ジョブショップスケジューリングにおける多目的評価のペタル図による視覚化

概要

論文の詳細を見る
This paper proposes a visualization method for multi-purpose job-shop scheduling. Recently, the petal diagram has been reported to visualize a multi-purpose function. However, it is difficult to make use of it in a real problem, because the petal diagram requests a normalizing procedure for an individual function, of which a multi-purpose function consists. Genetic algorithm (GA) is a useful method to solve a multi-purpose job-shop scheduling problem, in practice. One of GA's features is that it approximately searches a wide range of feasible solutions. In other words, it iteratively generates not only better solutions but also worse ones. This feature is adopted to normalize the individual function approximately. In the normalizing procedure, an obtained solution is mapped into [0, 1], where 0 and 1 respectively correspond to the worst and best solution searched in past. Some numerical experiments verify that visualization of multi-purpose job-shop scheduling solutions is intuitively achieved.
旭川工業高等専門学校の論文
2005-03-05