Abstract
The status selection planning system, which has been proposed by our group, is one of planning expert systems. The system uses two kinds of knowledge to solve a problem, that is, dispatching rules and a set of status selection rules. Dispatching rules mean fragmentary and convenient assignment algorithms. In this system, the most promising status from tentative statuses generated by applying the dispatching rules is selected by a set of status selection rules.
As the dispatching rules and the status selection knowledge are independent each other, it is easy to change the knowledge. Although the quality of the solution depends on the knowledge-base, it is usually difficult to acquire useful knowledge from human experts.
In this paper, we present a method of learning the status selection knowledge for flow shop problems. Using C 4.5 algorithm, which has its origin in ID 3, the status selection knowledge is created and formed into a decision tree. Training data are made from the path to the optimal solution. From the result of the simulation of the proposed method, we have confirmed that the learning method is so effective.
