Organ transplantation is an important treatment method. Despite recent advances in medicine and technology in this field and its great importance in health, in most countries, thousands of patients are still waiting to receive an organ, and many of them die before they can receive an organ. The main problem in organ transplantation is the imbalance between supply and demand, and the demand for a transplanted organ is always greater than its supply. Due to the severe shortage of organs against its demand, it seems vital to pay attention to donated organs and use them efficiently. The organ transplant network includes organ transplant centers, organ procurement centers, hospitals, donors, and volunteer patients for receiving transplants. By determining the appropriate location of each of these facilities, moving between them in the shortest time and optimal allocation can diminish the number of dying patients. In this thesis, a bi-objective mathematical model for the transplant network and location-allocation is presented. Alongside minimizing the total costs of establishing facilities and flow costs between centers, the model maximizes the flow between centers. Also, in the proposed model, the probability of failure of each of the transplant centers is considered. Therefore, a support transplant center is determined for each transplant center according to the objective function. Due to the Np-hard nature of the problem, a heuristic algorithm based on multi-objective simulated annealing is designed to solve large-scale instances. Finally, the results of solving the mathematical model and designed heuristic algorithm are reported in small, medium, and large dimensions instances, and several sensitivity analyzes have been performed on important parameters. The numerical experiments show that the ε-constraint method is not able to solve large instances in a logical time. Besides, a heuristic algorithm is used to solve large instances; The heuristic algorithm always performs better than the ε-constraint method in terms of computation time .According to other eva‎luation indices, it can be concluded that the ε-constraint method works better than the heuristic algorithm in solving small instances, but in medium and large instances the the heuristic algorithm performs relatively better than the ε-constraint method.

علي شاهنده نوك‌آبادي

مهدي ايرانپور

مرتضي راستي برزكي، مهدي مهنام