Dynamic optimisation of urban intermodal freight transport with random transit times, flexible tasks and time windows

Abstract

An improvement on drayage operations is necessary for intermodal freight transport to become competitive. When drayage takes place in cities or urban centres transit times are usually random, as a consequence finding the optimal fleet schedule is very difficult, and this schedule can even change during the day. The work we present here is a dynamic optimisation model which uses real-time knowledge of the fleet's position, permanently enabling the planner to reallocate tasks as the problem conditions change. Stochastic trip times are considered, both in the completion of each task and between tasks. Tasks can also be flexible or well-defined. We describe the algorithm in detail for a test problem and then apply it to a set of random drayage problems of different size and characteristics, obtaining significant cost reductions with respect to initial estimates.