A dimension-aware gaining-sharing knowledge algorithm for distributed hybrid flowshop scheduling with resource-dependent processing time

Abstract

Abstract The resource-assisted processing operation involves the coupling of multi-dimensional sub-problem, which poses a challenge in scheduling system. In this study, a dimension-aware gain-sharing knowledge algorithm (DGSK) is presented to address the distributed hybrid flowshop scheduling problem with resource-dependent processing times (DHFSP-RDPT), where the makespan is to be minimized. Firstly, by analyzing the mathematical model of the DHFSP-RDPT, four problem-specific lemmas and two novel resource reallocation rules are proposed. The DGSK begin with a high-performance initial population, which is generated by three knowledge-driven heuristics in hybrid way. Next, a discrete evolution-based search mechanism assists the DGSK to extend the search in solution space. Furthermore, a dimension-aware two-stage local search combined with meta-Lamarckian learning method is embedded to enhance the local search ability for the multidimensional problems. Finally, the proposed algorithm is measured on a series of instances based on real production data. The results demonstrate that the DGSK improves the performance by in solving DHFSP-RDPT compared to the state-of-the-art methods.

Keywords

• Hybrid flowshop scheduling problem
• Resource-dependent processing time
• Gain-sharing knowledge-based algorithm
• Dimension-aware local search