摘要

We consider the order acceptance and scheduling problem in a flow shop where the objective is to simultaneously minimize makespan and maximize total net revenue. We formulate the problem as a mixed integer linear programming model and develop an effective parallel neighborhood search algorithm. Two-string representation and three neighborhood structures are applied to generate new solutions. Parallelization is implemented by applying two independent searches and directly exchanging information between them. We assess the performance of the proposed method via computational experiments using an extensive set of instances. The experimental results show that the proposed method is highly effective and competitive when compared to other algorithms.