CITY LOGISTICS: FREIGHT MANAGEMENT WITH TIME DEPENDENT TRAVEL TIMES AND APPLICATION TO LARGE-SCALE NETWORKS

摘要

In this paper, we address freight management in large, congested urban settings. Congestioncreates a substantial variation in travel time (speeds) during morning peak and evening peakperiods. Urban route designs that ignore these significant variations are often found to beinefficient or even infeasible within a congested traffic environment. This limitation may lead todelays, higher operational costs, and inferior customer service. To take account of the dynamicsof traffic conditions, the proposed modeling framework includes an underlying dynamic networktraffic simulation-assignment (DTA) model which is able to compute time-dependent shortestpaths (TDSP) and simulate individual vehicles on a more realistic representation of the urbanroad network elements, including signalization, ramp metering, and other operational controls.An intuitive and efficient heuristic solution algorithm is proposed, which combines the classicalk-opt exchanges to improve the solution and a procedure to reduce the number of vehicles out forservice. No parameter tuning is required and no random choice is made. The integrated modelingframework is capable of addressing TDVRP of practical sizes on a real world urban network,demonstrated here on the Chicago area network, with problem instances of up to 500 customerswith TDSP updated at 10-minute intervals.