| The growing number of containers processed at marine centers and the increasing traffic congestion in metropolitan areas adjacent to marine ports necessitates the investigation of more efficient and reliable ways to handle the increasing container traffic. In this dissertation, it is shown that the problem of container movement by trucks in metropolitan networks can be modeled as an asymmetric “multi-Traveling Salesmen Problems with Time Windows” (m-TSPTW). The m-TSPTW entails the determination of routes for the fleet of trucks so that the total distribution costs are minimized while various requirements (constraints) are met.; In the first part of this dissertation, the deterministic m-TSPTW is investigated. A two-phase exact algorithm based on dynamic programming (DP) is proposed that finds the best routes for a fleet of trucks. Since the m-TSPTW problem is NP-hard, the computational time for optimally solving large size problems becomes prohibitive. For the case of medium to large size problems, two computationally feasible methods are developed: (1) a hybrid methodology consisting of DP in conjunction with genetic algorithms, and (2) a heuristic insertion algorithm.; It should be noted that in the real world operations in any traffic network contain a fairly high degree of uncertainties. The second part of this research is devoted to stochastic TSPTW in which the travel times along the arcs and the service times at the nodes of the traffic network are non-stationary stochastic processes. Means to estimate the arrival times at nodes of the network are investigated. A methodology is developed to address the existing non-linearity formed by hard time windows at nodes, and an approximate solution method is proposed to solve the stochastic TSPTW.; Simulation results are used to demonstrate the efficiency of the proposed methods. |
