Abstract
Casting production scheduling problem (CPSP) has attracted increasing research attention in recent years to facilitate the profits, efficiency, and environment issues of casting industry. Casting is often characterized by the properties of intensive energy consumption and complex process routes, which motivate the in-depth investigation on construction of practical multi-objective scheduling models and development of effective algorithms. In this paper, for the first time, the multi-objective casting production scheduling problem (MOCPSP) is constructed to simultaneously minimize objectives of defective rate, makespan, and total energy consumption. Moreover, a neighborhood structure enhanced discrete NSGA-II (N-NSGA-II) is designed to better cope with the proposed MOCPSP. In the N-NSGA-II, the advantage of selection mechanism of NSGA-II is fully utilized for selecting non-dominate solution, three neighborhood structures are elaborately designed to strengthen the ability of the local search, and a novel solution generating approach is proposed to increase the diversity of solutions for global search. Finally, a real-world case is illustrated to evaluate the performance of the N-NSGA-II. Computational results show that the proposed N-NSGA-II obtains a wider range of non-dominated solutions with better quality compared to other well-known multi-objective algorithms.
An effective multi-objective discrete grey wolf optimizer for a real-world scheduling problem in welding production
