Metaheuristics for multiobjective capacitated location allocation on logistics networks

Entitled: Metaheuristics for multiobjective capacitated location allocation on logistics networks and submitted in partial fulfillment of the requirements for the degree of complies with the regulations of the University and meets the accepted standards with respect to originality and quality. ABSTRACT Metaheuristics for multiobjective capacitated location allocation on logistics networks Yonglin Ren Logistics is vital to sustaining many industrial, commercial, and administrative activities. It is often composed of the logistics service providers and the customers being serviced. The goal of service providers is to maximize revenues by servicing customers efficiently within their preferred timelines. To achieve this goal, they are often involved in activities of location-allocation planning, that is, which logistics facilities be opened, where they should be opened, and how customer allocations should be performed to ensure timely service to customers at least delivery costs to logistics operators. In this thesis, we address the multiobjective capacitated location allocation problem on logistics networks. The distinction between the location allocation problem treated in this thesis and the traditional location allocation problem lies in its multiobjective and dynamic nature. The multiple objectives considered are travel time, travel distance, travel cost etc. and developed based on practical constraints such as presence of congestion, timing and access restrictions imposed by municipal administrations in urban areas etc. The dynamic aspect means the location allocation results are not fixed forever but vary iv with change in municipal access or timing regulations, congestion, or land, material and labor costs on logistics networks. Four metaheuristics namely Genetic algorithms (GA), Simulated annealing (SA), Tabu search (TS), and Ant colony optimization (ACO) based solution approaches are presented to treat the multiobjective facility location allocation problem. Two cases are studied. In the first case, opening costs of the facilities and only one criterion (distance) is used. In the second case, opening costs of the facilities and multiple criteria (distance, travel cost, travel time) are used. The proposed approaches are tested under various problem instances to verify and validate the model results.