Scheduling Aircraft Cabin Door Assembly Line via Q-Learning Strategy-Assisted Particle Swarm Optimizer

Unofficial AI-generated summary based on the public title and abstract. Not an official translation.

🔬研究者:Provides a novel Q-learning-based PSO for bi-objective scheduling with energy consumption, offering a benchmark for future research in sustainable manufacturing.

🏢実務担当者:Offers a scheduling optimizer and early-warning system that can reduce energy costs and mitigate supply chain disruptions in assembly lines.

The aircraft cabin door assembly line scheduling problems are challenging with complex constraints, such as the allocation of auxiliary processing tools, temperature related processes, and materials delay caused by supplier uncertainty. To tackle these challenges, this work formulates the problems as bi-objective distributed hybrid flowshop scheduling problems with limited auxiliary tools and speed-adjustment machines. First, a MILP model is developed and validated to characterize the actual constraints with the goal of minimizing total tardiness/earliness and total energy consumption. Second, an improved particle-swarm optimization framework with novel Q-Learning assistant strategies is proposed for solving it. The algorithms innovate in three key aspects. 1) A two-phase searching process with an adaptive adjustment mechanism for communication interval is designed for efficient Pareto front exploration; 2) Based on the problem-specific features, eight neighborhood structures are developed and the corresponding search operators are designed to speed up its solution process; and 3) Two Q-Learning-based adaptive enhancements are proposed to dynamically select optimal local search operators during the process. Third, the effectiveness of the proposed enhancement strategies is verified on 48 benchmark cases. The experimental results and discussions show that the proposed algorithm outperforms the state-of-the-art algorithms. Finally, we supply the analysis of critical-components’ early-warning scheduling for a real assembly instance. The permissible idle time windows for the critical-components are provided as early-warnings to avoid the instability of supply chains. Note to Practitioners—The aircraft cabin door assembly line represents a critical and complex manufacturing environment, where the production efficiency and energy sustainability related objectives should be addressed. This paper models its scheduling problem subject to various constraints such that total earliness/tardiness and total energy consumption are minimized. The key practical constraints are related to auxiliary tool allocation, temperature-sensitive processes, and supplier-induced part shortages, which directly impact real-world production stability. We propose an improved particle swarm optimizer to solve the problem. The method introduces several innovations. Experimental validation against industry-derived benchmarks demonstrates that the proposed optimizer outperforms compared algorithms in both solution quality and computational efficiency. Future extensions could incorporate dynamic uncertain supplier deliveries or integrate carbon emission into the optimization model. The framework is also generalizable to other distributed manufacturing systems requiring sustainability-aware scheduling, such as electronics assembly. For practitioners, we further develop an early-warning approach that identifies permissible idle time windows for critical jobs by examining post-scheduling Gantt charts. The visualization highlights idle time windows caused by potential disruptions, enabling planners to proactively adjust production schedules.

• semanticscholar https://doi.org/10.1109/tase.2026.3716037first seen 2026-07-31 06:31:59 · last seen 2026-08-02 06:13:01