radio in LTE Heterogeneous networkrdquo presentedat the FiCloud Barcelona Spain (Submitted) 2014

[6] H Ryan and J M Peha ldquoEnabling public safety priority use ofcommercial wireless networksrdquo Homeland Security Affairs vol9 article 13 2013

[7] R Blom P D Bruin J Eman et al ldquoPublic safety commu-nication using commercial cellular technologyrdquo in Proceedingsof the 2nd International Conference on Next Generation MobileApplications Services and Technologies (NGMAST rsquo08) pp 291ndash296 Cardiff UK September 2008

[8] X Lin J G Andrews A Ghosh and R Ratasuk ldquoAnoverview on 3GPP device-to-device proximity servicesrdquo CoRRabs13100116 2013

[9] B Raghothaman E Deng R Pragada G Sternberg T Dengand K Vanganuru ldquoArchitecture and protocols for LTE-baseddevice to device communicationrdquo in Proceedings of the Interna-tional Conference on Computing Networking and Communica-tions (ICNC 13) pp 895ndash899 San Diego Calif USA January2013

[10] C Tata and M Kadoch ldquoCAM courteous bandwidth con-straints allocation modelrdquo in Proceedings of the 20th Interna-tional Conference on Telecommunications (ICT rsquo13) pp 1ndash5 May2013

[11] Le Faucheur and W Lai ldquoMaximum Allocation BandwidthConstraints Model for Diffserv-aware MPLS Traffic Engineer-ingrdquo IETF RFC 4125 June 2005

[12] F Le Faucheur ldquoRussian dolls bandwidth constraints modelfor Diffserv-aware MPLS traffic engineeringrdquo IETF RFC 41272005

[13] M Qian Y Huang J Shi Y Yuan L Tian and E DutkiewiczldquoA novel radio admission control scheme for multiclass servicesin LTE systemsrdquo inProceedings of the IEEEGlobal Telecommuni-cations Conference (GLOBECOM rsquo09) pp 1ndash6 IEEE HonoluluHawaii USA December 2009

[14] R Kwan R Arnott R Trivisonno and M Kubota ldquoOn pre-emption and congestion control for LTE systemsrdquo in Proceed-ings of the IEEE 72nd Vehicular Technology Conference Fall(VTC2010-Fall rsquo10) September 2010

[15] S Chadchan and C Akki ldquoPriority-scaled preemption of radioresources for 3GPP LTE networksrdquo International Journal ofComputer Theory and Engineering vol 3 no 6 2011

[16] A Pyattaev K Johnsson S Andreev and Y Koucheryavyldquo3GPP LTE traffic offloading onto WiFi Directrdquo in Proceedingsof the IEEE Wireless Communications and Networking Confer-ence Workshops (WCNCW rsquo13) pp 135ndash140 Shanghai ChinaApril 2013

[17] W Yoon and B Jang ldquoEnhanced non-seamless offload for LTEandWLANnetworksrdquo IEEECommunications Letters vol 17 no10 pp 1960ndash1963 2013

[18] H Ishii X Cheng S Mukherjee and B Yu ldquoAn LTE offloadsolution using small cells with D2D linksrdquo in Proceedings of theIEEE International Conference on Communications Workshops(ICC rsquo13) pp 1155ndash1160 June 2013

[19] L Hu L L Sanchez M Maternia et al ldquoModeling of Wi-Fi IEEE 80211ac offloading performance for 1000x capacityexpansion of LTE-advancedrdquo in Proceedings of the 78th IEEEVehicular Technology Conference (VTC Fall 13) pp 1ndash6 LasVegas Nev USA September 2013

[20] M Kadoch and C Tata ldquoCourteous algorithm performanceoptimization in WiMAX networksrdquo in Proceedings of the 4thInternational Conference on Communications and InformationTechnology (CIT rsquo10) pp 23ndash32 July 2010

[21] C Tata and M Kadoch ldquoRBC reliable butterfly networkconstruction algorithm for network coding in wireless meshnetworkrdquo in Proceedings of the 13th WSEAS InternationalConference on Applied Informatics and Communications (AICrsquo13) August 2013

[22] C Tata and M Kadoch ldquoMultipath routing algorithm fordevice-to-device communications for public safety over LTEheterogeneous networksrdquo in Proceedings of the 1st InternationalConference on Information and Communication Technologies forDisaster Management (ICT-DM rsquo14) Algiers Algeria 2014

[23] C Tata and M Kadoch ldquoSecure network coding based datasplitting for public safety D2D communications over LTEheterogeneous networksrdquo in 14th IEEE International Conferenceon Computer and Information Technology (CIT rsquo14) TenerifeSpain 2014

[24] R Q Hu and Y Qian ldquoAn energy efficient and spectrumefficient wireless heterogeneous network framework for 5Gsystemsrdquo IEEECommunicationsMagazine vol 52 no 5 pp 94ndash101 2014

[25] Q C Li R Q Hu Y Xu and Y Qian ldquoOptimal fractional fre-quency reuse and power control in the heterogeneous wirelessnetworksrdquo IEEE Transactions on Wireless Communications vol12 no 6 pp 2658ndash2668 2013

[26] Q Li R Q Hu Y Qian and G Wu ldquoIntracell cooperationand resource allocation in a heterogeneous networkwith relaysrdquoIEEE Transactions on Vehicular Technology vol 62 no 4 pp1770ndash1784 2013

International Journal of

AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

RoboticsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014


Active and Passive Electronic Components

Control Scienceand Engineering

Journal of



RotatingMachinery


Hindawi Publishing Corporation httpwwwhindawicom

Journal ofEngineeringVolume 2014

Submit your manuscripts athttpwwwhindawicom

VLSI Design



Shock and Vibration


Civil EngineeringAdvances in

Acoustics and VibrationAdvances in



Electrical and Computer Engineering

Journal of

Advances inOptoElectronics


Volume 2014

The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014

SensorsJournal of


Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014


Chemical EngineeringInternational Journal of Antennas and

Propagation




Navigation and Observation



DistributedSensor Networks



instead of preempting the active bearers to release the neededradio resources Offloading scheme is applied as long asthe unlicensed network is not congested Otherwise thepreemption must be applied on the lower priority activebearers

Different radio resource allocation constraintsmodels arebuilt formanaging the frequency resources quantity allocatedto each traffic type in order to guarantee a good qualityof service for all bearers in the network Courteous band-width allocation constraints model for frequencies (CAMF)is used to design a new frequency allocation model withconstraints in terms of radio resources based on the courteousbandwidth allocation constraints (CAM) model [10] CAMis a bandwidth allocation constraints model built especiallybut not exclusively for the MPLS network It focuses onimproving the allocation of resources for the classes of lowerpriority traffics The CAM enhances the traffic allocationin the network compared to the two approaches maxi-mum bandwidth constraints allocation model (MAM) [11]and Russian dolls bandwidth constraints allocation model(RDM) [12] In this work we adapt the CAM model forthe frequencies allocation scheme Finally three models havebeen developed namely the basic model CAMF (CAM forfrequencies) the generalized CAMF model G-CAMF andRUS-CAMF model (radio usage situation based courteousallocation constraints model for frequencies) These modelsare detailed below

Note that RUS-CAMF and G-CAMF may be used byCPAwO solution besides CAMF is used by our previoussolution CPA algorithm [5] CPA (Courteous Priority Accessalgorithm) is one of our contributions published in [5] CPAaims at enhancing the radio resources allocation for theCommercial and public safety networks The main objectiveof CPA is to allow the PS network to access the sharedcommercial radio over LTE with priority without penalizingcommercial users The PS network may give an amountof its resources if the CN one suffers from a high rate ofpacket loss When comparing with the solution given by [4]the simulation results show that CPA algorithm reduces thenumber of the commercial blocked bearers delays the CNblocking mechanism increases the number of commercialactive bearers and keeps an acceptable level of PS blockingbearers Note that when resources are limited for PS networkit may not yield any part of its dedicated resources At thistime no more courteous allocation is done and a preemptionwill be invoked

This paper is organized as follows Section 2 presents therelated work In Section 3 the CPAwO algorithm is modeledSection 4 illustrates the different radio resource allocationconstraints models Section 5 summarizes the most impor-tant simulation results and their interpretation Section 6concludes this paper Finally Section 7 presents the futureworks relevant to this study

2 Related Work

Several studies in the literature deal with the radio resourcesallocation issue in LTE network [13ndash15] The objective is to

provide efficient schemes to manage the radio resources inthe LTE and LTE HetNets networks Article [13] offers a newsolution to the radio resources allocation through a novelradio admission control (RAC)mechanism RACproposes todivide bearers into three groups namely the first groupwhichincludes the bearers vulnerable to preemption the secondgroup including the nonpreempting bearers which can inter-rupt bearers from the first group and the third one whichcontains the nonpreempting bearers that cannot interruptothers bearers Furthermore bearers in the same group mayshare the same amount of radio resourcesTherefore once thesecond group resources are limited its bearers can interruptothers bearers from the first group Authors in [14] proposea control admission mechanism combined with a congestioncontrol scheme as a solution to reduce the blocking and theloss calls probability in the system In [15] high prioritybearers interrupt lower priority active bearers until gettingthe necessary resources to reach the minimum required QoSlevel This may decrease the number of bearerrsquos preemptionin the network Some other studies address the problem ofpublic safety userrsquos access to the shared commercial radioover the LTE and the LTE HetNets The two papers [4 5]develop solutions to provide a priority access to the sharedcommercial radio for the PS clients over the LTE networkIn the first paper PS users may use only a part of the sharedradio even if no commercial user requests the radio resourcesBesides commercial users can use all the radio bands butcan be interrupted if PS bands are requested by PS clientsTherefore commercial bearers may suffer from high levelof preemption The second paper offers additional resourcesto the commercial users from the PS radio while the QoSlevel of the PS traffic remains acceptable which delays thepreemption of commercial bearers and reduces call loss

Managing radio resources is efficient to enhance the radioresources allocation in the LTE network when resourcesare available But once all resources are allocated and nopreemptable bearer is active in the system the level ofQoS of different traffics will decrease To address this issuethe bearers offloading mechanism from the macrocell tothe small cells can be applied The study [16] discussesthe offloading of bearers from an LTE macrocell to WiFinetwork This approach considers the case where the smallcells position discovery is performed by the cellular networkin order to optimise the number of D2D connections Theauthors demonstrate that their model can provide significantimprovement in capacity and energy consumption Othersolutions are presented in to deal with the enhancement ofD2D communications [17ndash19]

3 Courteous Priority Access to the SharedCommercial Radio with MacrocellOffloading for PSN over LTE HetNets

In this section we detail the PS network Courteous PriorityAccess with offloading approach to the commercial radioover the LTE HetNets namely CPAwO We offer a solutionfor Commercial Radio Access with prioritization for thePSN Our scheme gives additional radio resources to the PS



















(3)




= 001 lowast 15 lowast (8 + 9) = 255

(4)













= 3 lowast 15 lowast (8 + 9) = 765


(8)


Also


















119894will








Set RR(FCG119894) = 0

Set RR1015840(FCG119894) = 0



ENDWHILE



1015840

(FCGiminus1))


119894) = RR (FCG

119894) + RRreq


(FCGi+1)




ENDIF





Ω =

119899

sum

119895=1

120574119895 (11)






119895




ENDIF












119894+1) = RR (FCG

119894+1) + RRreq

ENDIFENDIF





EndEndΩ = Ω + 120574j119895 = 119895 + 1

End





119894




119894) (12)












CP (NB) = +1

VP (BAvp) = +1

(13)


CP (NB) = minus1 or

VP (BAvp) = minus1

(14)


1

False if Condition2

(15)



(1) For 119894 = CN and 119895 = PS

RR119904 (119894) + RR

119904(119895) le 120593

119904 (119894) (16)


sum

119909=CNPSRR119889 (119909) + RR

119904 (119909) le 120593Max (17)

(3) Finally

sum

119909=CNPS120593119889 (119909) + 120593119904 (

119909) ge 120593Max (18)


120593d(PS)120593d(CN)

120593s(PS)

120593s(CN)

120593max


RRs(PS) + RRs(CN)


120593d(CN)120593d(PS)

120593s(CN)

120593max

RRd(PS) RRd(CN)

RRs(PS) + RRs(CN)

120593998400s(PS) 120593998400s(CN)

RR998400s(CN)RR998400

s(PS)

(b) CAMF approach







(1) For 119894 = CN and 119895 = PS

RR119904 (119894) + RR

119904(119895) le 120593

119904 (119894) (19)


sum

119909=CNPSRR119889 (119909) + RR

119904 (119909) le 120593Max (20)


1205932(FCG2)

1205939984002(FCG2)

120593998400i (FCGi)

120593i(FCGi)

1205939984001(FCG1)



120593n(FCGn)

120593998400n(FCGn)



RRs119899(FCGn)

RR998400 s 119899minus1(F

CGnminus1)

RR998400 s 119899minus2(F

CGnminus2)


CGiminus1)


(FCGnminus1)


(FCGi)





RR998400 s 1

(FCG

1)


(3) For 119894 = CN and 119895 = PS

RR119904(119895) = RR1015840

119904(119895) + RR1015840

119904(119894)

RR1015840119904(119895) le 119877th119904 (119895)

(21)

(4) Finally

sum

119909=CNPS120593119889 (119909) + 120593119904 (

119909) ge 120593Max (22)




119899 with 119899 isin




1and cp

2 where



120593s(CN)

120593se(PS) + 120593se(CN) + 120593sne(PS)

120593998400se(PS) + 120593998400se(CN) + 120593998400sne(PS)

120593se(PS) + 120593se(CN)

120593998400se(PS) + 120593998400se(CN)

120593se(PS)

120593998400se(PS)

120593998400sne(CN)

120593998400sne (PS)

120593998400se (CN)

RRs119890(PS)

RR998400 s 119890(P

S)

RR998400 s 119899119890(

CN)

RR998400 s 119899119890(

CN)

RRs119890(PS) + RRs119890

(CN)

RRs119890(PS) + RRs119890

(CN) + RRs119899119890(PS)

RRs119890(PS) + RRs119890

(CN) + RRs119899119890(PS) + RRs119899119890

(CN)










119899minus1


119899minus1) + 1205931015840

119899minus1(FCG

119899minus1)

resources

(iii) And so forth


119899minus1+


119899minus1(FCG

119899minus1) +

1205931015840

119899minus1(FCG

119899minus1) + 120593119899minus2(FCG

119899minus2) + 1205931015840

119899minus2(FCG


1205931(FCG

1) + 1205931015840

1(FCG

1)




119904119896

from 120593119896(FCG


119904119896+1

from 120593119896+1(FCG

119896+1)









FCG groups 119862 = 119899

(1)

119888

sum

119894=119896

RR119904119894

le

119888

sum

119894=119896

120593119894le 119872 (23)



RR119904119862

+

119888minus1

sum

119894=0

(RR119904119894

+ RR1015840119904119894

) le 119872 (24)


RR1015840119904119894

= RR119904119894

+ RR1015840119904119894+1

(25)

(4) For 119862 = 119899

RR1015840119904119899

= RR119904119899

(26)

(5) Finally for 119862 = 119899

119888

sum

119894=0

120593119904119894ge 119872 (27)






1

and cp2














(1)

119888

sum

119894=119896

RR119904119894

le

119888

sum

119894=119896

120593119894le 120593119904 (CN) (28)


RR119904119862

+

119888minus1

sum

119894=0

(RR119904119894

+ RR1015840119904119894

) le 120593119904 (CN) (29)


RR1015840119904119894

= RR119904119894

+ RR1015840119904119894+1

(30)

(4) For 119862 = 119899

RR1015840119904119899

= RR119904119899

(31)

(5) Finally for 119862 = 119899

119888

sum

119894=0

120593119904119894ge 120593119904 (CN) (32)










00

50

100

100

150

200

200

250

300

300

350

400

400 500 600

CPA

Classical

CPAwO

Classical

Simulation time (s)


Num

ber o

f act

ive b

eare

rs








00

100

100

200

200

300

300

500

600

400

400 500 600

CPA

Classical

CPAwO

Classical

Simulation time (s)


Num

ber o

f act

ive b

eare

rs



00

50

100

100

150

200

200

250

300

300

350

400 500 600

CPA

Classical

CPAwO

Classical

Simulation time (s)


Num

ber o

f act

ive b

eare

rs





00

2000

400

100

6000

800

200

1000

1200

300 400 500 600

CPA

Classical

Classical

Simulation time (s)


Num

ber o

f act

ive b

eare

rs

CPAwO



00

50

100

100

150

200

200

250

300 400 500 600

CPA

Classical

Classical

Simulation time (s)


Num

ber o

f blo

cked

bea

rers






00

50

100

100

150

200

200

250

300

300 400 500 600

Classical

Simulation time (s)


Num

ber o

f blo

cked

bea

rers

CPA

Classical

CPAwO = null graph


CPA

Classical

00

50

100

100

150

200

200

250

300 400 500 600

Classical

Simulation time (s)


Num

ber o

f blo

cked

bea

rers


CPAwO = null graph




0 100 200 300 400 500 6000

100

200

300

400

500

600

CPA

Classical

Classical

Simulation time (s)


Num

ber o

f pre

empt

ed b

eare

rs

CPAwO = null graph



6 Conclusion




7 Future Works









for ES120593119904119899119890


120593119904119899119890


1205931015840






1205931015840















119904(CN)





1205931015840













References































RoboticsJournal of





Journal of



RotatingMachinery





VLSI Design



Shock and Vibration







Journal of



Volume 2014


SensorsJournal of





Propagation



























(3)




= 001 lowast 15 lowast (8 + 9) = 255

(4)













= 3 lowast 15 lowast (8 + 9) = 765


(8)


Also


















119894will








Set RR(FCG119894) = 0

Set RR1015840(FCG119894) = 0



ENDWHILE



1015840

(FCGiminus1))


119894) = RR (FCG

119894) + RRreq


(FCGi+1)




ENDIF





Ω =

119899

sum

119895=1

120574119895 (11)






119895




ENDIF












119894+1) = RR (FCG

119894+1) + RRreq

ENDIFENDIF





EndEndΩ = Ω + 120574j119895 = 119895 + 1

End





119894




119894) (12)












CP (NB) = +1

VP (BAvp) = +1

(13)


CP (NB) = minus1 or

VP (BAvp) = minus1

(14)


1

False if Condition2

(15)



(1) For 119894 = CN and 119895 = PS

RR119904 (119894) + RR

119904(119895) le 120593

119904 (119894) (16)


sum

119909=CNPSRR119889 (119909) + RR

119904 (119909) le 120593Max (17)

(3) Finally

sum

119909=CNPS120593119889 (119909) + 120593119904 (

119909) ge 120593Max (18)


120593d(PS)120593d(CN)

120593s(PS)

120593s(CN)

120593max


RRs(PS) + RRs(CN)


120593d(CN)120593d(PS)

120593s(CN)

120593max

RRd(PS) RRd(CN)

RRs(PS) + RRs(CN)

120593998400s(PS) 120593998400s(CN)

RR998400s(CN)RR998400

s(PS)

(b) CAMF approach







(1) For 119894 = CN and 119895 = PS

RR119904 (119894) + RR

119904(119895) le 120593

119904 (119894) (19)


sum

119909=CNPSRR119889 (119909) + RR

119904 (119909) le 120593Max (20)


1205932(FCG2)

1205939984002(FCG2)

120593998400i (FCGi)

120593i(FCGi)

1205939984001(FCG1)



120593n(FCGn)

120593998400n(FCGn)



RRs119899(FCGn)

RR998400 s 119899minus1(F

CGnminus1)

RR998400 s 119899minus2(F

CGnminus2)


CGiminus1)


(FCGnminus1)


(FCGi)





RR998400 s 1

(FCG

1)


(3) For 119894 = CN and 119895 = PS

RR119904(119895) = RR1015840

119904(119895) + RR1015840

119904(119894)

RR1015840119904(119895) le 119877th119904 (119895)

(21)

(4) Finally

sum

119909=CNPS120593119889 (119909) + 120593119904 (

119909) ge 120593Max (22)




119899 with 119899 isin




1and cp

2 where



120593s(CN)

120593se(PS) + 120593se(CN) + 120593sne(PS)

120593998400se(PS) + 120593998400se(CN) + 120593998400sne(PS)

120593se(PS) + 120593se(CN)

120593998400se(PS) + 120593998400se(CN)

120593se(PS)

120593998400se(PS)

120593998400sne(CN)

120593998400sne (PS)

120593998400se (CN)

RRs119890(PS)

RR998400 s 119890(P

S)

RR998400 s 119899119890(

CN)

RR998400 s 119899119890(

CN)

RRs119890(PS) + RRs119890

(CN)

RRs119890(PS) + RRs119890

(CN) + RRs119899119890(PS)

RRs119890(PS) + RRs119890

(CN) + RRs119899119890(PS) + RRs119899119890

(CN)










119899minus1


119899minus1) + 1205931015840

119899minus1(FCG

119899minus1)

resources

(iii) And so forth


119899minus1+


119899minus1(FCG

119899minus1) +

1205931015840

119899minus1(FCG

119899minus1) + 120593119899minus2(FCG

119899minus2) + 1205931015840

119899minus2(FCG


1205931(FCG

1) + 1205931015840

1(FCG

1)




119904119896

from 120593119896(FCG


119904119896+1

from 120593119896+1(FCG

119896+1)









FCG groups 119862 = 119899

(1)

119888

sum

119894=119896

RR119904119894

le

119888

sum

119894=119896

120593119894le 119872 (23)



RR119904119862

+

119888minus1

sum

119894=0

(RR119904119894

+ RR1015840119904119894

) le 119872 (24)


RR1015840119904119894

= RR119904119894

+ RR1015840119904119894+1

(25)

(4) For 119862 = 119899

RR1015840119904119899

= RR119904119899

(26)

(5) Finally for 119862 = 119899

119888

sum

119894=0

120593119904119894ge 119872 (27)






1

and cp2














(1)

119888

sum

119894=119896

RR119904119894

le

119888

sum

119894=119896

120593119894le 120593119904 (CN) (28)


RR119904119862

+

119888minus1

sum

119894=0

(RR119904119894

+ RR1015840119904119894

) le 120593119904 (CN) (29)


RR1015840119904119894

= RR119904119894

+ RR1015840119904119894+1

(30)

(4) For 119862 = 119899

RR1015840119904119899

= RR119904119899

(31)

(5) Finally for 119862 = 119899

119888

sum

119894=0

120593119904119894ge 120593119904 (CN) (32)










00

50

100

100

150

200

200

250

300

300

350

400

400 500 600

CPA

Classical

CPAwO

Classical

Simulation time (s)


Num

ber o

f act

ive b

eare

rs








00

100

100

200

200

300

300

500

600

400

400 500 600

CPA

Classical

CPAwO

Classical

Simulation time (s)


Num

ber o

f act

ive b

eare

rs



00

50

100

100

150

200

200

250

300

300

350

400 500 600

CPA

Classical

CPAwO

Classical

Simulation time (s)


Num

ber o

f act

ive b

eare

rs





00

2000

400

100

6000

800

200

1000

1200

300 400 500 600

CPA

Classical

Classical

Simulation time (s)


Num

ber o

f act

ive b

eare

rs

CPAwO



00

50

100

100

150

200

200

250

300 400 500 600

CPA

Classical

Classical

Simulation time (s)


Num

ber o

f blo

cked

bea

rers






00

50

100

100

150

200

200

250

300

300 400 500 600

Classical

Simulation time (s)


Num

ber o

f blo

cked

bea

rers

CPA

Classical

CPAwO = null graph


CPA

Classical

00

50

100

100

150

200

200

250

300 400 500 600

Classical

Simulation time (s)


Num

ber o

f blo

cked

bea

rers


CPAwO = null graph




0 100 200 300 400 500 6000

100

200

300

400

500

600

CPA

Classical

Classical

Simulation time (s)


Num

ber o

f pre

empt

ed b

eare

rs

CPAwO = null graph



6 Conclusion




7 Future Works









for ES120593119904119899119890


120593119904119899119890


1205931015840






1205931015840















119904(CN)





1205931015840













References































RoboticsJournal of





Journal of



RotatingMachinery





VLSI Design



Shock and Vibration







Journal of



Volume 2014


SensorsJournal of





Propagation










Also


















119894will








Set RR(FCG119894) = 0

Set RR1015840(FCG119894) = 0



ENDWHILE



1015840

(FCGiminus1))


119894) = RR (FCG

119894) + RRreq


(FCGi+1)




ENDIF





Ω =

119899

sum

119895=1

120574119895 (11)






119895




ENDIF












119894+1) = RR (FCG

119894+1) + RRreq

ENDIFENDIF





EndEndΩ = Ω + 120574j119895 = 119895 + 1

End





119894




119894) (12)












CP (NB) = +1

VP (BAvp) = +1

(13)


CP (NB) = minus1 or

VP (BAvp) = minus1

(14)


1

False if Condition2

(15)



(1) For 119894 = CN and 119895 = PS

RR119904 (119894) + RR

119904(119895) le 120593

119904 (119894) (16)


sum

119909=CNPSRR119889 (119909) + RR

119904 (119909) le 120593Max (17)

(3) Finally

sum

119909=CNPS120593119889 (119909) + 120593119904 (

119909) ge 120593Max (18)


120593d(PS)120593d(CN)

120593s(PS)

120593s(CN)

120593max


RRs(PS) + RRs(CN)


120593d(CN)120593d(PS)

120593s(CN)

120593max

RRd(PS) RRd(CN)

RRs(PS) + RRs(CN)

120593998400s(PS) 120593998400s(CN)

RR998400s(CN)RR998400

s(PS)

(b) CAMF approach







(1) For 119894 = CN and 119895 = PS

RR119904 (119894) + RR

119904(119895) le 120593

119904 (119894) (19)


sum

119909=CNPSRR119889 (119909) + RR

119904 (119909) le 120593Max (20)


1205932(FCG2)

1205939984002(FCG2)

120593998400i (FCGi)

120593i(FCGi)

1205939984001(FCG1)



120593n(FCGn)

120593998400n(FCGn)



RRs119899(FCGn)

RR998400 s 119899minus1(F

CGnminus1)

RR998400 s 119899minus2(F

CGnminus2)


CGiminus1)


(FCGnminus1)


(FCGi)





RR998400 s 1

(FCG

1)


(3) For 119894 = CN and 119895 = PS

RR119904(119895) = RR1015840

119904(119895) + RR1015840

119904(119894)

RR1015840119904(119895) le 119877th119904 (119895)

(21)

(4) Finally

sum

119909=CNPS120593119889 (119909) + 120593119904 (

119909) ge 120593Max (22)




119899 with 119899 isin




1and cp

2 where



120593s(CN)

120593se(PS) + 120593se(CN) + 120593sne(PS)

120593998400se(PS) + 120593998400se(CN) + 120593998400sne(PS)

120593se(PS) + 120593se(CN)

120593998400se(PS) + 120593998400se(CN)

120593se(PS)

120593998400se(PS)

120593998400sne(CN)

120593998400sne (PS)

120593998400se (CN)

RRs119890(PS)

RR998400 s 119890(P

S)

RR998400 s 119899119890(

CN)

RR998400 s 119899119890(

CN)

RRs119890(PS) + RRs119890

(CN)

RRs119890(PS) + RRs119890

(CN) + RRs119899119890(PS)

RRs119890(PS) + RRs119890

(CN) + RRs119899119890(PS) + RRs119899119890

(CN)










119899minus1


119899minus1) + 1205931015840

119899minus1(FCG

119899minus1)

resources

(iii) And so forth


119899minus1+


119899minus1(FCG

119899minus1) +

1205931015840

119899minus1(FCG

119899minus1) + 120593119899minus2(FCG

119899minus2) + 1205931015840

119899minus2(FCG


1205931(FCG

1) + 1205931015840

1(FCG

1)




119904119896

from 120593119896(FCG


119904119896+1

from 120593119896+1(FCG

119896+1)









FCG groups 119862 = 119899

(1)

119888

sum

119894=119896

RR119904119894

le

119888

sum

119894=119896

120593119894le 119872 (23)



RR119904119862

+

119888minus1

sum

119894=0

(RR119904119894

+ RR1015840119904119894

) le 119872 (24)


RR1015840119904119894

= RR119904119894

+ RR1015840119904119894+1

(25)

(4) For 119862 = 119899

RR1015840119904119899

= RR119904119899

(26)

(5) Finally for 119862 = 119899

119888

sum

119894=0

120593119904119894ge 119872 (27)






1

and cp2














(1)

119888

sum

119894=119896

RR119904119894

le

119888

sum

119894=119896

120593119894le 120593119904 (CN) (28)


RR119904119862

+

119888minus1

sum

119894=0

(RR119904119894

+ RR1015840119904119894

) le 120593119904 (CN) (29)


RR1015840119904119894

= RR119904119894

+ RR1015840119904119894+1

(30)

(4) For 119862 = 119899

RR1015840119904119899

= RR119904119899

(31)

(5) Finally for 119862 = 119899

119888

sum

119894=0

120593119904119894ge 120593119904 (CN) (32)










00

50

100

100

150

200

200

250

300

300

350

400

400 500 600

CPA

Classical

CPAwO

Classical

Simulation time (s)


Num

ber o

f act

ive b

eare

rs








00

100

100

200

200

300

300

500

600

400

400 500 600

CPA

Classical

CPAwO

Classical

Simulation time (s)


Num

ber o

f act

ive b

eare

rs



00

50

100

100

150

200

200

250

300

300

350

400 500 600

CPA

Classical

CPAwO

Classical

Simulation time (s)


Num

ber o

f act

ive b

eare

rs





00

2000

400

100

6000

800

200

1000

1200

300 400 500 600

CPA

Classical

Classical

Simulation time (s)


Num

ber o

f act

ive b

eare

rs

CPAwO



00

50

100

100

150

200

200

250

300 400 500 600

CPA

Classical

Classical

Simulation time (s)


Num

ber o

f blo

cked

bea

rers






00

50

100

100

150

200

200

250

300

300 400 500 600

Classical

Simulation time (s)


Num

ber o

f blo

cked

bea

rers

CPA

Classical

CPAwO = null graph


CPA

Classical

00

50

100

100

150

200

200

250

300 400 500 600

Classical

Simulation time (s)


Num

ber o

f blo

cked

bea

rers


CPAwO = null graph




0 100 200 300 400 500 6000

100

200

300

400

500

600

CPA

Classical

Classical

Simulation time (s)


Num

ber o

f pre

empt

ed b

eare

rs

CPAwO = null graph



6 Conclusion




7 Future Works









for ES120593119904119899119890


120593119904119899119890


1205931015840






1205931015840















119904(CN)





1205931015840













References































RoboticsJournal of





Journal of



RotatingMachinery





VLSI Design



Shock and Vibration







Journal of



Volume 2014


SensorsJournal of





Propagation











Set RR(FCG119894) = 0

Set RR1015840(FCG119894) = 0



ENDWHILE



1015840

(FCGiminus1))


119894) = RR (FCG

119894) + RRreq


(FCGi+1)




ENDIF





Ω =

119899

sum

119895=1

120574119895 (11)






119895




ENDIF












119894+1) = RR (FCG

119894+1) + RRreq

ENDIFENDIF





EndEndΩ = Ω + 120574j119895 = 119895 + 1

End





119894




119894) (12)












CP (NB) = +1

VP (BAvp) = +1

(13)


CP (NB) = minus1 or

VP (BAvp) = minus1

(14)


1

False if Condition2

(15)



(1) For 119894 = CN and 119895 = PS

RR119904 (119894) + RR

119904(119895) le 120593

119904 (119894) (16)


sum

119909=CNPSRR119889 (119909) + RR

119904 (119909) le 120593Max (17)

(3) Finally

sum

119909=CNPS120593119889 (119909) + 120593119904 (

119909) ge 120593Max (18)


120593d(PS)120593d(CN)

120593s(PS)

120593s(CN)

120593max


RRs(PS) + RRs(CN)


120593d(CN)120593d(PS)

120593s(CN)

120593max

RRd(PS) RRd(CN)

RRs(PS) + RRs(CN)

120593998400s(PS) 120593998400s(CN)

RR998400s(CN)RR998400

s(PS)

(b) CAMF approach







(1) For 119894 = CN and 119895 = PS

RR119904 (119894) + RR

119904(119895) le 120593

119904 (119894) (19)


sum

119909=CNPSRR119889 (119909) + RR

119904 (119909) le 120593Max (20)


1205932(FCG2)

1205939984002(FCG2)

120593998400i (FCGi)

120593i(FCGi)

1205939984001(FCG1)



120593n(FCGn)

120593998400n(FCGn)



RRs119899(FCGn)

RR998400 s 119899minus1(F

CGnminus1)

RR998400 s 119899minus2(F

CGnminus2)


CGiminus1)


(FCGnminus1)


(FCGi)





RR998400 s 1

(FCG

1)


(3) For 119894 = CN and 119895 = PS

RR119904(119895) = RR1015840

119904(119895) + RR1015840

119904(119894)

RR1015840119904(119895) le 119877th119904 (119895)

(21)

(4) Finally

sum

119909=CNPS120593119889 (119909) + 120593119904 (

119909) ge 120593Max (22)




119899 with 119899 isin




1and cp

2 where



120593s(CN)

120593se(PS) + 120593se(CN) + 120593sne(PS)

120593998400se(PS) + 120593998400se(CN) + 120593998400sne(PS)

120593se(PS) + 120593se(CN)

120593998400se(PS) + 120593998400se(CN)

120593se(PS)

120593998400se(PS)

120593998400sne(CN)

120593998400sne (PS)

120593998400se (CN)

RRs119890(PS)

RR998400 s 119890(P

S)

RR998400 s 119899119890(

CN)

RR998400 s 119899119890(

CN)

RRs119890(PS) + RRs119890

(CN)

RRs119890(PS) + RRs119890

(CN) + RRs119899119890(PS)

RRs119890(PS) + RRs119890

(CN) + RRs119899119890(PS) + RRs119899119890

(CN)










119899minus1


119899minus1) + 1205931015840

119899minus1(FCG

119899minus1)

resources

(iii) And so forth


119899minus1+


119899minus1(FCG

119899minus1) +

1205931015840

119899minus1(FCG

119899minus1) + 120593119899minus2(FCG

119899minus2) + 1205931015840

119899minus2(FCG


1205931(FCG

1) + 1205931015840

1(FCG

1)




119904119896

from 120593119896(FCG


119904119896+1

from 120593119896+1(FCG

119896+1)









FCG groups 119862 = 119899

(1)

119888

sum

119894=119896

RR119904119894

le

119888

sum

119894=119896

120593119894le 119872 (23)



RR119904119862

+

119888minus1

sum

119894=0

(RR119904119894

+ RR1015840119904119894

) le 119872 (24)


RR1015840119904119894

= RR119904119894

+ RR1015840119904119894+1

(25)

(4) For 119862 = 119899

RR1015840119904119899

= RR119904119899

(26)

(5) Finally for 119862 = 119899

119888

sum

119894=0

120593119904119894ge 119872 (27)






1

and cp2














(1)

119888

sum

119894=119896

RR119904119894

le

119888

sum

119894=119896

120593119894le 120593119904 (CN) (28)


RR119904119862

+

119888minus1

sum

119894=0

(RR119904119894

+ RR1015840119904119894

) le 120593119904 (CN) (29)


RR1015840119904119894

= RR119904119894

+ RR1015840119904119894+1

(30)

(4) For 119862 = 119899

RR1015840119904119899

= RR119904119899

(31)

(5) Finally for 119862 = 119899

119888

sum

119894=0

120593119904119894ge 120593119904 (CN) (32)










00

50

100

100

150

200

200

250

300

300

350

400

400 500 600

CPA

Classical

CPAwO

Classical

Simulation time (s)


Num

ber o

f act

ive b

eare

rs








00

100

100

200

200

300

300

500

600

400

400 500 600

CPA

Classical

CPAwO

Classical

Simulation time (s)


Num

ber o

f act

ive b

eare

rs



00

50

100

100

150

200

200

250

300

300

350

400 500 600

CPA

Classical

CPAwO

Classical

Simulation time (s)


Num

ber o

f act

ive b

eare

rs





00

2000

400

100

6000

800

200

1000

1200

300 400 500 600

CPA

Classical

Classical

Simulation time (s)


Num

ber o

f act

ive b

eare

rs

CPAwO



00

50

100

100

150

200

200

250

300 400 500 600

CPA

Classical

Classical

Simulation time (s)


Num

ber o

f blo

cked

bea

rers






00

50

100

100

150

200

200

250

300

300 400 500 600

Classical

Simulation time (s)


Num

ber o

f blo

cked

bea

rers

CPA

Classical

CPAwO = null graph


CPA

Classical

00

50

100

100

150

200

200

250

300 400 500 600

Classical

Simulation time (s)


Num

ber o

f blo

cked

bea

rers


CPAwO = null graph




0 100 200 300 400 500 6000

100

200

300

400

500

600

CPA

Classical

Classical

Simulation time (s)


Num

ber o

f pre

empt

ed b

eare

rs

CPAwO = null graph



6 Conclusion




7 Future Works









for ES120593119904119899119890


120593119904119899119890


1205931015840






1205931015840















119904(CN)





1205931015840













References































RoboticsJournal of





Journal of



RotatingMachinery





VLSI Design



Shock and Vibration







Journal of



Volume 2014


SensorsJournal of





Propagation
















119894+1) = RR (FCG

119894+1) + RRreq

ENDIFENDIF





EndEndΩ = Ω + 120574j119895 = 119895 + 1

End





119894




119894) (12)












CP (NB) = +1

VP (BAvp) = +1

(13)


CP (NB) = minus1 or

VP (BAvp) = minus1

(14)


1

False if Condition2

(15)



(1) For 119894 = CN and 119895 = PS

RR119904 (119894) + RR

119904(119895) le 120593

119904 (119894) (16)


sum

119909=CNPSRR119889 (119909) + RR

119904 (119909) le 120593Max (17)

(3) Finally

sum

119909=CNPS120593119889 (119909) + 120593119904 (

119909) ge 120593Max (18)


120593d(PS)120593d(CN)

120593s(PS)

120593s(CN)

120593max


RRs(PS) + RRs(CN)


120593d(CN)120593d(PS)

120593s(CN)

120593max

RRd(PS) RRd(CN)

RRs(PS) + RRs(CN)

120593998400s(PS) 120593998400s(CN)

RR998400s(CN)RR998400

s(PS)

(b) CAMF approach







(1) For 119894 = CN and 119895 = PS

RR119904 (119894) + RR

119904(119895) le 120593

119904 (119894) (19)


sum

119909=CNPSRR119889 (119909) + RR

119904 (119909) le 120593Max (20)


1205932(FCG2)

1205939984002(FCG2)

120593998400i (FCGi)

120593i(FCGi)

1205939984001(FCG1)



120593n(FCGn)

120593998400n(FCGn)



RRs119899(FCGn)

RR998400 s 119899minus1(F

CGnminus1)

RR998400 s 119899minus2(F

CGnminus2)


CGiminus1)


(FCGnminus1)


(FCGi)





RR998400 s 1

(FCG

1)


(3) For 119894 = CN and 119895 = PS

RR119904(119895) = RR1015840

119904(119895) + RR1015840

119904(119894)

RR1015840119904(119895) le 119877th119904 (119895)

(21)

(4) Finally

sum

119909=CNPS120593119889 (119909) + 120593119904 (

119909) ge 120593Max (22)




119899 with 119899 isin




1and cp

2 where



120593s(CN)

120593se(PS) + 120593se(CN) + 120593sne(PS)

120593998400se(PS) + 120593998400se(CN) + 120593998400sne(PS)

120593se(PS) + 120593se(CN)

120593998400se(PS) + 120593998400se(CN)

120593se(PS)

120593998400se(PS)

120593998400sne(CN)

120593998400sne (PS)

120593998400se (CN)

RRs119890(PS)

RR998400 s 119890(P

S)

RR998400 s 119899119890(

CN)

RR998400 s 119899119890(

CN)

RRs119890(PS) + RRs119890

(CN)

RRs119890(PS) + RRs119890

(CN) + RRs119899119890(PS)

RRs119890(PS) + RRs119890

(CN) + RRs119899119890(PS) + RRs119899119890

(CN)










119899minus1


119899minus1) + 1205931015840

119899minus1(FCG

119899minus1)

resources

(iii) And so forth


119899minus1+


119899minus1(FCG

119899minus1) +

1205931015840

119899minus1(FCG

119899minus1) + 120593119899minus2(FCG

119899minus2) + 1205931015840

119899minus2(FCG


1205931(FCG

1) + 1205931015840

1(FCG

1)




119904119896

from 120593119896(FCG


119904119896+1

from 120593119896+1(FCG

119896+1)









FCG groups 119862 = 119899

(1)

119888

sum

119894=119896

RR119904119894

le

119888

sum

119894=119896

120593119894le 119872 (23)



RR119904119862

+

119888minus1

sum

119894=0

(RR119904119894

+ RR1015840119904119894

) le 119872 (24)


RR1015840119904119894

= RR119904119894

+ RR1015840119904119894+1

(25)

(4) For 119862 = 119899

RR1015840119904119899

= RR119904119899

(26)

(5) Finally for 119862 = 119899

119888

sum

119894=0

120593119904119894ge 119872 (27)






1

and cp2














(1)

119888

sum

119894=119896

RR119904119894

le

119888

sum

119894=119896

120593119894le 120593119904 (CN) (28)


RR119904119862

+

119888minus1

sum

119894=0

(RR119904119894

+ RR1015840119904119894

) le 120593119904 (CN) (29)


RR1015840119904119894

= RR119904119894

+ RR1015840119904119894+1

(30)

(4) For 119862 = 119899

RR1015840119904119899

= RR119904119899

(31)

(5) Finally for 119862 = 119899

119888

sum

119894=0

120593119904119894ge 120593119904 (CN) (32)










00

50

100

100

150

200

200

250

300

300

350

400

400 500 600

CPA

Classical

CPAwO

Classical

Simulation time (s)


Num

ber o

f act

ive b

eare

rs








00

100

100

200

200

300

300

500

600

400

400 500 600

CPA

Classical

CPAwO

Classical

Simulation time (s)


Num

ber o

f act

ive b

eare

rs



00

50

100

100

150

200

200

250

300

300

350

400 500 600

CPA

Classical

CPAwO

Classical

Simulation time (s)


Num

ber o

f act

ive b

eare

rs





00

2000

400

100

6000

800

200

1000

1200

300 400 500 600

CPA

Classical

Classical

Simulation time (s)


Num

ber o

f act

ive b

eare

rs

CPAwO



00

50

100

100

150

200

200

250

300 400 500 600

CPA

Classical

Classical

Simulation time (s)


Num

ber o

f blo

cked

bea

rers






00

50

100

100

150

200

200

250

300

300 400 500 600

Classical

Simulation time (s)


Num

ber o

f blo

cked

bea

rers

CPA

Classical

CPAwO = null graph


CPA

Classical

00

50

100

100

150

200

200

250

300 400 500 600

Classical

Simulation time (s)


Num

ber o

f blo

cked

bea

rers


CPAwO = null graph




0 100 200 300 400 500 6000

100

200

300

400

500

600

CPA

Classical

Classical

Simulation time (s)


Num

ber o

f pre

empt

ed b

eare

rs

CPAwO = null graph



6 Conclusion




7 Future Works









for ES120593119904119899119890


120593119904119899119890


1205931015840






1205931015840















119904(CN)





1205931015840













References































RoboticsJournal of





Journal of



RotatingMachinery





VLSI Design



Shock and Vibration







Journal of



Volume 2014


SensorsJournal of





Propagation














CP (NB) = +1

VP (BAvp) = +1

(13)


CP (NB) = minus1 or

VP (BAvp) = minus1

(14)


1

False if Condition2

(15)



(1) For 119894 = CN and 119895 = PS

RR119904 (119894) + RR

119904(119895) le 120593

119904 (119894) (16)


sum

119909=CNPSRR119889 (119909) + RR

119904 (119909) le 120593Max (17)

(3) Finally

sum

119909=CNPS120593119889 (119909) + 120593119904 (

119909) ge 120593Max (18)


120593d(PS)120593d(CN)

120593s(PS)

120593s(CN)

120593max


RRs(PS) + RRs(CN)


120593d(CN)120593d(PS)

120593s(CN)

120593max

RRd(PS) RRd(CN)

RRs(PS) + RRs(CN)

120593998400s(PS) 120593998400s(CN)

RR998400s(CN)RR998400

s(PS)

(b) CAMF approach







(1) For 119894 = CN and 119895 = PS

RR119904 (119894) + RR

119904(119895) le 120593

119904 (119894) (19)


sum

119909=CNPSRR119889 (119909) + RR

119904 (119909) le 120593Max (20)


1205932(FCG2)

1205939984002(FCG2)

120593998400i (FCGi)

120593i(FCGi)

1205939984001(FCG1)



120593n(FCGn)

120593998400n(FCGn)



RRs119899(FCGn)

RR998400 s 119899minus1(F

CGnminus1)

RR998400 s 119899minus2(F

CGnminus2)


CGiminus1)


(FCGnminus1)


(FCGi)





RR998400 s 1

(FCG

1)


(3) For 119894 = CN and 119895 = PS

RR119904(119895) = RR1015840

119904(119895) + RR1015840

119904(119894)

RR1015840119904(119895) le 119877th119904 (119895)

(21)

(4) Finally

sum

119909=CNPS120593119889 (119909) + 120593119904 (

119909) ge 120593Max (22)




119899 with 119899 isin




1and cp

2 where



120593s(CN)

120593se(PS) + 120593se(CN) + 120593sne(PS)

120593998400se(PS) + 120593998400se(CN) + 120593998400sne(PS)

120593se(PS) + 120593se(CN)

120593998400se(PS) + 120593998400se(CN)

120593se(PS)

120593998400se(PS)

120593998400sne(CN)

120593998400sne (PS)

120593998400se (CN)

RRs119890(PS)

RR998400 s 119890(P

S)

RR998400 s 119899119890(

CN)

RR998400 s 119899119890(

CN)

RRs119890(PS) + RRs119890

(CN)

RRs119890(PS) + RRs119890

(CN) + RRs119899119890(PS)

RRs119890(PS) + RRs119890

(CN) + RRs119899119890(PS) + RRs119899119890

(CN)










119899minus1


119899minus1) + 1205931015840

119899minus1(FCG

119899minus1)

resources

(iii) And so forth


119899minus1+


119899minus1(FCG

119899minus1) +

1205931015840

119899minus1(FCG

119899minus1) + 120593119899minus2(FCG

119899minus2) + 1205931015840

119899minus2(FCG


1205931(FCG

1) + 1205931015840

1(FCG

1)




119904119896

from 120593119896(FCG


119904119896+1

from 120593119896+1(FCG

119896+1)









FCG groups 119862 = 119899

(1)

119888

sum

119894=119896

RR119904119894

le

119888

sum

119894=119896

120593119894le 119872 (23)



RR119904119862

+

119888minus1

sum

119894=0

(RR119904119894

+ RR1015840119904119894

) le 119872 (24)


RR1015840119904119894

= RR119904119894

+ RR1015840119904119894+1

(25)

(4) For 119862 = 119899

RR1015840119904119899

= RR119904119899

(26)

(5) Finally for 119862 = 119899

119888

sum

119894=0

120593119904119894ge 119872 (27)






1

and cp2














(1)

119888

sum

119894=119896

RR119904119894

le

119888

sum

119894=119896

120593119894le 120593119904 (CN) (28)


RR119904119862

+

119888minus1

sum

119894=0

(RR119904119894

+ RR1015840119904119894

) le 120593119904 (CN) (29)


RR1015840119904119894

= RR119904119894

+ RR1015840119904119894+1

(30)

(4) For 119862 = 119899

RR1015840119904119899

= RR119904119899

(31)

(5) Finally for 119862 = 119899

119888

sum

119894=0

120593119904119894ge 120593119904 (CN) (32)










00

50

100

100

150

200

200

250

300

300

350

400

400 500 600

CPA

Classical

CPAwO

Classical

Simulation time (s)


Num

ber o

f act

ive b

eare

rs








00

100

100

200

200

300

300

500

600

400

400 500 600

CPA

Classical

CPAwO

Classical

Simulation time (s)


Num

ber o

f act

ive b

eare

rs



00

50

100

100

150

200

200

250

300

300

350

400 500 600

CPA

Classical

CPAwO

Classical

Simulation time (s)


Num

ber o

f act

ive b

eare

rs





00

2000

400

100

6000

800

200

1000

1200

300 400 500 600

CPA

Classical

Classical

Simulation time (s)


Num

ber o

f act

ive b

eare

rs

CPAwO



00

50

100

100

150

200

200

250

300 400 500 600

CPA

Classical

Classical

Simulation time (s)


Num

ber o

f blo

cked

bea

rers






00

50

100

100

150

200

200

250

300

300 400 500 600

Classical

Simulation time (s)


Num

ber o

f blo

cked

bea

rers

CPA

Classical

CPAwO = null graph


CPA

Classical

00

50

100

100

150

200

200

250

300 400 500 600

Classical

Simulation time (s)


Num

ber o

f blo

cked

bea

rers


CPAwO = null graph




0 100 200 300 400 500 6000

100

200

300

400

500

600

CPA

Classical

Classical

Simulation time (s)


Num

ber o

f pre

empt

ed b

eare

rs

CPAwO = null graph



6 Conclusion




7 Future Works









for ES120593119904119899119890


120593119904119899119890


1205931015840






1205931015840















119904(CN)





1205931015840













References































RoboticsJournal of





Journal of



RotatingMachinery





VLSI Design



Shock and Vibration







Journal of



Volume 2014


SensorsJournal of





Propagation










1205932(FCG2)

1205939984002(FCG2)

120593998400i (FCGi)

120593i(FCGi)

1205939984001(FCG1)



120593n(FCGn)

120593998400n(FCGn)



RRs119899(FCGn)

RR998400 s 119899minus1(F

CGnminus1)

RR998400 s 119899minus2(F

CGnminus2)


CGiminus1)


(FCGnminus1)


(FCGi)





RR998400 s 1

(FCG

1)


(3) For 119894 = CN and 119895 = PS

RR119904(119895) = RR1015840

119904(119895) + RR1015840

119904(119894)

RR1015840119904(119895) le 119877th119904 (119895)

(21)

(4) Finally

sum

119909=CNPS120593119889 (119909) + 120593119904 (

119909) ge 120593Max (22)




119899 with 119899 isin




1and cp

2 where



120593s(CN)

120593se(PS) + 120593se(CN) + 120593sne(PS)

120593998400se(PS) + 120593998400se(CN) + 120593998400sne(PS)

120593se(PS) + 120593se(CN)

120593998400se(PS) + 120593998400se(CN)

120593se(PS)

120593998400se(PS)

120593998400sne(CN)

120593998400sne (PS)

120593998400se (CN)

RRs119890(PS)

RR998400 s 119890(P

S)

RR998400 s 119899119890(

CN)

RR998400 s 119899119890(

CN)

RRs119890(PS) + RRs119890

(CN)

RRs119890(PS) + RRs119890

(CN) + RRs119899119890(PS)

RRs119890(PS) + RRs119890

(CN) + RRs119899119890(PS) + RRs119899119890

(CN)










119899minus1


119899minus1) + 1205931015840

119899minus1(FCG

119899minus1)

resources

(iii) And so forth


119899minus1+


119899minus1(FCG

119899minus1) +

1205931015840

119899minus1(FCG

119899minus1) + 120593119899minus2(FCG

119899minus2) + 1205931015840

119899minus2(FCG


1205931(FCG

1) + 1205931015840

1(FCG

1)




119904119896

from 120593119896(FCG


119904119896+1

from 120593119896+1(FCG

119896+1)









FCG groups 119862 = 119899

(1)

119888

sum

119894=119896

RR119904119894

le

119888

sum

119894=119896

120593119894le 119872 (23)



RR119904119862

+

119888minus1

sum

119894=0

(RR119904119894

+ RR1015840119904119894

) le 119872 (24)


RR1015840119904119894

= RR119904119894

+ RR1015840119904119894+1

(25)

(4) For 119862 = 119899

RR1015840119904119899

= RR119904119899

(26)

(5) Finally for 119862 = 119899

119888

sum

119894=0

120593119904119894ge 119872 (27)






1

and cp2














(1)

119888

sum

119894=119896

RR119904119894

le

119888

sum

119894=119896

120593119894le 120593119904 (CN) (28)


RR119904119862

+

119888minus1

sum

119894=0

(RR119904119894

+ RR1015840119904119894

) le 120593119904 (CN) (29)


RR1015840119904119894

= RR119904119894

+ RR1015840119904119894+1

(30)

(4) For 119862 = 119899

RR1015840119904119899

= RR119904119899

(31)

(5) Finally for 119862 = 119899

119888

sum

119894=0

120593119904119894ge 120593119904 (CN) (32)










00

50

100

100

150

200

200

250

300

300

350

400

400 500 600

CPA

Classical

CPAwO

Classical

Simulation time (s)


Num

ber o

f act

ive b

eare

rs








00

100

100

200

200

300

300

500

600

400

400 500 600

CPA

Classical

CPAwO

Classical

Simulation time (s)


Num

ber o

f act

ive b

eare

rs



00

50

100

100

150

200

200

250

300

300

350

400 500 600

CPA

Classical

CPAwO

Classical

Simulation time (s)


Num

ber o

f act

ive b

eare

rs





00

2000

400

100

6000

800

200

1000

1200

300 400 500 600

CPA

Classical

Classical

Simulation time (s)


Num

ber o

f act

ive b

eare

rs

CPAwO



00

50

100

100

150

200

200

250

300 400 500 600

CPA

Classical

Classical

Simulation time (s)


Num

ber o

f blo

cked

bea

rers






00

50

100

100

150

200

200

250

300

300 400 500 600

Classical

Simulation time (s)


Num

ber o

f blo

cked

bea

rers

CPA

Classical

CPAwO = null graph


CPA

Classical

00

50

100

100

150

200

200

250

300 400 500 600

Classical

Simulation time (s)


Num

ber o

f blo

cked

bea

rers


CPAwO = null graph




0 100 200 300 400 500 6000

100

200

300

400

500

600

CPA

Classical

Classical

Simulation time (s)


Num

ber o

f pre

empt

ed b

eare

rs

CPAwO = null graph



6 Conclusion




7 Future Works









for ES120593119904119899119890


120593119904119899119890


1205931015840






1205931015840















119904(CN)





1205931015840













References































RoboticsJournal of





Journal of



RotatingMachinery





VLSI Design



Shock and Vibration







Journal of



Volume 2014


SensorsJournal of





Propagation










120593s(CN)

120593se(PS) + 120593se(CN) + 120593sne(PS)

120593998400se(PS) + 120593998400se(CN) + 120593998400sne(PS)

120593se(PS) + 120593se(CN)

120593998400se(PS) + 120593998400se(CN)

120593se(PS)

120593998400se(PS)

120593998400sne(CN)

120593998400sne (PS)

120593998400se (CN)

RRs119890(PS)

RR998400 s 119890(P

S)

RR998400 s 119899119890(

CN)

RR998400 s 119899119890(

CN)

RRs119890(PS) + RRs119890

(CN)

RRs119890(PS) + RRs119890

(CN) + RRs119899119890(PS)

RRs119890(PS) + RRs119890

(CN) + RRs119899119890(PS) + RRs119899119890

(CN)










119899minus1


119899minus1) + 1205931015840

119899minus1(FCG

119899minus1)

resources

(iii) And so forth


119899minus1+


119899minus1(FCG

119899minus1) +

1205931015840

119899minus1(FCG

119899minus1) + 120593119899minus2(FCG

119899minus2) + 1205931015840

119899minus2(FCG


1205931(FCG

1) + 1205931015840

1(FCG

1)




119904119896

from 120593119896(FCG


119904119896+1

from 120593119896+1(FCG

119896+1)









FCG groups 119862 = 119899

(1)

119888

sum

119894=119896

RR119904119894

le

119888

sum

119894=119896

120593119894le 119872 (23)



RR119904119862

+

119888minus1

sum

119894=0

(RR119904119894

+ RR1015840119904119894

) le 119872 (24)


RR1015840119904119894

= RR119904119894

+ RR1015840119904119894+1

(25)

(4) For 119862 = 119899

RR1015840119904119899

= RR119904119899

(26)

(5) Finally for 119862 = 119899

119888

sum

119894=0

120593119904119894ge 119872 (27)






1

and cp2














(1)

119888

sum

119894=119896

RR119904119894

le

119888

sum

119894=119896

120593119894le 120593119904 (CN) (28)


RR119904119862

+

119888minus1

sum

119894=0

(RR119904119894

+ RR1015840119904119894

) le 120593119904 (CN) (29)


RR1015840119904119894

= RR119904119894

+ RR1015840119904119894+1

(30)

(4) For 119862 = 119899

RR1015840119904119899

= RR119904119899

(31)

(5) Finally for 119862 = 119899

119888

sum

119894=0

120593119904119894ge 120593119904 (CN) (32)










00

50

100

100

150

200

200

250

300

300

350

400

400 500 600

CPA

Classical

CPAwO

Classical

Simulation time (s)


Num

ber o

f act

ive b

eare

rs








00

100

100

200

200

300

300

500

600

400

400 500 600

CPA

Classical

CPAwO

Classical

Simulation time (s)


Num

ber o

f act

ive b

eare

rs



00

50

100

100

150

200

200

250

300

300

350

400 500 600

CPA

Classical

CPAwO

Classical

Simulation time (s)


Num

ber o

f act

ive b

eare

rs





00

2000

400

100

6000

800

200

1000

1200

300 400 500 600

CPA

Classical

Classical

Simulation time (s)


Num

ber o

f act

ive b

eare

rs

CPAwO



00

50

100

100

150

200

200

250

300 400 500 600

CPA

Classical

Classical

Simulation time (s)


Num

ber o

f blo

cked

bea

rers






00

50

100

100

150

200

200

250

300

300 400 500 600

Classical

Simulation time (s)


Num

ber o

f blo

cked

bea

rers

CPA

Classical

CPAwO = null graph


CPA

Classical

00

50

100

100

150

200

200

250

300 400 500 600

Classical

Simulation time (s)


Num

ber o

f blo

cked

bea

rers


CPAwO = null graph




0 100 200 300 400 500 6000

100

200

300

400

500

600

CPA

Classical

Classical

Simulation time (s)


Num

ber o

f pre

empt

ed b

eare

rs

CPAwO = null graph



6 Conclusion




7 Future Works









for ES120593119904119899119890


120593119904119899119890


1205931015840






1205931015840















119904(CN)





1205931015840













References































RoboticsJournal of





Journal of



RotatingMachinery





VLSI Design



Shock and Vibration







Journal of



Volume 2014


SensorsJournal of





Propagation











RR119904119862

+

119888minus1

sum

119894=0

(RR119904119894

+ RR1015840119904119894

) le 119872 (24)


RR1015840119904119894

= RR119904119894

+ RR1015840119904119894+1

(25)

(4) For 119862 = 119899

RR1015840119904119899

= RR119904119899

(26)

(5) Finally for 119862 = 119899

119888

sum

119894=0

120593119904119894ge 119872 (27)






1

and cp2














(1)

119888

sum

119894=119896

RR119904119894

le

119888

sum

119894=119896

120593119894le 120593119904 (CN) (28)


RR119904119862

+

119888minus1

sum

119894=0

(RR119904119894

+ RR1015840119904119894

) le 120593119904 (CN) (29)


RR1015840119904119894

= RR119904119894

+ RR1015840119904119894+1

(30)

(4) For 119862 = 119899

RR1015840119904119899

= RR119904119899

(31)

(5) Finally for 119862 = 119899

119888

sum

119894=0

120593119904119894ge 120593119904 (CN) (32)










00

50

100

100

150

200

200

250

300

300

350

400

400 500 600

CPA

Classical

CPAwO

Classical

Simulation time (s)


Num

ber o

f act

ive b

eare

rs








00

100

100

200

200

300

300

500

600

400

400 500 600

CPA

Classical

CPAwO

Classical

Simulation time (s)


Num

ber o

f act

ive b

eare

rs



00

50

100

100

150

200

200

250

300

300

350

400 500 600

CPA

Classical

CPAwO

Classical

Simulation time (s)


Num

ber o

f act

ive b

eare

rs





00

2000

400

100

6000

800

200

1000

1200

300 400 500 600

CPA

Classical

Classical

Simulation time (s)


Num

ber o

f act

ive b

eare

rs

CPAwO



00

50

100

100

150

200

200

250

300 400 500 600

CPA

Classical

Classical

Simulation time (s)


Num

ber o

f blo

cked

bea

rers






00

50

100

100

150

200

200

250

300

300 400 500 600

Classical

Simulation time (s)


Num

ber o

f blo

cked

bea

rers

CPA

Classical

CPAwO = null graph


CPA

Classical

00

50

100

100

150

200

200

250

300 400 500 600

Classical

Simulation time (s)


Num

ber o

f blo

cked

bea

rers


CPAwO = null graph




0 100 200 300 400 500 6000

100

200

300

400

500

600

CPA

Classical

Classical

Simulation time (s)


Num

ber o

f pre

empt

ed b

eare

rs

CPAwO = null graph



6 Conclusion




7 Future Works









for ES120593119904119899119890


120593119904119899119890


1205931015840






1205931015840















119904(CN)





1205931015840













References































RoboticsJournal of





Journal of



RotatingMachinery





VLSI Design



Shock and Vibration







Journal of



Volume 2014


SensorsJournal of





Propagation

















00

50

100

100

150

200

200

250

300

300

350

400

400 500 600

CPA

Classical

CPAwO

Classical

Simulation time (s)


Num

ber o

f act

ive b

eare

rs








00

100

100

200

200

300

300

500

600

400

400 500 600

CPA

Classical

CPAwO

Classical

Simulation time (s)


Num

ber o

f act

ive b

eare

rs



00

50

100

100

150

200

200

250

300

300

350

400 500 600

CPA

Classical

CPAwO

Classical

Simulation time (s)


Num

ber o

f act

ive b

eare

rs





00

2000

400

100

6000

800

200

1000

1200

300 400 500 600

CPA

Classical

Classical

Simulation time (s)


Num

ber o

f act

ive b

eare

rs

CPAwO



00

50

100

100

150

200

200

250

300 400 500 600

CPA

Classical

Classical

Simulation time (s)


Num

ber o

f blo

cked

bea

rers






00

50

100

100

150

200

200

250

300

300 400 500 600

Classical

Simulation time (s)


Num

ber o

f blo

cked

bea

rers

CPA

Classical

CPAwO = null graph


CPA

Classical

00

50

100

100

150

200

200

250

300 400 500 600

Classical

Simulation time (s)


Num

ber o

f blo

cked

bea

rers


CPAwO = null graph




0 100 200 300 400 500 6000

100

200

300

400

500

600

CPA

Classical

Classical

Simulation time (s)


Num

ber o

f pre

empt

ed b

eare

rs

CPAwO = null graph



6 Conclusion




7 Future Works









for ES120593119904119899119890


120593119904119899119890


1205931015840






1205931015840















119904(CN)





1205931015840













References































RoboticsJournal of





Journal of



RotatingMachinery





VLSI Design



Shock and Vibration







Journal of



Volume 2014


SensorsJournal of





Propagation










00

100

100

200

200

300

300

500

600

400

400 500 600

CPA

Classical

CPAwO

Classical

Simulation time (s)


Num

ber o

f act

ive b

eare

rs



00

50

100

100

150

200

200

250

300

300

350

400 500 600

CPA

Classical

CPAwO

Classical

Simulation time (s)


Num

ber o

f act

ive b

eare

rs





00

2000

400

100

6000

800

200

1000

1200

300 400 500 600

CPA

Classical

Classical

Simulation time (s)


Num

ber o

f act

ive b

eare

rs

CPAwO



00

50

100

100

150

200

200

250

300 400 500 600

CPA

Classical

Classical

Simulation time (s)


Num

ber o

f blo

cked

bea

rers






00

50

100

100

150

200

200

250

300

300 400 500 600

Classical

Simulation time (s)


Num

ber o

f blo

cked

bea

rers

CPA

Classical

CPAwO = null graph


CPA

Classical

00

50

100

100

150

200

200

250

300 400 500 600

Classical

Simulation time (s)


Num

ber o

f blo

cked

bea

rers


CPAwO = null graph




0 100 200 300 400 500 6000

100

200

300

400

500

600

CPA

Classical

Classical

Simulation time (s)


Num

ber o

f pre

empt

ed b

eare

rs

CPAwO = null graph



6 Conclusion




7 Future Works









for ES120593119904119899119890


120593119904119899119890


1205931015840






1205931015840















119904(CN)





1205931015840













References































RoboticsJournal of





Journal of



RotatingMachinery





VLSI Design



Shock and Vibration







Journal of



Volume 2014


SensorsJournal of





Propagation











00

50

100

100

150

200

200

250

300

300 400 500 600

Classical

Simulation time (s)


Num

ber o

f blo

cked

bea

rers

CPA

Classical

CPAwO = null graph


CPA

Classical

00

50

100

100

150

200

200

250

300 400 500 600

Classical

Simulation time (s)


Num

ber o

f blo

cked

bea

rers


CPAwO = null graph




0 100 200 300 400 500 6000

100

200

300

400

500

600

CPA

Classical

Classical

Simulation time (s)


Num

ber o

f pre

empt

ed b

eare

rs

CPAwO = null graph



6 Conclusion




7 Future Works









for ES120593119904119899119890


120593119904119899119890


1205931015840






1205931015840















119904(CN)





1205931015840













References































RoboticsJournal of





Journal of



RotatingMachinery





VLSI Design



Shock and Vibration







Journal of



Volume 2014


SensorsJournal of





Propagation










7 Future Works









for ES120593119904119899119890


120593119904119899119890


1205931015840






1205931015840















119904(CN)





1205931015840













References































RoboticsJournal of





Journal of



RotatingMachinery





VLSI Design



Shock and Vibration







Journal of



Volume 2014


SensorsJournal of





Propagation


































RoboticsJournal of





Journal of



RotatingMachinery





VLSI Design



Shock and Vibration







Journal of



Volume 2014


SensorsJournal of





Propagation









research article efficient priority access to the shared...

Documents