Soft Computing Approach to Discrete Transport System Management

Abstract

AbstractA serious problem in the transport system analysis is to find a suitable methodology for modelling the management presented in real systems. It is hard to create an ”intelligent” algorithm of dispatcher - an algorithm giving significantly better results from pure random algorithms. In the paper we propose a multilayer perceptron approach to solve this problem. The neural network is learned using a genetic algorithm. A fitness function is defined by business service requirements of discrete transport system (DTS). The proposed approach is based on modelling and simulating of the system behaviour. Monte Carlo simulation is a tool for DTS performance metric calculation. No restriction on the system structure and on a kind of distribution is the main advantage of the method. The system is described by the formal model, which includes reliability and functional parameters of DTS. The proposed, novelty approach can serve for practical solving of essential management problems related to an organization of transport systems.KeywordsTransport SystemCentral NodeSoft ComputingBusiness ServiceDestination AddressThese keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.