Distributed Multi-Agent System Development for Regional Business Processes Information Support Текст научной статьи по специальности «Экономика и бизнес»

Информатизация

Masloboev A.V., Shishaev M.G.

DEVELOPMENT OF DISTRIBUTED MULTI-AGENT SYSTEM FOR INFORMATION SUPPORT OF REGIONAL BUSINESS PROCESSES

Institute for Informatics and Mathematical Modeling Kola Science Center RAS,

Murmansk Region, Apatity

This paper presents development experience of architecture, principles of functioning and software engineering and modeling of distributed multi-agent system (DMASICC) which provides an interactive innovation chains construction for information support of regional scale innovation infrastructure. Special attention is given to methods and technologies for developing, modeling and constructing systems of the given type which provides control of innovation and business processes. We also describe how to design such adaptive multi-agent systems using an innovation process multi-agent model, which defines the entities and relationships of a typical business process. By designing a system using the multi-agent model, the system can determine the best mapping of agents to roles, based on their current capabilities, for the current system goals.

1. Introduction

In the new millennium just countries susceptible to innovations will stand up to competitive struggle on the global wares and services market. Therefore, transition to innovation economics is very relevant for Russia and other countries which have a great tendency in the field of raw materials. A great support could be provided here by Russian Academy of Sciences which has already accumulated innovation activity immense experience in planned economy frameworks. During these years promotional chains “science-industry” formed on basis of administrative and research organizational arrangements which bound academic science with sectoral, ministerial design offices and plants in stable budget and self-supporting financing frameworks. Transition to market economy entailed to organizational bindings in science-industry chains disruption, and now they form in market element [1].

Thus, for today it is conventional and quite obvious fact that innovations are critical element of modern business system development which provides their competitiveness. Business process participants are geographically distributed. Modern economic and scientific relations not only overcome wide distances, but very often “remove” any intergovernmental frontier. Growing up day after day number of innovation activity fellows and their geographic distribution condition on the fact that just lesser part of potentially efficient innovation projects are practically realized. Innovation markets peculiarities put forward new requirements for management problem-solving quality and efficiency which influence on socioeconomic systems development process. These claims settlement is not possible without adequate information support of business processes within the systems. We can propose many other arguments to validity of proposals listed above.

Innovation structures information support workable system has to provide

wide and active expansion of information and common features (scientific and/or industrial potential) and all of innovation activity participants needs in the region, in order to create frameworks for accumulated knowledge materialization in the form of new technologies and products.

Thereupon, we see the urgency of innovations information support systems development which provide, specifically, potential business partners automation seeking and innovation chains “researcher - technologist - investor - producer -customer” formation.

In this paper we describe innovation chains formation mechanism and their formation rules. Descriptions of architecture, functioning principles and programming realization experience of distributed multi-agent system for innovation chains construction are given

2. Business Processes Information Support Problem

Innovations information support is a difficult and very dynamic multivariate task. In such frameworks it is not possible to rely on full management problemsolving process automation. However, this process can be substantially supported by availability of adequate to the current task information collections, formalistic methods and innovation processes simulation modeling computer systems, which in complex constitute innovation activity information support system [2, 3].

Complexity and dynamics such type of tasks cause to necessity of large information volumes for securing of which we need methods, technologies and tools of existing geographically distributed information resources maximum integration. Task dynamics advances specific requirements to functional features of these technologies - realization of rapid and minimum cost-is-no-object reconfiguration and providing information systems adaptation created with their help. Thus, we can formulate following requirements to innovations information support system features spectrum:

• Flexible recustomizable integration of distributed information resources.

• Unitized interface for heterogeneous data sources securing

• Seeking for data sources on basis of their content

• Appropriate access to infobases

• Availability of socio-economic systems reaction modeling methods and tools on external and internal changes.

In our opinion for today to the utmost must be paid great attention to following technologies which provide the requirements listed above. There are heterogeneous data-processing resources integration technologies based on Grid systems (Grid technologies) and multi-agent technologies which provide distributed data-processing tasks solving.

These technologies provide unitized access to geographically distributed heterogeneous data sources composing necessary infobase for current situation analysis for the purpose of management problem-solving and formation.

3. Innovation Chains Formation Mechanism Description

We propose well known potential innovation chains formation approach [1] based on innovation projects, products and resources identification (ontology) using unified terminology database which is realized as ontology of innovation activity.

Innovation chains variants, which potentially provide the realization of separately taking innovation project, form using terminologically derived unified description

Innovation chains formation occurs properly by searching innovations fellows which dispose of necessary resources for innovation product development.

Terminology database incorporates following concept groups: innovators (enterprises, scientific organizations, etc.), investors, innovation process phases (basic research, research, development, marketing, manufacture and other), innovations objects and subjects. All these concepts are connected with such relations as “innovation process phase result”, “phase sequence”, “phase initial values”, ”phase participant”, etc.

Innovation chain construction is carried out by automated produced scenario which structure is defined by innovation chain structure and innovation request form.

Innovation request form depends on user’s type (researcher, producer, investor, etc.) and can consist, for example, in advancing of any scientific idea or producing a new device or material organization.

4. System Architecture

In contrast to existing innovations information support systems developed system has a distributed hybrid architecture represented on Figure 1. System components are central node which plays the role of main innovation supplies data storage and program agents which represent ultimate users interests in the system and implement on client hosts.

Fig. 1. Distributed multi-agent system for innovations support hybrid architecture

• Innovator - innovation process participant’s computer which use software-based intelligent agents for seeking potential business partners and innovation chains formation

• Internet - group of devices (bank) which give access to the Internet or local area network (LAN);

• Agent Environment Server - mobile agent server, innovation and business supplies storage (central node);

• Lockal Network - local network equipment which provide connection with server

• DataBase Server - database server (for example MS SQL Server 2000, Oracle9i, DB/2 Sybase) which contains information about innovation projects, their participants, etc.;

• File Server - file server which store all necessary information and additional materials.

For designing business process participants or their agents negotiation model in our developed multi-agent system we constructed a business processes multiagent model and realized it in the form of virtual round-table discussion which can be implement both in the local network and in the Internet.

In consideration of the fact that agents like their owners are geographically distributed and just possessed of a short part of information needed to efficient solving of the assigned goal in our work. System state of distribution gives the main advantage which consists in the possibility of geographically distributed heterogeneous information resources integration in a single whole and their use in a single whole information environment.

Agent allocation on the client side (innovation processes participants hosts) improves the interactivity level of communications both between user and agent, and user and system in general. In one's turn decentralized hybrid type of system architecture allows to organize more efficient search of potential business partners and their business supplies information.

Thus, designed system architecture provides asynchronous agent communication and interaction character that gives the possibility of their work in the frameworks of heterogeneous and unreliable communications which are extremely relevant regional scale systems. Active nature of developed program agents position multi-agent systems as an efficient feature for distributed data processing (DDP). Agent autonomy and their cooperation ability allowed us to develop flexible and easily reconfigurable distributed system.

5. System Functioning Principles

New information technologies dynamic development concerned with global network Internet provides innovation activity information support distributed multi-agent systems creation on basis of agent technologies and Semantic Web Project concepts and principles. The Semantic Web Project [6] main purpose consists in information systems for global network development for autonomous program agent work and interaction that allows to automatize a lot of task solving

with use of Internet, for example in following spheres: e-business, medicine, financial operations, electronic libraries, distance education, etc. [7].

System functioning principles showed on Figure 2 reproduce in many respects ideas laid under Semantic Web concept [8]. But in contrast to this Semantic Web concept where initially it is implied that seeking, registration, update and data processing processes are initiated by user in our developed system such processes initiators, analogous to the same in Semantic Web concept, are agents which are directly representatives of their owner within the system. Thus, innovation subject has just to register his own node (host) in the distributed virtual business environment, configure his agents required options and wait for results of agent activity. All the work related with seeking for business partners and preliminary innovation chains potential efficiency analysis is in the agent competence. During the system functioning agents communicate with ultimate users representing results of their own activity for consideration to them and requesting for qualifying information about innovation supplies declared user in case of structure or innovation field attributes (agent functioning scene) changing or business supplies information insufficiency. At the same time user can choose how it will be carried out: in interactive mode in way of question-answer or in automatic mode.

Real World

Agent World

hformaicri represertation Environment

Pa itid pa rt o r hi s in nowti on sup pi y hto nrat » n s eek r g abo ut pot emia I f'

Registration lj)daeintorntfiononthepgtaI p at rers

Recieve message tan other seidmeuage tfeqi«t)tootter agerts age it

Prepaation res lits of seeking Chans formaicn /^gent Str^eges and Incticns .Agent request or^neaiion to

confguation olher chan inks

Real World

о

Results nepreseritaicn

Lhain links and chans selection

Business process Paticipant

Fig. 2. System functioning scheme

Developed software screenshots are represented on Figure 3.

Agent Manager: HelloWarld

Е* а^!йп* ї

щ^ЭЁЕЕЕЭ Е32302здз

УДдвпІ ЛІГЙГЙІГІ'ВЙЙТЙ'

pRfHriH

FAG* РАС ko tare

ачгАїглйло»

Асі-г-лп

GotmttraniB

Fkiln

ІЬ—j^jtnri

rt*s >jnrti*»aa4^kBlpaofef інгігріівні азекоириі

І іію шві ■ КвсС'ттл-йріЬггіНі'Г пГ'Гпти ГІ о IЕІ Гі bum the

Чіеогін:

-UlbWnikd

~’гк±:ъ+ъ'

iuch T-mji LTt4-:tr,'

WsutOntofo®

bXSfcrt

S?ATVy*.T\K №» i7;d?:MP5T \e*$

Iferdor R5J

5t-*to rafaLs Dl-?• L■“■■t

-J

Рґ-.-Іиїві Consumer tall liHiiiiA^utliLd

Consumer Producer Г Mister tom netrnijr

Con sum и pTOdliOii tell String Do мої mi int.

Producer Consumer tell ItemBMi.iesI

Cuii£uiiiCi Produce ailiiwfii IDainAMiuiicLii

C«isum« tao>duc4f uWVjiir.y,- ют.паЕкиЬг...

Consumer Producer ash—H ІЬетЯмиеі-і

РАС Editor: HclloMJorld

№ Ш !¥>n*nre

P*0 S’TwHrt

Narre №Ро№ОТ»т

l№

Full Ntarn і

n Mlioitai ■

Егаа halki^i'DiU Halbrt'niklFmT

Deoofrnn: *

it bnr'i't« L.r: - ir-^ к -

>.lbf«-d№chl Cluid-. Tnir Lhtobqy

Uvli-ci! I'Aui

\|ЄТі

■V-irJkiia

РадОолптЯДОт

himCaiTiT^ib

ТіГЛе-

EJta.....£dlt К|імкнк

:'ші^'йіі'!аііа|нц'*-н«ггі дне | Conditions

^peratcr=> * Insisnwe...

Bui о Editor t Frinb БрнвЫпе

Utlp

Defined var йЫе...

! <Vdui

Г-Jew

Utsnga Cmc :im:

i-ihKorno^Hwit^^nder F.T/WL5 -HeloW^WPACT* I ?hcb.rnngMJ? fWIDIГПr<jV? 6XIіLS )

ncnm Г&Л«мер. co.nl e п I Гу p e EQUALS iSrng |

7ІУ12ІЇІ trji'00 лI•?ПI ЁQUALS ‘Se/HMf )

HtenW Oondfaons

Up

Up

Cw

fluent Engine Optlprn

J A.

*|й DU h к:

fi&itii SriiwK.

CkiiaUilh ■uritaaLji^ii'd' ■ ■ іф-лшЧ=-зІгр-ліІіЬл1 (■г.^лп-^тсА.ЦргіЙикйпіЬ

UtflJHdSpifem РпрМіОфгі EWUfl

Del 510

Fig. 3. System Graphic User Interface

6. Conclusion

The problem of innovation activity information support has a multidimensional character. High dynamics of innovation-oriented regional economics from the one side and inertness of socio-economic system in reaction on innovation management from the other side, improve the role and significance of innovations information support task. For specific claims settlement advanced to corresponding information support systems it is necessary to use modern progressive methods and technologies of distributed information systems development, such as Grid, Peer-to-Peer (P2P), Multi-agent systems (MAS) technologies and some methods of system dynamics.

In this paper we proposed architecture and functioning principles of distributed multi-agent system for innovation chains construction, their interactive formation mechanism in the innovation activity information support network regional system, which is accessible for users in the form of specialized business-portal.

This mechanism is based on applying the agent-oriented approach to nonlinear model of innovation process which we didn’t describe in this paper. For its realization is used knowledge and data integration subsystem incoming into business-portal with help of which integration of geographically distributed heterogeneous information resources related with innovations and semantic search within them are provided.

System is developed on basis of multi-agent system technology that allows to carry out active seeking of potential efficient innovation chains which consist of geographically distributed elements.

System functions on basis of interaction scenarios of business process participants and is oriented to use in large computer networks with wide service diversity.

7. Acknowledgements

This work is sponsored by the Institute of Informatics and Mathematical Modelling Kola Science Center Russian Academy of Sciences (IIMM KSC RAS) and Russian Foundation for Basic Research (RFBR) under grant number № 0507-90050.

References

1. Zagorulko Y.A.., Bulgakov S.V. Use of ontology for innovation chains construction in innovation activity support system in the region // Management and Simulation problems in complex systems: Proceedings of VI International conference, Samara, 14-17 June 2004 - Samara: Samara Science Center RAS, 2004. - P. 328-333.

2. Tumanov I.V. Innovation services market as an element of market infrastructure // Scientific materials of SevKavGTU. Series «Economics», version.6. - Stavropol: SevKavGTU, 2002.- P. 52-63

3. Peter M. Milling. Modeling innovation processes for decision support and management simulation // System Dynamics Review, Vol. 12 1996. - P. 214-234.

4. I. Foster, C Kesselman, J. Nick, S. Tuecke Grid Services Architecture for Distributed Systems Integration // Extensible Systems. 2003. №1.-P. 20-26.

5. The Physiology of the Grid: An Open Grid Services Architecture for Distributed Systems Integration. I. Foster, C Kesselman, J. Nick, S. Tuecke, Open Grid Service Infrastructure WG, Global Grid Forum, June 22, 2002. http://www.giobus.org/research/papers/ogsa.pdf

6. Terziyan V., Kononenko O.: Semantic Web Enabled Web Services: State-of-Art and Industrial Challenges. In: M. Jeckle and L.-J. Zhang (eds.), Web Services - ICWS-Europe 2003, Lecture Notes in Computer Science, Vol. 2853, Springer-Verlag, 2003. - pp. 183-197.

7. McIlraith S.T. Cao Son and H. Zeng, «Semantic Web Services», IEEE Intelligent Systems, March/April, 2001. - pp. 46-53.

8. Hendler J. Agents and the Semantic Web // IEEE Intelligent Systems, Vol. 16, №2, March/April 2001. P. 130-151.