Soft and Parallel Computing for Multi-objective Combinatorial Optimization

Under recent socio-technical systems interaction, many problems are formulated as mixed-integer programs whose solution will refer to a class of NP-hard combinatorial problems. If concerned with multiple objectives additionally, obtaining an optimal solution becomes specially tedious and difficult in real world applications. To cope with the problem, this study calls attention on outstanding recent progress of soft computing methods like genetic algorithm, tabu search, simulated annealing, neural networks, and so on. Besides parallel computing implementation of the algorithm, a hybrid adaptation of those methods will be examined so that we can develop novel and practical solution methods depending on the situation.

Network optimization has been a core problem domain in operations research, computer science, applied mathematics, and many fields of engineering and management as well. In this study, we consider the location of hub facilities, which is an important issue arising in the design of communication networks, airline passengers flow and parcel delivery networks. A new heuristic method based on the hybrid tabu search is developed for the problem of locating p hub facilities in a network consisting of n clients and m servers. As closely related aspects, this study also considers the problem of designing minimum cost networks satisfying certain design criteria about network survivability or reliability.