Abstract
In recent years, Japan has been faced with the serious problems in the relief distribution in the aftermaths of large-scale disasters. The affected areas experienced supply shortage due to surge in demand. Therefore, it is necessary to make the relief distribution plan in advance. This helps appropriate allocation of relief supplies to each refuge and facilitate rapid delivery. This paper presents a model to decide the quantity of relief supplies and visiting order of refuges in a state of excess-demand. The model features two objectives: minimizing the total travel distance and the penalty for the shortage of the relief supply. In this model, the non-dominated sorting genetic algorithm is used for finding the pareto optimal front.
First, this algorithm is applied to a test problem for the verification. All feasible solutions of the penalty are enumerated on the problem where the total travel distance has been minimized. The performance of this solution method is confirmed by comparing the computational result with the enumerated solutions. Finally, we applied this model to a realistic problem in the case of locations of refuges and total demand. In the result of this case study, many non-dominated solutions are obtained. These results show the importance and necessity of planning for relief supplies. This model can propose valid options for relief supplies and contribute to distribution for relief supplies in the aftermath of a disaster.
