Introduction of a Weighting Factor Setting Technique for Multi-Objective Optimization
Hirokazu Kobayashi, Ryosuke Tachi, Hisashi Tamaki
In production planning and logistics planning, there are multiple purposes which are to reduce cost of procurement and to shorten transportation time, and so on. These are sometimes contrary to each other and have different units. To adjust the balance of these purposes is important to the satisfaction of operators for the plan. Although various methods have been proposed in these multi-objective optimizations, there exists a problem it is impossible to derive the satisfactory result for operators, because the reason for the difficulty of judging for each purpose. Therefore, in this paper, we consider the achievement as operators’ preferred solution and aim to obtain a satisfactory plan for operators by proposing a method to estimate the importance of each objective to obtain this preferred solution. In the proposed method, the weighting coefficient of the problem is estimated as an inverse problem. In the cases the preferred solution is a Pareto optimal solution and a base solution, or the preferred solution is not a Pareto optimal solution or a basis solution, but a Pareto optimal constraint is given, it was confirmed that we can estimate the weighting coefficient to make the optimal solution is consistent with the preferred solution.