Static scheduling method for aircraft flat-tail assembly production based on improved bi-level genetic algorithm

Abstract

Abstract Aircraft flat-tail assembly is a complex process that involves multiple assembly processes, multiple parallel frames, and multi-configuration mixed flow assembly, and its assembly processes exhibit extended processing times (typically measured in days) combined with temporal fluctuations arising from human factors, leading to a certain degree of uncertainty in single-process durations, thereby presenting a complicated flexible flow-shop scheduling problem (FFSP), which is a typical NP-hard problem. Despite its significance, the research on FFSP in aircraft flat-tail assembly production scheduling is limited. This study proposes an improved bi-level genetic algorithm to address the two sub-problems of flat-tail assembly production scheduling: frame assignment and assembly task sequencing. The objective is to minimize the maximum delay penalty cost. A two-stage coding scheme is introduced for frame assignment and task sequencing, respectively. To mitigate genetic algorithms’ convergence to local optima and enhance positive feedback, we implement a variable neighborhood search mechanism combined with elite retention. The efficacy of the improved bi-level genetic algorithm is evaluated through experiments and case studies in enterprises, indicating a significant impact on the assembly production scheduling of flat-tail, with potential applications to similar large and complex equipment. Overall, this study contributes to FFSP research in aircraft flat-tail assembly production scheduling by offering a novel solution approach to effectively address the sub-problems of frame assignment and assembly task sequencing.