Semiconductor final testing scheduling using Q-learning based hyper-heuristic
作者：
Highlights：
• A QHH algorithm is proposed for solving the SFTSP with makespan criterion.
• An efficient encoding and decoding pair is presented to generate feasible schedules.
• Eight simple heuristic rules are designed to construct a set of LLHs.
• Q-learning algorithm is employed as a high-level strategy to intelligently select the LLHs.
• The performance of the proposed QHH is evaluated on a set of benchmark scenarios.
摘要
•A QHH algorithm is proposed for solving the SFTSP with makespan criterion.•An efficient encoding and decoding pair is presented to generate feasible schedules.•Eight simple heuristic rules are designed to construct a set of LLHs.•Q-learning algorithm is employed as a high-level strategy to intelligently select the LLHs.•The performance of the proposed QHH is evaluated on a set of benchmark scenarios.
论文关键词：Q-learning,Hyper-heuristic,Semiconductor final testing,Scheduling
论文评审过程：Received 3 February 2021, Revised 26 July 2021, Accepted 23 September 2021, Available online 28 September 2021, Version of Record 16 October 2021.
论文官网地址：https://doi.org/10.1016/j.eswa.2021.115978
