Simulations of transportation logistic systems utilising agent-based architecture

Abstract

This paper presents a methodology related to rapid and flexible prototyping of simulation models reflecting large-scale transportation logistic systems. This methodology is based on a proprietary architecture (called ABAsim) inspired by the paradigm of autonomous agents. The architecture was designed within the framework of research and development of an integrated computer environment, which is specialized for simulations of a wide class of transportation logistic systems. Basic characteristics of the mentioned architecture are briefly described, whereas a flexibility of simulation models, developed within that architecture, is emphasised. The high degree of flexibility is enabled due to consistent hierarchical structure of relevant conceptual models and indirect addressing mechanism utilised for inter-agent communication. Because of the applied approaches, the simulation model configuration can be built by replacing relevant sub-models, individual agents or their internal components. In addition, exploitation of Petri nets for definitions of control procedures and utilisation of a proprietary CASE-tool also supports rapid and flexible model prototyping. There is also presented a complex simulation tool Villon (developed within ABAsim architecture) and corresponding experience with its utilisation within simulations mirroring compound systems from the field of transportation logistics systems.