Knowledge-based scheduling under unpredicted conditions: Two approaches

Control of a flexible manufacturing system (FMS) is a complicated task due to the wealth of feasible alternatives. Such control is done hierarchically: in the cell level, part scheduling is considered; in lower levels the equipment control strategy is looked at. In building a knowledge-based control system the conditions and variables which affect the control mechanism have to be defined and represented in the knowledge base. Since the control mechanism defines conditins and actions in the environment, it is possible to reach conditions for which the action is not sufficiently defined, is not appropriate to control the system, or missing altogether. This paper presents two methods for knowledge-based scheduling, one for each level, and evaluates their ability to function under varying conditions that deviate from the expected ones for which the system was designed.