Apply Greedy Randomized Adaptive Search Procedures for Solving Capacitated Single Allocation p-Hub Median Problem with Multiple Capacity Levels

• Main Authors: Kuo-Rui Lu, 呂國瑞
• Other Authors: Ching-JungTing
• Format: Others
• Language: zh-TW
• Online Access: http://ndltd.ncl.edu.tw/handle/93833516384463660235

碩士 === 元智大學 === 工業工程與管理學系 === 100 === Hub and spoke network has been used in telecommunications and transportation systems. In such a network, the decisions on the location transshipment points, where the flows between origin-destination (O/D) pairs are consolidated, and the paths for sending the flows between the O/D pairs have to be determined. The capacitated single allocation p-hub median problem (CSApHMP) is to decide the locations of the p hubs with capacity restriction and the allocation of non-hub nodes. In this research, an extension of the classical capacitated single allocation p-hub median problem is studied, in which the size of the hubs is part of the decision making process, called capacitated single allocation p-hub median problem with multiple capacity levels (CSApHMPMC). A set of capacities is assumed to be available to select for each candidate hub. We formulate the problem as a 0-1 integer programming model. A Lagrangian relaxation (LR) approach and a greedy randomized adaptive search procedure (GRASP) algorithm are proposed to solve the CSApHMPMC). The Lagrangian function that we formulated decomposed the original problem into smaller subproblems that can be solved easier. Computational experiments with Australia Post (AP) data set (ranging from 10 to 50 nodes and five capacity levels) from literature are performed for both proposed algorithms. We also use Gurobi to solve the small size instance for comparison. For all tested instance, GRASP can obtain optimal solutions in most of the instances and obtain more optimal solutions than those by LR approach. In the future, larger size instances should be tested to confirm the efficiency and robustness of our proposed algorithms.