A Hybrid Method of Heuristic Algorithm and Constraint Programming for No-wait Integrated Scheduling Problem

Abstract

No-wait Integrated Scheduling Problem (NISP) describes a real-life process of the non-standard products where the consideration is given to the great structure differences, processing parameter differences, no-wait constraint, and the need for further deep processing after assembly of jobs. To deal with the dynamic orders of non-standard products, the scheduling algorithm to be design should be a dynamic algorithm with the ability to deal with the above conditions. At first, the dynamic scheduling problem is transformed to a series of continuous static scheduling problem by adoption of window-based event-driven strategy, thus establishing constraint programming model targeted at minimal total tardiness and thereby proposing a hybrid method of Heuristic Algorithm and Constraint Programming (HA-CP) for the problem. In order to enhance the ability to response the dynamic orders of non-standard products, HA-CP adopts heuristic algorithm to generate a pre-scheduling solution at each dynamic event moment, so that jobs that fall into the window period are labelled as dispatched jobs, while the remaining jobs are labelled as jobs to be dispatched. To improve solution quality, the jobs to be dispatched are mapped into an operation-based constraint programming model, then, during the execution interval of dispatched jobs, constraint programming solver starts to solve the jobs to be dispatched and update the current solution if the solver gets a better solution within the execution interval. The above procedures are repeated until all jobs are scheduled. Finally, the results of simulation experiment show that the proposed algorithm is effective and feasible.