An Intelligent Mobile Information Consultation and Sharing Multi-Agent System

Sheng-Yuan Yang Department of Information and Communication

St. John’s University New Taipei City, Taiwan, R.O.C.

Kune-Yao Chen Department of Information Management

St. John’s University New Taipei City, Taiwan, R.O.C.

Fu-Hsien Chen Department of Electrical Engineering

St. John’s University New Taipei City, Taiwan, R.O.C.

Chun-Liang Hsu Department of Electrical Engineering

St. John’s University New Taipei City, Taiwan, R.O.C.

Abstract—This paper presents Dr. What-Info II as an upgrade of Dr. What-Info, a prototype master multi-agent system asking what information it is. WIAS-supported universal application interface (UAI) technology is analyzed, designed and researched for different data formats, as a way to construct the network service system APIs of desired information with the support of OpenData@Taiwan/Taipei and domain ontology. This move not only upgrades the quality of mobile information consultation and sharing, but also enhances the correctness, authenticity and integrity of location-based information provision. Dr. What-Info II is experimentally demonstrated as a successful technology integration, and stands as an innovative piece of work in the literature.

Keywords—Open Data; Universal Application Interfaces; Multi-agent Systems

I. INTRODUCTION

The prototype system of Dr. What-Info [1] has been completed and the results are summarized as follows:

(1) The Cloud operating environment WIAS (Web-services-based Information Agent System) [2] of Cloud information exchange processing, analysis, integrated operation, decision support and network service support is constructed actively, so as to implement ubiquitous Cloud information access of mobile information equipments.

(2) The agent technology is imported to construct "tourism" related ontology (described later) to assist the users to obtain appropriate and current Cloud information rapidly, accurately and effectively. The Ontological Information Agent Shell (OntoIAS) [3] developed by this study is used in the Cloud tourism information domain, so as to augment this information agent shell EOntoIAS (Extension of OntoIAS) into CEOntoIAS (Cloud EOntoIAS, see Fig. 1) [2], supported by OntoDMA (Ontological Data Mining Agent) [4], OntoCBRA (Ontological Case-Based Reasoning Agent) [5] and OntoIAS, i.e., the three-stage

intelligent decision making, the research target of this Dr. What-Info for finding and sharing the optimal Cloud information solution is reached easily.

(a) Front-end system operation process

(b) Back-end system framework

Fig. 1. Dr. What-Info prototype system

(3) Using the photographing or video recording function of mobile information equipment, and the ontology-supported semantic analysis and Chinese and English multimedia recognition technology [1], with the support of network and back end powerful information agent system CEOntoIAS, an accurate, rapid and effective intelligent

mobile "tourism" information consultation and sharing multi-agent prototype system is built, supporting 1) QR/Bar Code reading; 2) recognize and display decoding result, the back-end system support provides suitable recommended information; 3) local GPS location capture and display; 4) target route planning and display, as shown in Fig. 2.

Fig. 2. Execution procedure of Dr. What-Info system

II. SYSTEM RESEARCH AND DEVELOPMENT TECHNIQUE AND

ARCHITECTURE

A. UAI Technology

When the Open Data platform [6,7,8] has not provided the access APIs of the corresponding database, only provides open database download hyperlink, this paper provides the solution via Dropbox Cloud space, the JSON format of Kaohsiung City government open data platform is taken as an example, described below:

(1) This access point saves the link as the corresponding .json file;

(2) Upload the file to Dropbox Cloud space, and share this access link;

(3) Modify the first half of the shared access link "https://www.dropbox.com/" as "https://dl.dropboxusercontent.com/" for APP to access the open database directly, instead of skipping to the Dropbox file download page directly, this step is of great importance;

(4) Finally, the corrected access hyperlink is copied to the object corresponding program code of access interface directly, the same access mode as access APIs of corresponding database provided by open data platform is completed for the system to implement later processing.

Therefore, for subsequent UAI access construction for the open database, as long as the user's local GPS location is converted into the corresponding address, and the suitable open database is found according to the keywords of user's current event, the corresponding map is converted according to the data capture of "address" field, the location-based mobile

information service can be improved easily. The design philosophy is: "the space is exchanged for time, the Keywords in the original address, such as the city/county of the first layer, district/town/township of the second layer, road/street of the third layer, are indexed as different Web elements in advance". In addition, the back-end system must construct the ontology corresponding to the event, linked to suitable open database; the map corresponding access APIs is converted from the open database query and query result, completing the construction of the overall open database UAI. Furthermore, how to use the back end WIAS to provide corresponding network service APIs [2] is summarized in TABLE I, the UAI consists of Connect, Compute, Search and Storage [9]. Finally, the name of address field, format and relative location of field of different open data platforms must be solved. The Web elements of MIT APP INVENTOR II have Get method, this problem can be solved easily by the JsonTextDecode method of GetText event, and the display of map can complete the display of corresponding Google map by Action.android.intent.action.VIEW with complete address, as shown in Fig. 3.

TABLE I

PARTIAL CLOUD NETWORK SERVICE APIS FUNCTION OUTLINE SERVICE NAME 

SERVICE DESCRIPTION 

TO WHOM 

UAI_ConnectCrossDomain  Connection of cross‐domain Open DB   UAI_Connect 

UAI_ConnectAlertReg  Connect registration prompt   UAI_Connect 

UAI_ConnectRuleSetting  Connect rule setting   UAI_Connect 

UAI_ComputeFileAccess  File access Compute   UAI_Compute

UAI_ComputeTimeSeriesAccess  Time series Compute   UAI_Compute

UAI_ComputeTriggerConfig  Trigger reconfiguration Compute   UAI_Compute

UAI_SearchTimeSeries  Time series search   UAI_Search 

UAI_SearchBuildIndexFields  Create index field   UAI_Search 

UAI_SearchStatistics  Search statistics   UAI_Search 

UAI_StorageFFStatus  Data folder and file status   UAI_Storage 

UAI_StorageFFManipulation  Data folder and file manipulation   UAI_Storage 

UAI_StorageFTransimission Data folder and file transmission and 

status  UAI_Storage 

UAI_StorageFFSharing Data folder and file sharing and 

integration  UAI_Storage 

     

 

Fig. 3. Three-layer address result presentation, case study of Taipei City hotel and Tainan greening restaurant open databases

B. Dr. What-Info II System Operation

The mobile equipment user uses the photographing or video recording function of mobile information equipment, the Dr. What-Info II APP (as shown in Fig. 1(a)) uses the Ubi-IA

(Ubiquitous Interface Agent in [10]) for ontology supported multimedia semantic analysis subsystem for recognition and conversion and the description of corresponding query semantic content is completed by XML, the corresponding Query Data are fed back to the Cloud server system CEOntoIAS for related Cloud information processing and query decision support operation [11], as shown in Fig. 4. The Ubi-IA provides Cloud query information processing and conversion and intelligent query decision. The WIAS uses SQL software IC concept to build universal SQL access templates, transmitting information parameters, like hardware IC provides the network service function for various Cloud information among the aforesaid subsystems, to access Cloud system information rapidly via Internet. The Content Databases include system Ontological Databases constructed by system processed index information and recommended user related databases supporting information sharing. When the Dr. What-Info II is in operation, the system calculates the most frequent and the least frequent query information according to time series analysis technique in answer to the frequency of query information periodically. The OntoCBRA produces related case information, and uses Full Scan with PHP algorithm [12] to trigger OntoDMA to correct the corresponding prediction rules. With the support of UAI technology, according to local GPS location capture and address conversion, with the support of OpenData@Taiwan/Taipei and domain ontology, the suitable corresponding information is captured to upgrade the quality of mobile information consultation and sharing effectively, so as to enhance the correctness, authenticity and integrity of location-based information provision. Moreover, if the aforesaid two agents cannot provide appropriate Cloud information solution, Dr. What-Info II will trigger OntoIAS to look for appropriate Cloud information solution outside internet by information search, information acquisition, information classification and information presentation or ordering, and the domain experts increase and modify preset rules, constructing the learning cycle in answer to query information thoroughly.

Fig. 4. Architecture of the proposed system

III. SYSTEM EFFECTIVENESS EVALUATION

A. System Display

The UAI is developed based on Taiwan government open data with local GPS location capture (Fig. 5) and the support of OpenData@Taiwan/Taipei and domain ontology. The suitable information is intercepted, such as distribution of good

restaurants and corresponding parking lots in journey based on local GPS location at St. John’s University, Taiwan, as shown in Figs. 6 and 7, respectively. Each bottom in the lower part of the aforesaid figures is clicked and its corresponding website is shown as Figs. 8 and 9, for understanding all of corresponding restaurant and parking lot official information in advance, respectively. Finally, if the user decides to go to the target, for

example, No. II Warehouse Café (⼆號倉庫咖啡), and the system can autonomously do the target route planning and display, as shown in Fig. 10; at the same time, the user can share the special and good recommendation by the Face Book, as shown in Fig. 11. The aforesaid block information recommendation based on the special event shall be the major contribution of Dr. What-Info II extended research.

Fig. 5. Local GPS location

Fig. 6. Good restaurant recommendations

Fig. 7. Parking lots information

Fig. 8. Restaurant website of No. II Warehouse Café

Fig. 9. Parking lot website of Repulse Bay in New Taipei City, Taiwan (新北

市淺水灣停車場)

Fig. 10. Target route planning and display

Fig. 11. Sharing the website of No. II Warehouse Café by the FB

B. System Evaluation

The "usability" of content provided by system webpage and "ease of use" (easy to use) of system operation are analyzed. The former one analyzes the ISO 9241, this definition uses effectiveness, efficiency and satisfaction dimensions to evaluate the usability. The Whitney Quesenbery (http://www.wqusability.com/, 2009) recommends a 5E function for measuring usability. For the ease of use of system operation, Jakob Nielsen (http://www.useit.com/jakob/, 2005) suggested ten basic principles of user interface design, as tabulated in TABLE II. The presented system user interface receives a collective satisfaction score of 80% (((4/5)+(8/10))/2). This tremendous difference in satisfaction score is caused by superior engaging, including (1) ~ (8) of 10 quantities by Nielsen as a consequence of an intelligent user friendly interface design. However, it must be pointed out that there is still room for the improvement of error tolerant, (9), and (10) in both user interfaces.

The design concept of this human-machine interface (HMI) uses the "Design preference to Importance Ratio" (DIR, see Eq. (1)) proposed by Ha [13], and use "Balancing Index" (BI, see Eq. (2)) to define the perfection of interface setting. In other words, when the BI value is zero, meaning all HMI elements are balanced in this design. Its physical significance is: this

HMI design conforms to the concept of important design preference and meeting the user's operating requirement in system operation. Hence, the DIR’s value should approach unity for optimal balance.

n

i ik

ik

n

i ij

ij

ijk

IIDP

DP

DIR

1

1

(1)

n

n

i ijk

jk

DIRBI

1 10log (2)

where DIRijk represents the DIR value of the HMI interface element i in design attribute j and important attribute k; DPij represents the design preference of the HMI interface element i in design attribute j; Iik represents the design importance of the HMI interface element i in important attribute k; BIjk represents the BI in design attribute j and important attribute k; n represents the total number of HMI interface elements.

TABLE II

A COLLECTIVE SATISFACTION OF THE PROPOSED USER INTERFACE

  Standard Items The presented 

system 

UsabilityQuesenbery

5E 

Efficient  O 

Effective  O 

Engaging  O 

Error Tolerant  X 

Easy to Learn  O 

Easy to use Nielsen 

Visibility of system status  O 

Match  between  system  and  the  real world 

O 

User control and freedom  O 

Consistency and standards  O 

Error prevention  O 

Recognition rather than recall  O 

Flexibility and efficiency of use  O 

Aesthetic and minimalist design  O 

Help  users  recognize,  diagnose,  and recover from errors 

X 

Help and documentation  X 

Total comparisons  80% 

Legend: “O” means to have this function; while “X” means none.

TABLE III EVALUATION OF HMI ELEMENTS AND THEIR

CRRESPONDING IMPORTANCE HMI 

ElementsDescription 

Design Preference 

Informational Importance 

Remarks 

GMAP  Google Map  5  0.166667 Graphical display and 

points 

GSP  GSP Location  3  0.233333  Label in Text 

BOT Recommendation 

Bottoms 3  0.233333  Bottom in Text 

NAVI  Navigation  3  0.183333  Bottom in Text 

FB  Sharing by Face Book  3  0.183333  Label in Text 

A 5-scale evaluation scheme is utilized to evaluate each of design preference (DPij). In other words, the design preferences include very good, good, moderate, weak and very weak, and their corresponding values are 5, 4, 3, 2 and 1, respectively. TABLE III illustrates the evaluation of HMI elements in our design attributes and their corresponding informational importance, evaluated by using the analytic hierarchy process [13], while TABLE IV shows the evaluation results of DIR with BI=0.003165 (BI should approach zero for the best balance). The verification results show that the human-machine interface of our proposed system can not only meet the important design preference, but also provide approximately optimal balance.

TABLE IV EVALUATION RESULTS OF DIR WITH BI=0.003165

HMI Elements  DIR  Remarks 

GMAP  1.266055  A little bit need to improve 

GSP  0.904325  A little bit need to improve 

BOT  0.904325  A little bit need to improve 

NAVI  1.000834  Approximately optimal balance 

FB  1.000834  Approximately optimal balance 

IV. CONCLUSION AND DISCUSSION

The Cloud intelligent Multi-agent Systems of information processing, exchange, communication, operation, integration, analysis and decision support is built by OntoDMA, OntoCBRA and OntoIAS three-stage intelligent decision, the technological research and development of integrating this system Dr. What-Info II with Taiwan government open data and UAI technology to look for and share optimal Cloud (mobile) information solution are implemented easily. It not only upgrades the quality of mobile information consultation and sharing effectively, but also enhances the correctness, authenticity and integrity of location-based information provision, so as to build the accurate, rapid, robust, pervasive and active intelligent mobile information consultation and sharing Multi-agent Systems. The system effectiveness analysis and validation results show the feasibility of the techniques and corresponding architecture of this system.

ACKNOWLEDGMENT

The authors would like to thank Tung-Han Tsai and Ji-Ling Gu for their assistance in system implementation and preliminary experiments. This research is sponsored under grant 106-2221-E-129-008 and 106-2632-E-129-002 by the Ministry of Science and Technology, Taiwan. The authors feel deeply indebted to the Department of Information and Communication, St. John’s University, Taiwan, for all aspects of assistance provided.

REFERENCES [1] S.Y. Yang and C.L. Hsu, "Dr.What-Info: An Intelligent Multi-agent

System Equipped with a User Friendly Interface Using the LBS and

Google Maps-based Techniques for Tour Guiding," Proc. 2015 International Conference on Applied System Innovation, Osaka, Japan, Paper-ID: 861, 2015.

[2] S.Y. Yang, "A Novel Cloud Information Agent System with Web Service Techniques: Example of an Energy-saving Multi-agent system," Expert Systems with Applications, 40(5), 2013, pp. 1758-1785.

[3] S.Y. Yang, "OntoIAS: an Ontology-Supported Information Agent Shell for Ubiquitous Services," Expert Systems with Applications, 38(6), 2011, pp. 7803-7816.

[4] S.Y. Yang, "Developing A Cloud Energy-saving and Data-Mining Information Agent with Web Service and Ontology Techniques," Submitted to Journal of Internet Technology, No. 826, 2014.

[5] S.Y. Yang, "Developing an Energy-saving and Case-Based Reasoning Information Agent with Web Service and Ontology Techniques," Expert Systems with Applications, 40(9), 2013, pp. 3351-3369.

[6] O. Curé, “On the design of a self-medication web application built on linked open data,” Web Semantics: Science, Services and Agents on the World Wide Web, 24, 2014, pp. 27-32.

[7] P. Fafalios and P. Papadakos, “Theophrastus: On demand and real-time automatic annotation and exploration of (web) documents using open linked data,” Web Semantics: Science, Services and Agents on the World Wide Web, 29(2014), 2014, pp. 31-38.

[8] A. Zuiderwijk and M. Janssen, “Open data policies, their implementation and impact: A framework for comparison,” Government Information Quarterly, 31(1), 2014, pp. 17-29.

[9] S. Tsai, “ASUS Cloud employs ASUS Cloud Platform to accelerate innovation and promote the interconnection of all things for vibrating new economic system of APIs,” White Paper, ASUS Cloud Solution Day, Taipei, Taiwan, 2014.

[10] S.Y. Yang, "Developing of an Ontological Interface Agent with Template-based Linguistic Processing Technique for FAQ Services," Expert Systems with Applications, 36(2), 2009, pp. 4049-4060.

[11] S.Y. Yang, D.L. Lee, and W.H. Chang, “Dr. What-Info II: Developing an Intelligent Multi-agent System with Open Government Data in Taiwan and Universal Application Interface Techniques for Mobile Information Consultation and Sharing,” Proc. of 2015 Cross-Straits Conference on Economics and Management, Yanbian, Chian, 2015.

[12] S.Y. Yang, "How Does Ontology Help Information Management Processing," WSEAS Transactions on Computers, 5(9), 2006, pp. 1843-1850.

[13] J.S. Ha, “A Human-machine Interface Evaluation Method Based on Balancing Principles,” Procedia Engineering, 69, 2014, pp. 13-19.