A smart multi-objective differential evolution algorithm for energy-efficient scheduling in parallel batch processing machines.
Uloženo v:
| Název: | A smart multi-objective differential evolution algorithm for energy-efficient scheduling in parallel batch processing machines. |
|---|---|
| Autoři: | Chen, Yarong1 (AUTHOR) yarongchen@126.com, Su, Hansen1 (AUTHOR) 18072083688@163.com, Rauf, Mudassar1 (AUTHOR) mudassar@wzu.edu.cn, Zhao, Xue1 (AUTHOR) zx2710750937@163.com, Wang, Chen1 (AUTHOR) 1958813143@qq.com |
| Zdroj: | Applied Soft Computing. Mar2026, Vol. 189, pN.PAG-N.PAG. 1p. |
| Témata: | Differential evolution, Scheduling, Manufacturing industries, Heuristic, Soft computing, Multi-objective optimization, Batch processing, Greenhouse gas mitigation |
| Abstrakt: | In the context of the pressing need for sustainable engineering practices and addressing environmental challenges, integrated energy-efficient production technologies have become essential for achieving carbon neutrality. This study leverages soft computing to investigate the parallel batch processing machine scheduling problem (PBPMSP) under complex constraints, encompassing heterogeneous job attributes (such as processing times, due dates, sizes, and arrival times) and limited machine capabilities. To effectively handle the multi-criteria decision-making inherent in this problem, we propose a smart multi-objective differential evolution algorithm based on decomposition (SMODE/D). This algorithm incorporates three improvements: (1) an adaptive smart learning-based strategy is designed to dynamically optimizes control parameters and mutation operators within the differential evolution framework, (2) hybrid heuristics that integrate domain knowledge to effectively coordinate batch formation and scheduling, (3) a tri-objective optimization framework based on decomposition to minimize three objectives concurrently: makespan, total earliness and tardiness, and overall energy consumption. The comparative experiments show that SMODE/D surpasses four leading algorithms (NSGA-II, NSGA-III, MODE/D, and MOABC) regarding its convergence, diversity, and robustness, achieving average improvements of 79.2 % in the diversity (inverted generational distance), 87.3 % in (non-dominated Rate), and 50.4 % in convergence (c-matrix). These findings provide valuable insights for balancing operational efficiency and energy savings, ultimately contributing to carbon footprint reduction in industrial manufacturing. • A smart MOEA/D with novel SLBS and problem-specific heuristic is proposed for PBPMSP. • Energy, earliness/tardiness, and makespan are optimized simultaneously. • SLBS dynamically adjusts control parameters and mutation operator selection. • SMODE/D outperforms MODE/D, NSGA-II, and NSGA-III in convergence and robustness. [ABSTRACT FROM AUTHOR] |
| Databáze: | Supplemental Index |
Full Text Finder 
Nájsť tento článok vo Web of Science Buďte první, kdo okomentuje tento záznam!