A New Approach to Transportation Equilibrium Assignment Problem by Neural Network Model

Abstract

This paper considers the method for calculating the transportation equilibrium assignment problem on the “neural network”. The method proposed in this paper represents link flows by many “neurons”, which have parallel dynamics. First, the equations of neuron dynamics are derived for deterministic multi-commodity equilibrium assignment problem from the correspondence between Lagrangian of the problem and Lyapunov function of neural network. Next, the method are implemented in a moderately large size network and the numerical properties are investigated. Furthermore, relation between the method and the stochastic equilibrium theory are considered.