Optimal scheduling system with multiple status selection rules

In the status selection planning system, which is one of planning expert systems proposed by the authors' research group, the most promising status from tentative statuses that are generated by applying the dispatching rule is selected by a status selection rule. As the dispatching rules and the status selection knowledge are independent, it is easy to change the knowledge-base. Though quality of the solution depends on the status selection rule, it is difficult to acquire an excellent status selection rule. This paper presents an extended method of the status selection that can deal with multiple status selection rules as status selection knowledge. Since each status selection rule is simple and corresponds to a measure for evaluation, the status selection knowledge becomes simple. From the result of the application to a simple job shop problem, the authors confirm that this method is effective.<<ETX>>