An algorithm for the stochastic user equilibrium problem

In this paper we present a new algorithm for the approximate solution of the logit-based stochastic user equilibrium problem. The main advantage of this algorithm is that it provides route flows explicitly, of particular interest in the evaluation of route guidance and information systems; in previously proposed methods for the stochastic user equilibrium problem, only link flows are provided. The proposed algorithm alternates between two main phases. In the subproblem phase, profitable routes are generated. In the restricted master problem phase, a descent method is used to solve the restriction to the original problem to the subset of the total set of routes generated so far. We present and evaluate alternative strategies for generating routes algorithmically, and discuss the possibility of utilizing such strategies for reducing the inherent problem of overlapping route flows in logit-based traffic models.