Research On Disruption Management For Production Scheduling When Uncertain Disruption Event Occurred

In the actual production of enterprise, an original plan of production scheduling due to sudden disruption evernts may be no longer optimal, or even no feasible, which affects follow-up execution of the whole procduction scheduling process. How to generate a revising plan after disruption events to minimize the negative impact which is cauesed by disruption events on the production scheduling system is a core problem. The enterprise should consider many factors to make the revising plan, such as recovery time, recovery cost and so on. However, the method for solving the problem often consider only the recovery cost to make the new plan which will not to meet the actual require of enterprises. In this case, this paper uses disruption management to effectively measure disruption events, studies and discusses the model and solving method for production scheduling problem which is affected by the objective disruption events and the subjective-objective disruption events. This research includes the following aspects.(1) The method to measure disruption events is proposed since the existing method is difficult to measure the deviation in production scheduling system. Combining the prospect theory which is good at analyzing behavior with fuzzy theory which can express uncertain information, the deviation is measured by analyzing the effects of disruption events on the customer, manager and worker. The value functions of the different objectives are constructed and normalized respectively. The dissatisfactory membership functions of the different objectives are obtained. In consideration of human behavior, the functions to measure the behavior perception are constructed, which is helpful to contribute to the theory and method of disruption management.(2) The model of disruption management in production scheduling problem affected by disruption events is constructed. According to difference of coupling between machines, the production scheduling problem which is affected by the disruption events may simplify into more single machine scheduling problems or a job shop scheduling problem. The models of disruption management of production scheduling problem which is affected by objective disruption events and the subjective-objective disruption events are constructed respectively based on the different methods of measuring disruption events. It provides the theory foundation for requiring revising plan which has minimum disruption. (3) An improved muti-objective particle swarm optimization is proposed to meet characteristics of the above mode solution. Adjusting the inertia weight is used to speed up velocity to meet need of real-time solution of disruption management model of production scheduling problem; Updating the extremum, variancing optimization solution and so on are adopted to expand the solution space and avoid prematurity. Experiment indicates our method may satisfy solving need of disruption management model of production scheduling problem.