Postoptimizing equilibrium flows on large scale networks

In a large variety of applications, equilibrium traffic flows corresponding to a set of slightly modified input data must be computed sequentially. Until now, it is believed that only a disaggregate decomposition approach, that works explicitly on the path flow space, offers the postoptimization capability. This note proposes a new postoptimization method to deal with perturbations of the traffic demand input that does not require path information. Numerical experiments on practical size networks show a drastic reduction in the number of iterations with respect to the naive restart approach.