In: Lecture Notes in Computer Science : Artificial Intelligence and Soft Computing - ICAISC 2006, Volume 4029, 2006, pages 1040-1049. 2006. URL: http://dx.doi.org/10.1007/11785231109.
Abstract: The following analytical approaches: queuing network models, stochastic timed Petri nets, stochastic process algebra, Markov chains are used in performance eevaluation of multi-agent systems. In this paper, new approach which is based on PERT networks is proposed. This approach is applied in performance evaluation of layered multi-agent system. These layers are associated with the following types of agents: manager, bidder, and searcher ones. Our method is based on approximation using Erlang distribution. Accuracy of our approximation method is verified using simulation experiments.