research article vertical handover decision algorithm

Research ArticleVertical Handover Decision Algorithm Using MulticriteriaMetrics in Heterogeneous Wireless Network

Gita Mahardhika Mahamod Ismail and Rosdiadee Nordin

Department of Electrical Electronics and System Faculty of Engineering and Built Environment Universiti Kebangsaan Malaysia43000 Bangi Malaysia

Correspondence should be addressed to Gita Mahardhika gitaengukmmy

Received 27 July 2014 Revised 25 November 2014 Accepted 25 November 2014

Academic Editor Rui Zhang

Copyright copy 2015 Gita Mahardhika et al 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

Heterogeneous network concept was introduced to satisfy the demands of networkrsquos traffic capacity and data rate It consists ofmultiplatformnetworkswith various radio access technologies Conventionally amobile usermay roamand accomplish the verticalhandover using single criteria such as received signal strength (RSS) Single criteria vertical handover decision however may causeinefficient handoff unbalancednetwork load and service interruptionThis paper proposed an improved vertical handover decisionusing multicriteria metrics in the environment of heterogeneous network consisting of three network interfaces (i) wireless localarea network (WLAN) (ii) wideband code division multiple access (WCDMA) and (iii) worldwide interoperability for microwaveaccess (WiMAX) In the vertical handover decision four metrics are considered (i) RSS (ii) mobile speed (iii) traffic class and (iv)network occupancyThere are three types of the vertical handover decision algorithm (i) equal priority (ii) mobile priority and (iii)network priority Equal priority multicriteria handover algorithm improved the number of handoffs by 4660 while mobile prioritymulticriteria algorithm improved the number of handoffs by 9041 and improved the balance index by 009 Network prioritymulticriteria method improved the number of handoffs by 8460 balance index by 1803 and average blocking probability by2023

1 Introduction

The demand of data rate and traffic capacity of mobilecommunication is growing rapidly thus the concept ofheterogeneous network is introduced tomeet this demand Ina heterogeneous networkmobility feature is essential becausemobile stationsmust be able to roam throughout the networkand able to connect to various radio access technologies(RATs) Conventionally mobile station considers the pointof attachment (PoA) based on single criteria such as receivedsignal strength (RSS)

The method to decide handover based on RSS is consid-ered as the simplestmethod to decide handover [1] but on theother hand itmay not have sufficient reliability because of theRSS fluctuation [2] Each network involved in heterogeneousnetwork has different threshold of RSS thus the RSS-basedmethod leads to inefficient handoff unbalanced load andservice interruption

There are issues to be addressed in vertical handoverdecision First the algorithm should have the reliability asinaccurate vertical handover decision may cost the excessiveusage of network resource Second the algorithm should havethe function as a network balancer The vertical handoverdecision algorithm should provide accuracy to balance thetraffic load for all involved networks Third the algorithmshould be accurate that it may reduce the network blockingprobability

Multicriteria decision making is an appropriate methodfor vertical handover decision because there is more thanone target network as the decision alternatives Furthermoremulticriteria decision making provides flexibility to considerseveral criteria to decide the best network

2 Related Works

There are several methods of vertical handover decisionalgorithm Generally it may be classified into five categories

Hindawi Publishing CorporationJournal of Computer Networks and CommunicationsVolume 2015 Article ID 539750 8 pageshttpdxdoiorg1011552015539750

2 Journal of Computer Networks and Communications

(i) RSS-based (ii) multicriteria (iii) context-aware (iv)cost function and (v) fuzzy logic RSS-based method is aconventional handover decision algorithmwhich uses RSS asthe handover trigger [3] and as the main criteria to decidehandover [4] RSS-based method has been optimized byadapting RSS threshold [5] and by combining RSS thresholdwith userrsquos velocity and location [6] Context-aware strategiesuse the signal quality and the context of the mobile userand the network to decide handover [7] There is a broaddefinition of context It may be defined as any informationthat can be used to characterize the situation of an entity [8]or it may be a location environment identity and time [9]

There are two approaches in cost function strategynetwork-related cost function and user-related cost function[10] Time unit monetary cost and user bit rate are severalmetrics involved in user-related cost function [11 12] Infuzzy logic strategy there are two steps of the procedurethe fuzzification and weighting procedure and the decisionmaking [13] In decisionmaking fuzzy logic methodmay useMADM (multiattributes decision making) [14] or multicrite-riamethod [15] By combiningmulticriteriamethod the han-dover decision algorithm achieved lower power consumption[16]

Inmost cases RSS-basedmethod has the lowest complex-ity compared to other methods even though the accuracy ofRSS-based method may be the least On the contrary highcomplexity methods such as fuzzy logic and cost functionprovide higher accuracy and network efficiency

There are extensive works on multicriteria vertical han-dover decision algorithmMulticriteria method has an abilityto make quantitative calculation on decision using severalcriteria between several candidates [13] The methods ofmultiplicative exponent weighting (MEW) simple additiveweighting (SAW) technique for order preference by similar-ity to ideal solution (TOPSIS) and grey relational analysis(GRA) [17] are compared and their performance (in termsof handover delay) is evaluated [18] The comparison wassimulated in the heterogeneous network environment ofWLAN UMTS and GPRS Network performance (BERdelay jitter and bandwidth) was compared

There is also a comparison and performance evalua-tion between SAW and weighted product model (WPM)in terms of processing delay [19] in the environment ofWLAN and WiMAX The results indicated that WPM hasbetter accuracy in choosing target network compared toSAW Another multicriteria method ELECTRE has beenimplemented as vertical handover decision and evaluatedusing numerical analysis ELECTRE (Elimination et CroixTraduisant la Realite or elimination and choice expressingreality) is compared to the method of SAW and TOPSIS[20] In multicriteria vertical handover decision the chosencriteria have to be appropriate to ensure the accuracy of thedecision There are many criteria user-related or network-related such as RSS mobility application and bandwidth

3 Simulation Methodology

To obtain performance results proposed method is imple-mented using MATLAB simulation There are four stages

100

100

90

90

80

80

70

70

60

60

50

50

40

40

30

30

20

20

10

10

0

0

Network topology

Distance (m)

Dist

ance

(m)

WLANWCDMAWiMAX

times102

times102

Figure 1 Network topology

of the simulation (i) modelling of heterogeneous network(ii) multicriteria vertical handover decision (iii) implemen-tation of TOPSIS multicriteria decision making into theheterogeneous network environment and (iv) performanceevaluation of the network In the first stage heterogeneousnetwork topology base station parameters andmobile stationparameters are defined In the second stage multicriteriavertical handover decision based on the technique for orderpreferences by similarity to an ideal solution (TOPSIS) is pro-posed In the third stage proposed algorithm is implementedin heterogeneous network environment In the fourth stagethe network performance is measured in terms of number ofhandoffs balance index and average blocking probability

31 Heterogeneous Network In this study heterogeneousnetwork consists of WLAN WCDMA and WiMAX asillustrated in Figure 1 The cell radius of WLAN is 700mWCDMA 23 km and WiMAX 5 km WLAN covers 7854of the simulation area meanwhile WCDMA covers 6648and WiMAX covers 7543 Base stations parameter ispresented in Table 2

There are different propagation models for each typeof network (WLAN WCDMA and WiMAX) For WiMAXnetwork empirical WiMAX propagation model is used tocalculate the coverage dimensioning

119860 = 119860119900 + 10120574 log( 119889119889119900) (1)

119860119900 is 10545 dB measured at 119889119900 = 200m Path loss exponent120574 is 3911 Lower value of path loss exponent represents themore line of sight (LoS) propagation

Journal of Computer Networks and Communications 3

In WCDMA network the network propagation modelfollows COST 231 Walfish-Ikegami model as it is a standardmodel for 3G IMT 2000UMTS [25]

119871 = 426 + 26 log (119889) + 20 log (119891) (2)

where 119889 is distance from base station to mobile user in kmand 119891 is operating frequency in MHz

Empirical one-slope model is used as the propagationmodel for WLAN [26] with formula as follows

119871 (119889) = 119871119900 + 10120574 log (119889) (3)

where 119871119900 is reference loss value for the distance of 1m 120574 ispath loss exponent and 119889 is distance in meters For operatingfrequency 245GHz 119871119900 is 40 dB and 120574 is 35 [26]

There are two mobile station parameters in this studyspeed and traffic type There are 4 (four) speed classes (i)stationary mobile stations (ii) pedestrian mobile stations(iii) low speed vehicular mobile stations and (iv) high speedvehicular mobile stations Pedestrian mobile station has thespeed of 13ms low speed vehicular mobile station of 3msand high speed vehicular mobile station of 47ms [27]

There are 3 (three) traffic classes (i) conventional service(ii) streaming service and (iii) best effort service Conven-tional service has a low delay but may be compromised inbit error rate Streaming service has average quality in bothdelay and bit error rate Best effort service has low bit errorratewithout concern in delay Bit rate for conventional serviceis 32 kbps streaming service 512 kbps and best effort service128 kbps for best effort service [8]

32 TOPSIS Multicriteria Decision Making Compared toother multicriteria methods the technique for order prefer-ences by similarity to an ideal solution (TOPSIS) has severaladvantages It has conceptual simplicity and good compu-tational efficiency and it has the ability to measure relativeperformance for each alternative [15] TOPSIS only requiresone subjective input the weightage to calculate the decisionIn simulation TOPSIS provides higher throughput and lowerpacket loss compared to other multicriteria decision makingmethods [20]

33 Multicriteria Vertical Handover Decision AlgorithmThere are 4 (four) criteria involved in the handover decisionalgorithm Those are RSSI traffic class speed and networkoccupancy Input parameters of the algorithm are basestations parameters (network topology and radio parameters)and mobile stations parameters (speed class and traffic type)The flowchart of the algorithm is presented in Figure 2

TOPSIS method provides flexibility in defining theweightage of themulticriteria priorityThere are three types ofpriority in multicriteria vertical handover decision (i) equalpriority (ii)mobile priority and (iii) network priorityMobilepriority method emphasizes mobile parameter (speed classand traffic type) meanwhile network priority emphasizesnetwork occupancy Each priority has a certain weightage aspresented in Table 3

34 Performance Measurement There are three performanceparameters (i) number of handoffs (ii) balance indexand (iii) average blocking probability (Figure 11) Numberof handoffs is a total vertical handover occurrence duringactive call Number of handoffs is an essential performanceparameter in the mobile network because it is correlated withthe signalling load and delivered QoS Handover executionrequires certain amount of network resources thus unneces-sary handover may cause inefficiency

Load balance is also essential as unbalanced load maycause congestion in one network while excess networkresources in other networks are unused [28] By creating abalanced network more mobile users may be served andthe network will have a higher throughput and lower packetdelay [29]The value of balance index represents the networkbalance where the imbalance of the network is greater withthe higher balance index value There is no unit of balanceindex and the formulation is presented in (4) Load networkis the network throughput in bps

119861 = (ΔloadWLAN-WCDMA + ΔloadWLAN-WiMAX

+ ΔloadWCDMA-WiMAX) times (3)minus1

(4)

ΔloadWLAN-WCDMA =1003816100381610038161003816loadWLAN minus loadWCDMA

1003816100381610038161003816 (5)

ΔloadWLAN-WiMAX =1003816100381610038161003816loadWLAN minus loadWiMAX

1003816100381610038161003816 (6)

ΔloadWCDMA-WiMAX =1003816100381610038161003816loadWCDMA minus loadWiMAX

1003816100381610038161003816 (7)

Blocking probability model in this study follows Erlang Bmodel withMMc queuingmodel with multiserverThe callarrival model follows Poisson process with limited servernumber 119888 Erlang B blocking probability complies with lostcall clearedmodel (Erlang lossmodel) where blocked calls areimmediately abandoned [30] Average blocking probabilityrepresents the blocking probability for the heterogeneousnetwork which consists of three types of network WLANWCDMA and WiMAX

4 Results and Discussions

There are three types of priority in multicriteria verticalhandover decision algorithm equal priority mobile priorityand network priority implemented in heterogeneous net-work environment (Table 1) The performance of each typeof priority is measured and compared to the conventionalmethod Conventional method uses RSSI as single crite-ria meanwhile multicriteria method considers traffic classspeed and network occupancy

Figure 3 presents the number of handoff distributionsof equal priority multicriteria for 1800 simulation time and200 mobile users Equal priority multicriteria method hasreduced the number of handoffs by 4660 Using conven-tional method the mean of number of handoffs is 85 anddecreased to 46 after the implementation of equal prioritymulticriteria Number of handoffs has significantly decreasedso the network efficiency is improved and extra resource ofthe network is available for other usage


Start

Read

RSSI (WLAN WCDMA and WiMAX)Traffic classMobile speed

channel occupancy(WLAN WCDMA and WiMAX)

Set priority weightage

Multicriteria decision making TOPSIS

RSSI traffic class mobile speed and network occupancy

WLAN WCDMA and WiMAX

Sort target network (descending)

Channel availability



No

No

No

Yes

Yes

YesHandover to

Handover to

Handover to

End

End

End

End

Drop call

∙ Mobile station parameter

∙ Base station parameter

∙ Criteria

∙ Alternatives

∙ Result

ge1

ge1

ge1

network number 1 network number 1



network number 1 (target number 1)network number 2network number 3 (target number 3)

Figure 2 Multicriteria vertical handover decision algorithm

Figure 4 presents the distribution function of numberof handoffs after the implementation of mobile prioritymulticriteria Mobile priority multicriteria reduced numberof handoffs by 9041 Number of handoff means by usingconventionalmethod is 63 and the implementation ofmobilepriority multicriteria has successfully dropped the number to6

Number of handoffs for network priority multicriteria ispresented in Figure 5 The improvement is 8460 wherenumber of handoff means for conventional method is 67 and

number of handoff means for network priority multicriteriais 11

Conventional method vertical handover decision createslarger number of handoffs and this may lead to additionalsignalling load on the network Mobile priority multicriteriamethod has the best performance in decreasing the numberof handoffs compared to equal priority and network priorityThe weightage of mobile priority multicriteria has a largeproportion on mobile speed and traffic class Mobile speedis linked to the number of handoffs where mobile users


Table 1 Comparison of vertical handover decision algorithms

Heuristic Advantages DisadvantagesRSS-based

Bing et al [4]Eshanta et al [3]

Low complexity Low reliability and no user consideration

MulticriteriaAlkhawlani et al [21] Low handover failure No support on fuzzy decision

Context-awareZekri et al [1]Hasswa et al [22]

High throughput Additional delay on handover information gathering process

Cost functionCalvagna and Modica [11]Ong and Khan [23]

High user satisfaction High complexity

Fuzzy logicRibeiro [14]Pahlavan et al [24]

High reliability High complexity

Table 2 Base station parameter

Parameter Unit WiMAX WCDMA WLANOperating frequency MHz 3500 2100 2450Transmitting power dBm 48 15 20Antenna gain dB 15 6 3Loss (cable combiner) dB minus3 minus3 minus3EIRP 60 18 20Path loss 16012 11845 13958Receiver sensitivity dBm minus100 minus100 minus118Antenna gain dB 3 3 3Loss (cable) dB minus3 minus3 minus3Cell radius km 5 23 07Cell edge receiving level dB minus10012 minus10045 minus11958

1

09

08

07

06

05

04

03

02

01

0

0 100 200 300 400 500 600 700

Number of handoffs

P(n

umbe

r of h

ando

ffsab

sciss

a)

Conventional methodMulticriteria equal priority

⩽

Figure 3 Number of handoff distributions of equal priority multi-criteria

with higher speed are more likely to experience ping pongeffect The emphasis of traffic class and mobile speed hassuccessfully diminished unnecessary handovers

The value of balance index indicates the traffic load bal-ance between different types of network (WLAN WCDMAand WiMAX) Figure 6 presents balance index distribution

1

09

08

07

06

05

04

03

02

01

0

0 50 100 150 200 250 300 350

Number of handoffs

P(n

umbe

r of h

ando

ffsab

sciss

a)

Conventional methodMulticriteria mobile priority

⩽

Figure 4 Number of handoff distributions of mobile prioritymulticriteria

1

09

08

07

06

05

04

03

02

01

0

0 50 100 150 200 250 300 350 400

Number of handoffs

P(n

umbe

r of h

ando

ffsab

sciss

a)

Conventional methodMulticriteria network priority

⩽

Figure 5 Number of handoff distributions of network prioritymulticriteria

for equal prioritymulticriteria For conventional (RSS-based)method the mean of balance index is 34254 Equal prioritymulticriteria slightly increased the balance index by 366 to35507


Table 3 Priority weightage

Priority type RSSI Traffic class Speed Network occupancyEqual priority 025 025 025 025Mobile priority 01 04 04 01Network priority 01 01 01 07

1

09

08

07

06

05

04

03

02

01

0

0 100 200 300 400 500 600 700 800 900

Balance index

P(b

alan

ce in

dex

absc

issa)


⩽

Figure 6 Balance index distribution of equal priority multicriteria

The implementation of mobile priority multicriteriaslightly improved the performance of balance index by 009The mean of balance index of conventional method is 28104and the mean of balance index for mobile priority multicri-teria method is 28078 The distribution of balance index formobile priority multicriteria is presented in Figure 7

Balance index is not improved by the implementationof equal priority or mobile priority multicriteria method Inequal priority weight proportion of the network criterion(network occupancy) is equal to other criteria (RSSI traffictype and speed class) meanwhile in mobile priority itemphasized the weight on mobile criteria (traffic type andspeed class) Balance index is a network-related performancemetric thus larger proportion of network criterion should begiven to improve the balance index

Network priority multicriteria improved the balanceindex by 1803Themean of balance index for conventionalmethod is 21943 and the mean of balance index for networkpriority multicriteria method is 17987 The balance indexdistribution of network priority multicriteria is presented inFigure 8

Figure 9 presents the distribution of average blockingprobability for equal priority multicriteria method Blockingprobability is a fundamental performance metric because itindicates the ability of the network to serve incoming mobileusers However equal priority multicriteria degraded theaverage blocking probability by 705 The mean of averageblocking probability for conventional method is 012 andthe mean of average blocking probability for equal prioritymulticriteria is 013 Equal priority has equal proportion forall criteria (mobile criteria and network criteria) meanwhileblocking probability is closely related to the network Largerproportion of network occupancy should be given to improveblocking probability

1

09

08

07

06

05

04

03

02

01

0

0 100 200 300 400 500 600

Balance index

P(b

alan

ce in

dex

absc

issa)


⩽

Figure 7 Balance index distribution of mobile priority multicrite-ria

1

09

08

07

06

05

04

03

02

01

0

0 100 200 300 400 500 600

Balance index

P(b

alan

ce in

dex

absc

issa)


⩽

Figure 8 Balance index distribution of network priority multicrite-ria

1

09

08

07

06

05

04

03

02

01

0


Average blocking probability0 005 01 015 02 025

P(a

vera

ge b

lock

ing

prob

abili

tyab

sciss

a)⩽

Figure 9 Average blocking probability distribution of equal prioritymulticriteria


Table 4 Performance improvement comparisons

Performance improvement ()Equal priority

Multicriteria versusconventional method

Mobile priorityMulticriteria versusconventional method

Network priorityMulticriteria versusconventional method

Number of handoffs 4660 9041 8460Balance index minus366 009 1803Average blocking probability minus705 minus869 2023

1

09

08

07

06

05

04

03

02

01

0


Average blocking probability0 004002 006 01 014 018008 012 016 02

P(a

vera

ge b

lock

ing

prob

abili

tyab

sciss

a)⩽

Figure 10 Average blocking probability distribution of mobilepriority multicriteria

1

09

08

07

06

05

04

03

02

01

0



P(a

vera

ge b

lock

ing

prob

abili

tyab

sciss

a)⩽

Figure 11 Average blocking probability distribution of networkpriority multicriteria

In the same way mobile priority multicriteria alsodegrade blocking probability Mean value of average blockingprobability increased by 869 Mean value for conventionalmethod is 011 and 012 for mobile priority multicriteriamethod The distribution is presented in Figure 10

Unlike equal priority and mobile priority network pri-ority multicriteria significantly improved average blockingprobability by 2023 Mean value of average blockingprobability for conventional method is 011 meanwhile meanvalue of average blocking probability for network prioritymulticriteria method is 009 Average blocking probabilityis linked to network occupancy and in network prioritymulticriteria method the weightage of network occupancy

is higher compared to other criteria (RSSI traffic type andspeed class) thus it successfully reduced the average blockingprobability Table 4 presents performance improvement forall multicriteria methods and performance metrics

5 Conclusion and Future Works

Compared to equal priority and mobile priority networkpriority multicriteria vertical handover decision algorithmhas improved network performance in terms of number ofhandoffs network balance and average blocking probabilityThe number of handoffs decreased by 8460 while networkbalance is improved by 2023 and average blocking proba-bility improved by 2023compared to conventionalmethodGenerally network priority multicriteria method has thebest performance of balance index and average blockingprobability compared to equal priority and mobile priority

In future works vertical handover algorithms are to beoptimized by combining multicriteria method with othervertical handover methods such as cost function or fuzzylogic The combination of multiple strategies will improvenetwork performance

Conflict of Interests

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

Acknowledgment

The paper research is supported by the Malaysian Min-istry of Higher Education under Research Grant no LRGSTD2011UKMICT0202

References

[1] M Zekri B Jouaber and D Zeghlache ldquoContext aware verticalhandover decisionmaking in heterogeneouswireless networksrdquoin Proceedings of the 35th IEEE Conference on Local ComputerNetworks (LCN rsquo10) pp 764ndash768 Denver Colo USA October2010

[2] X Yan Y A Sekercioglu and S Narayanan ldquoA survey ofvertical handover decision algorithms in Fourth Generationheterogeneous wireless networksrdquo Computer Networks vol 54no 11 pp 1848ndash1863 2010

[3] OM EshantaM Ismail K Jumari and P Yahaya ldquoVHOStrat-egy for QoS-provisioning in the WiMAXWLAN interworkingsystemrdquo Asian Journal of Applied Sciences vol 2 no 6 pp 511ndash520 2009


[4] H Bing C He and L Jiang ldquoPerformance analysis of verticalhandover in a UMTS-WLAN integrated networkrdquo in Proceed-ings of the 14th IEEE International Symposium on PersonalIndoor andMobile RadioCommunications (PIMRC 03) pp 187ndash191 September 2003

[5] W T Chen J C Liu and H K Huang ldquoAn adaptive schemefor vertical handoff in wireless overlay networksrdquo in Proceedingsof the 10th International Conference on Parallel and DistributedSystems (ICPADS rsquo04) pp 541ndash548 July 2004

[6] M Liu Z-C Li X-B Guo and H-Y Lach ldquoDesign and eval-uation of vertical handoff decision algorithm in heterogeneouswireless networksrdquo in Proceedings of the 14th IEEE InternationalConference on Networks (ICON rsquo06) pp 1ndash6 IEEE SingaporeSeptember 2006

[7] Q Wei K Farkas C Prehofer P Mendes and B PlattnerldquoContext-aware handover using active network technologyrdquoComputer Networks vol 50 no 15 pp 2855ndash2872 2006

[8] A K Dey andGD Abowd ldquoTowards a better understanding ofcontext and context-awarenessrdquo in Proceedings of the 1st Inter-national Symposium on Handheld and Ubiquitous Computingpp 1ndash12 1999

[9] N Ryan J Pascoe and D Morse ldquoEnhanced reality fieldworkthe context-aware archaeological assistantrdquo in Proceedings of the25th Anniversary Conference on the Computer Applications andQuantitativeMethods in Archaeology (CAA rsquo97) vol 750 of BARInternational Series pp 269ndash274 University of BirminghamApril 1997

[10] H-D Chu H Kim and S-J Seok ldquoFlow based 3GWLANvertical handover scheme using MIH modelrdquo in Proceedingsof the 27th International Conference on Information Networking(ICOIN 13) pp 658ndash663 Bangkok Thailand January 2013

[11] A Calvagna and G D Modica ldquoA user-centric analysis ofvertical handoversrdquo inProceedings of the 2ndACM InternationalWorkshop on Wireless Mobile Applications and Services onWLAN Hotspots pp 137ndash146 2004

[12] F Zhu and J McNair ldquoOptimizations for vertical handoffdecision algorithmsrdquo in Proceedings of the IEEE Wireless Com-munications and Networking Conference (WCNC rsquo04) pp 867ndash872 Atlanta Ga USA March 2004

[13] M Kassar B Kervella and G Pujolle ldquoAn overview of ver-tical handover decision strategies in heterogeneous wirelessnetworksrdquo Computer Communications vol 31 no 10 pp 2607ndash2620 2008

[14] R A Ribeiro ldquoFuzzy multiple attribute decision making areview and new preference elicitation techniquesrdquo Fuzzy Setsand Systems vol 78 no 2 pp 155ndash181 1996

[15] C C Hung and L H Chen ldquoA fuzzy TOPSIS decision makingmodel with entropy weight under intuitionistic fuzzy environ-mentrdquo in Proceedings of the International MultiConference ofEngineers and Computer Scientists pp 1ndash4 Hong Kong 2009

[16] I Chamodrakas and D Martakos ldquoA utility-based fuzzy TOP-SIS method for energy efficient network selection in heteroge-neous wireless networksrdquo Applied Soft Computing vol 12 no 7pp 1929ndash1938 2012

[17] D Manjaiah and P Payaswini ldquoA review of vertical handoffalgorithms based on multi attribute decision methodrdquo Interna-tional Journal of Advanced Research in Computer Engineeringand Technology vol 2 pp 2005ndash2008 2013

[18] E Stevens-Navarro and V W S Wong ldquoComparison betweenvertical handoff decision algorithms for heterogeneous wirelessnetworksrdquo in Proceedings of the 63rd IEEE Vehicular Technology

Conference (VTC rsquo06) pp 947ndash951 Melbourne Australia July2006

[19] K Savitha and C Chandrasekar ldquoVertical handover decisionschemes using SAW andWPM for network selection in hetero-geneous wireless networksrdquo Global Journal of Computer Scienceand Technology vol 11 pp 19ndash24 2011

[20] A Ismail and R Byeong-hee ldquoAdaptive handovers in hetero-geneous networks using fuzzy MADMrdquo in Proceedings of theInternational Conference onMobile IT-Convergence (ICMIC rsquo11)pp 99ndash104 September 2011

[21] M M Alkhawlani K A Alsalem and A A Hussein ldquoMulti-criteria vertical handover by TOPSIS and fuzzy logicrdquo inProceedings of the International Conference on Communicationsand Information Technology (ICCIT rsquo11) pp 96ndash102 2011

[22] A Hasswa N Nasser and H Hassanein ldquoTramcar a context-aware cross-layer architecture for next generation heteroge-neous wireless networksrdquo in Proceedings of the IEEE Interna-tional Conference on Communications (ICC rsquo06) pp 240ndash245July 2006

[23] E H Ong and J Y Khan ldquoOn optimal network selectionin a dynamic multi-RAT environmentrdquo IEEE CommunicationsLetters vol 14 no 3 pp 217ndash219 2010

[24] K Pahlavan P Krishnamurthy A Hatami et al ldquoHandoff inhybrid mobile data networksrdquo IEEE Personal Communicationsvol 7 no 2 pp 34ndash47 2000

[25] K Staniec M J Grzybkowski and K Erlebach ldquoPropagationmodelingrdquo in Understanding UMTS Radio Network ModellingPlanning and Automated Optimisation Theory and Practice MJ Nawrocki M Dohler and A H Aghvami Eds pp 69ndash113John Wiley amp Sons West Sussex UK 1st edition 2006

[26] S Zvanovec P Pechac andM Klepal ldquoWireless LAN networksdesign site survey or propagation modelingrdquo Radioengineer-ing vol 2 pp 42ndash49 2003

[27] T Casey D Denieffe and G M Muntean ldquoInfluence of mobileuser velocity on data transfer in a multi-network wirelessenvironmentrdquo in Proceedings of the 9th IFIP InternationalConference on Mobile Wireless Communications Networks pp126ndash130 2007

[28] C Hebbi and S V Saboji ldquoVertical handoff in heterogeneouswireless mobile networksrdquo in Proceedings of the 1st InternationalConference onNetworks andCommunications (NetCoM rsquo09) pp66ndash71 IEEE Chennai India December 2009

[29] H Velayos V Aleo and G Karlsson ldquoLoad balancing inoverlapping wireless LAN cellsrdquo in Proceedings of the IEEEInternational Conference on Communications pp 3833ndash3836June 2004

[30] V B Iversen Teletraffic Engineering and Network Planning 1stedition 2004

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



(i) RSS-based (ii) multicriteria (iii) context-aware (iv)cost function and (v) fuzzy logic RSS-based method is aconventional handover decision algorithmwhich uses RSS asthe handover trigger [3] and as the main criteria to decidehandover [4] RSS-based method has been optimized byadapting RSS threshold [5] and by combining RSS thresholdwith userrsquos velocity and location [6] Context-aware strategiesuse the signal quality and the context of the mobile userand the network to decide handover [7] There is a broaddefinition of context It may be defined as any informationthat can be used to characterize the situation of an entity [8]or it may be a location environment identity and time [9]

There are two approaches in cost function strategynetwork-related cost function and user-related cost function[10] Time unit monetary cost and user bit rate are severalmetrics involved in user-related cost function [11 12] Infuzzy logic strategy there are two steps of the procedurethe fuzzification and weighting procedure and the decisionmaking [13] In decisionmaking fuzzy logic methodmay useMADM (multiattributes decision making) [14] or multicrite-riamethod [15] By combiningmulticriteriamethod the han-dover decision algorithm achieved lower power consumption[16]

Inmost cases RSS-basedmethod has the lowest complex-ity compared to other methods even though the accuracy ofRSS-based method may be the least On the contrary highcomplexity methods such as fuzzy logic and cost functionprovide higher accuracy and network efficiency

There are extensive works on multicriteria vertical han-dover decision algorithmMulticriteria method has an abilityto make quantitative calculation on decision using severalcriteria between several candidates [13] The methods ofmultiplicative exponent weighting (MEW) simple additiveweighting (SAW) technique for order preference by similar-ity to ideal solution (TOPSIS) and grey relational analysis(GRA) [17] are compared and their performance (in termsof handover delay) is evaluated [18] The comparison wassimulated in the heterogeneous network environment ofWLAN UMTS and GPRS Network performance (BERdelay jitter and bandwidth) was compared

There is also a comparison and performance evalua-tion between SAW and weighted product model (WPM)in terms of processing delay [19] in the environment ofWLAN and WiMAX The results indicated that WPM hasbetter accuracy in choosing target network compared toSAW Another multicriteria method ELECTRE has beenimplemented as vertical handover decision and evaluatedusing numerical analysis ELECTRE (Elimination et CroixTraduisant la Realite or elimination and choice expressingreality) is compared to the method of SAW and TOPSIS[20] In multicriteria vertical handover decision the chosencriteria have to be appropriate to ensure the accuracy of thedecision There are many criteria user-related or network-related such as RSS mobility application and bandwidth

3 Simulation Methodology

To obtain performance results proposed method is imple-mented using MATLAB simulation There are four stages

100

100

90

90

80

80

70

70

60

60

50

50

40

40

30

30

20

20

10

10

0

0

Network topology

Distance (m)

Dist

ance

(m)

WLANWCDMAWiMAX

times102

times102

Figure 1 Network topology

of the simulation (i) modelling of heterogeneous network(ii) multicriteria vertical handover decision (iii) implemen-tation of TOPSIS multicriteria decision making into theheterogeneous network environment and (iv) performanceevaluation of the network In the first stage heterogeneousnetwork topology base station parameters andmobile stationparameters are defined In the second stage multicriteriavertical handover decision based on the technique for orderpreferences by similarity to an ideal solution (TOPSIS) is pro-posed In the third stage proposed algorithm is implementedin heterogeneous network environment In the fourth stagethe network performance is measured in terms of number ofhandoffs balance index and average blocking probability

31 Heterogeneous Network In this study heterogeneousnetwork consists of WLAN WCDMA and WiMAX asillustrated in Figure 1 The cell radius of WLAN is 700mWCDMA 23 km and WiMAX 5 km WLAN covers 7854of the simulation area meanwhile WCDMA covers 6648and WiMAX covers 7543 Base stations parameter ispresented in Table 2

There are different propagation models for each typeof network (WLAN WCDMA and WiMAX) For WiMAXnetwork empirical WiMAX propagation model is used tocalculate the coverage dimensioning

119860 = 119860119900 + 10120574 log( 119889119889119900) (1)

119860119900 is 10545 dB measured at 119889119900 = 200m Path loss exponent120574 is 3911 Lower value of path loss exponent represents themore line of sight (LoS) propagation



119871 = 426 + 26 log (119889) + 20 log (119891) (2)



119871 (119889) = 119871119900 + 10120574 log (119889) (3)











(4)


1003816100381610038161003816 (5)


1003816100381610038161003816 (6)


1003816100381610038161003816 (7)






Start

Read











No

No

No

Yes

Yes

YesHandover to

Handover to

Handover to

End

End

End

End

Drop call



∙ Criteria

∙ Alternatives

∙ Result

ge1

ge1

ge1
























1

09

08

07

06

05

04

03

02

01

0

0 100 200 300 400 500 600 700

Number of handoffs

P(n

umbe

r of h

ando

ffsab

sciss

a)


⩽




1

09

08

07

06

05

04

03

02

01

0

0 50 100 150 200 250 300 350

Number of handoffs

P(n

umbe

r of h

ando

ffsab

sciss

a)


⩽


1

09

08

07

06

05

04

03

02

01

0

0 50 100 150 200 250 300 350 400

Number of handoffs

P(n

umbe

r of h

ando

ffsab

sciss

a)


⩽






1

09

08

07

06

05

04

03

02

01

0

0 100 200 300 400 500 600 700 800 900

Balance index

P(b

alan

ce in

dex

absc

issa)


⩽






1

09

08

07

06

05

04

03

02

01

0

0 100 200 300 400 500 600

Balance index

P(b

alan

ce in

dex

absc

issa)


⩽


1

09

08

07

06

05

04

03

02

01

0

0 100 200 300 400 500 600

Balance index

P(b

alan

ce in

dex

absc

issa)


⩽


1

09

08

07

06

05

04

03

02

01

0



P(a

vera

ge b

lock

ing

prob

abili

tyab

sciss

a)⩽









1

09

08

07

06

05

04

03

02

01

0



P(a

vera

ge b

lock

ing

prob

abili

tyab

sciss

a)⩽


1

09

08

07

06

05

04

03

02

01

0



P(a

vera

ge b

lock

ing

prob

abili

tyab

sciss

a)⩽










Acknowledgment


References



































RoboticsJournal of





Journal of



RotatingMachinery





VLSI Design



Shock and Vibration







Journal of



Volume 2014


SensorsJournal of





Propagation











119871 = 426 + 26 log (119889) + 20 log (119891) (2)



119871 (119889) = 119871119900 + 10120574 log (119889) (3)











(4)


1003816100381610038161003816 (5)


1003816100381610038161003816 (6)


1003816100381610038161003816 (7)






Start

Read











No

No

No

Yes

Yes

YesHandover to

Handover to

Handover to

End

End

End

End

Drop call



∙ Criteria

∙ Alternatives

∙ Result

ge1

ge1

ge1
























1

09

08

07

06

05

04

03

02

01

0

0 100 200 300 400 500 600 700

Number of handoffs

P(n

umbe

r of h

ando

ffsab

sciss

a)


⩽




1

09

08

07

06

05

04

03

02

01

0

0 50 100 150 200 250 300 350

Number of handoffs

P(n

umbe

r of h

ando

ffsab

sciss

a)


⩽


1

09

08

07

06

05

04

03

02

01

0

0 50 100 150 200 250 300 350 400

Number of handoffs

P(n

umbe

r of h

ando

ffsab

sciss

a)


⩽






1

09

08

07

06

05

04

03

02

01

0

0 100 200 300 400 500 600 700 800 900

Balance index

P(b

alan

ce in

dex

absc

issa)


⩽






1

09

08

07

06

05

04

03

02

01

0

0 100 200 300 400 500 600

Balance index

P(b

alan

ce in

dex

absc

issa)


⩽


1

09

08

07

06

05

04

03

02

01

0

0 100 200 300 400 500 600

Balance index

P(b

alan

ce in

dex

absc

issa)


⩽


1

09

08

07

06

05

04

03

02

01

0



P(a

vera

ge b

lock

ing

prob

abili

tyab

sciss

a)⩽









1

09

08

07

06

05

04

03

02

01

0



P(a

vera

ge b

lock

ing

prob

abili

tyab

sciss

a)⩽


1

09

08

07

06

05

04

03

02

01

0



P(a

vera

ge b

lock

ing

prob

abili

tyab

sciss

a)⩽










Acknowledgment


References



































RoboticsJournal of





Journal of



RotatingMachinery





VLSI Design



Shock and Vibration







Journal of



Volume 2014


SensorsJournal of





Propagation










Start

Read











No

No

No

Yes

Yes

YesHandover to

Handover to

Handover to

End

End

End

End

Drop call



∙ Criteria

∙ Alternatives

∙ Result

ge1

ge1

ge1
























1

09

08

07

06

05

04

03

02

01

0

0 100 200 300 400 500 600 700

Number of handoffs

P(n

umbe

r of h

ando

ffsab

sciss

a)


⩽




1

09

08

07

06

05

04

03

02

01

0

0 50 100 150 200 250 300 350

Number of handoffs

P(n

umbe

r of h

ando

ffsab

sciss

a)


⩽


1

09

08

07

06

05

04

03

02

01

0

0 50 100 150 200 250 300 350 400

Number of handoffs

P(n

umbe

r of h

ando

ffsab

sciss

a)


⩽






1

09

08

07

06

05

04

03

02

01

0

0 100 200 300 400 500 600 700 800 900

Balance index

P(b

alan

ce in

dex

absc

issa)


⩽






1

09

08

07

06

05

04

03

02

01

0

0 100 200 300 400 500 600

Balance index

P(b

alan

ce in

dex

absc

issa)


⩽


1

09

08

07

06

05

04

03

02

01

0

0 100 200 300 400 500 600

Balance index

P(b

alan

ce in

dex

absc

issa)


⩽


1

09

08

07

06

05

04

03

02

01

0



P(a

vera

ge b

lock

ing

prob

abili

tyab

sciss

a)⩽









1

09

08

07

06

05

04

03

02

01

0



P(a

vera

ge b

lock

ing

prob

abili

tyab

sciss

a)⩽


1

09

08

07

06

05

04

03

02

01

0



P(a

vera

ge b

lock

ing

prob

abili

tyab

sciss

a)⩽










Acknowledgment


References



































RoboticsJournal of





Journal of



RotatingMachinery





VLSI Design



Shock and Vibration







Journal of



Volume 2014


SensorsJournal of





Propagation























1

09

08

07

06

05

04

03

02

01

0

0 100 200 300 400 500 600 700

Number of handoffs

P(n

umbe

r of h

ando

ffsab

sciss

a)


⩽




1

09

08

07

06

05

04

03

02

01

0

0 50 100 150 200 250 300 350

Number of handoffs

P(n

umbe

r of h

ando

ffsab

sciss

a)


⩽


1

09

08

07

06

05

04

03

02

01

0

0 50 100 150 200 250 300 350 400

Number of handoffs

P(n

umbe

r of h

ando

ffsab

sciss

a)


⩽






1

09

08

07

06

05

04

03

02

01

0

0 100 200 300 400 500 600 700 800 900

Balance index

P(b

alan

ce in

dex

absc

issa)


⩽






1

09

08

07

06

05

04

03

02

01

0

0 100 200 300 400 500 600

Balance index

P(b

alan

ce in

dex

absc

issa)


⩽


1

09

08

07

06

05

04

03

02

01

0

0 100 200 300 400 500 600

Balance index

P(b

alan

ce in

dex

absc

issa)


⩽


1

09

08

07

06

05

04

03

02

01

0



P(a

vera

ge b

lock

ing

prob

abili

tyab

sciss

a)⩽









1

09

08

07

06

05

04

03

02

01

0



P(a

vera

ge b

lock

ing

prob

abili

tyab

sciss

a)⩽


1

09

08

07

06

05

04

03

02

01

0



P(a

vera

ge b

lock

ing

prob

abili

tyab

sciss

a)⩽










Acknowledgment


References



































RoboticsJournal of





Journal of



RotatingMachinery





VLSI Design



Shock and Vibration







Journal of



Volume 2014


SensorsJournal of





Propagation












1

09

08

07

06

05

04

03

02

01

0

0 100 200 300 400 500 600 700 800 900

Balance index

P(b

alan

ce in

dex

absc

issa)


⩽






1

09

08

07

06

05

04

03

02

01

0

0 100 200 300 400 500 600

Balance index

P(b

alan

ce in

dex

absc

issa)


⩽


1

09

08

07

06

05

04

03

02

01

0

0 100 200 300 400 500 600

Balance index

P(b

alan

ce in

dex

absc

issa)


⩽


1

09

08

07

06

05

04

03

02

01

0



P(a

vera

ge b

lock

ing

prob

abili

tyab

sciss

a)⩽









1

09

08

07

06

05

04

03

02

01

0



P(a

vera

ge b

lock

ing

prob

abili

tyab

sciss

a)⩽


1

09

08

07

06

05

04

03

02

01

0



P(a

vera

ge b

lock

ing

prob

abili

tyab

sciss

a)⩽










Acknowledgment


References



































RoboticsJournal of





Journal of



RotatingMachinery





VLSI Design



Shock and Vibration







Journal of



Volume 2014


SensorsJournal of





Propagation
















1

09

08

07

06

05

04

03

02

01

0



P(a

vera

ge b

lock

ing

prob

abili

tyab

sciss

a)⩽


1

09

08

07

06

05

04

03

02

01

0



P(a

vera

ge b

lock

ing

prob

abili

tyab

sciss

a)⩽










Acknowledgment


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 vertical handover decision algorithm

Documents