Distributed maintenance task scheduling for multiple technician teams considering uncertain durations and deterioration effects towards Industry 5.0
研究了工业5.0背景下多个技术团队的维护任务调度问题，考虑不确定工期和退化效应，提出改进变邻域搜索算法，在36个实例中平均提升完成概率39.7%。
Industry 5.0 emphasises a return to human-centricity to ensure outcomes for humans in the manufacturing system and create resilient and sustainable systems. This study explores the broader challenge of optimising maintenance scheduling in complex and dynamic industrial environments, focusing on developing human-centric and resilient approaches for enterprises under Industry 5.0. The combined impact of uncertain task durations and time-dependent deterioration effects are taken into consideration. The objective is to improve the system's robustness, quantified by multiplying completion probabilities. We propose Algorithm-1, which leverages the structural properties derived from analyzing the single maintenance team problem. Based on the structural properties of the multi-team scenario, we also propose Algorithm-2 to improve task assignment. Given the NP-hard nature of the problem, we further propose an improved variable neighbourhood search (IVNS), incorporating three local search methodologies and proposed algorithms to achieve near-optimal solutions within a reasonable timeframe. Comprehensive computational experiments on diverse problem instances indicate that the proposed IVNS algorithm outperforms other approaches, demonstrating significant robustness and efficiency. Specifically, across 36 experimental instances, IVNS achieved an average improvement of 39.7% relative to the lower bound, indicating a substantial enhancement in system-wide task completion probabilities.