Development of Logistical Model Based on Integration of Ontology, Multi-Agent approach and Simulation

摘要

The paper focuses on development of logistical model based on multi-agent model of a resource conversion process. The model is used for planning and dispatching of gas stations network. We compare existing approaches to multi-agent planning, including requirements and capabilities nets, multi-agent resource conversion process model, active and passive convertors model, agent-based simulation modeling, and their software implementations (Magenta, BPsim, Onto Modeler, AnyLogic). The most perspective approaches are implemented in Magenta and BPsim systems. However multi-agent approach based on distributed calculations has one disadvantage, which is frequent plan modification, resulting in certain instability of the model. To avoid this effect, we use situation filtration and diagnosis block. In order to take delays into consideration and analyze the bottlenecks in logistical chain, we use simulation modeling. We compare existing knowledge representation models, identify the benefits, and develop a methodology that considers all advantages.