Konstantin A. Aksyonov, Ph. D., IEEE member,Eugene A. Bykov, IEEE member
Elena F. Smoliy Maksim V. Popov
Leonid G. Dorosinskiy
Ural State Technical UniversityEkaterinburg, Russian Federation
10th PhD Workshop on Systems and ControlHluboka nad Vltavou, Czech Republic, 25 September 2009
Development of decision support system «BPsim.MSN»
[email protected] www.bpsim.ru
Multi-service telecommunication networks design and modeling application
Field of Interest
System analysis of business processesSimulation modelingMulti-agent systemsDecision making systemsComputer-aided software engineeringIntegration of simulation and CASE tools
[email protected] www.bpsim.ru
Outline
Theoretical aspectsMulti-agent approachDecision support over simulationDeployment experienceDemosConclusion
[email protected] www.bpsim.ru
BPsim products development background
Popular modeling tools comparisonCommon drawbacksAgent-based approach implementationRetail price
[email protected] www.bpsim.ru
Model of Resource Conversion Process
Transformation of resources
Controlled by agents
Operation in/out- puts
MechanizmsControl
commands and messages
A1Sender 1 Receiver 1PR 1
Res1, Res2
Res6, Res7, Res8
U1, U2,U3, U4
P1, P2
Op2
Res1
Res2
Ca2 U1
Mech1 Mech2 Mech3
Res3
Res5
Message 1
Message 2
Mech5
Mech4
Res6
Res7
Res8
Op4
Ca4 U4
Op3
Ca3 U3
Junction 1
Ca_J1 U2
A2
www.bpsim.ru
Resource Conversion Process Features
Structure and behavior algorithms complexity
Large amount of parameters
Complex hierarchyProbabilistic behaviorDecision-making
processes and decision-making people definition capability
L=0 (maximum topological definiteness graph)
L=1
L=2
Process Operation Agent
www.bpsim.ru
Agent Features
Control over RCP objectsEnvironment analysisSituation diagnosis and decision
search (interaction with a database)
System goals determination, delegation and achievement control
Message exchange (communication, negotiation, etc.)
[email protected] www.bpsim.ru
Agent Properties
Individual knowledge baseSet of goals for behavior
configuration settingPriority that defines agent order
in priority gaining queue
[email protected] www.bpsim.ru
Tool Application
Urals Industrial Group
Model designDefining agent
settingsMonitoring goals
achievement progress
Experiments output
[email protected] www.bpsim.ru
«Launch agent» knowledge base
[email protected] www.bpsim.ru
Intelligent reactive agent model
Frame-based approach Decision search diagramsFrame-concept (FC)
Frame name (FN)
Frame type (FT)
Application info (AI)
...
...
...
Behaviour scenarios structure (BSS)
Scenario selection block
Scenario 1
Scenario 2………
Slots structure (SLS)
Slot 1
Slot 2……
[email protected] www.bpsim.ru
Multi-service network model
[email protected] www.bpsim.ru
Conclusions
Integration of Simulation and multi-agent approach Expert system Situational modeling
Result – intelligent decision support system, applied to multi-service network design
[email protected] www.bpsim.ru