UDC 656.2+004.852
CREATING OF AN ARTIFICIAL NEURAL NETWORK FOR TRAIN SCHEDULE’S ELABORATING
A. V. Ignatenkov, Samara State Railway University, Railway Automation and Telemechanic Department, post-graduate student; LLC “Scientific and technological center on railway implementation and management”, Deputy Head on IT ; This email address is being protected from spambots. You need JavaScript enabled to view it.,
A. M. Olshansky, PhD (technical sciences), Samara State Railway University, Railway Automation and Telemechanic Department, doctoral student; LLC “Scientific and technological center on railway implementation and management”, Deputy Head on science; This email address is being protected from spambots. You need JavaScript enabled to view it.
The article is devoted to the consideration of transport process schedule creating problem (including train schedule creation) by an artificial neural network approach. The goal of investigation is to develop network architecture and topology and to construct major principles of neural network functioning to create train schedule on two-track railway sections. Within the framework of this goal traditional methods and techniques are analyzed and interneuron connections are offered. During the computation of neural network output the authors use soft competitive principle to activate connections between layers of the network. A specific adapted learning algorithm with variable learning rate is described. Also the authors present the results of numerical experiments with neural network.
Key words: train schedule, artificial neural network, variable learning rate, neuron links, error function, transport systems.