An investigation of mixed-model assembly line balancing problem with uncertain assembly time in remanufacturing

In recent decades, remanufacturing has emerged as an effective way to address resource crises and environmental pollution issues. Unlike traditional manufacturing, remanufacturing production is filled with various variable factors, especially in the assembly phase. Due to changes in part types, quality conditions, and assembly methods, the assembly time becomes highly uncertain. Assembly line balancing is a key challenge to achieve the stable operation of remanufacturing system. This study proposes an evaluation method for remanufacturing assembly time and establishes a multi-objective mathematical model for balancing remanufacturing mixed- model assembly (RMMA) line. The evaluation method utilizes the Fuzzy Graphical Evaluation Review Technique (FGERT) network to predict expected assembly time for each operation. The balancing model aims to optimize remanufacturing takt time and comprehensive balance rate (CBR). To effectively solve this model, an adaptive double-layer genetic algorithm (ADGA) is designed, where layer I ensures production efficiency and layer II optimizes assembly line balance. Finally, an assemble example of high-pressure common rail fuel pumps (HCRFP) is used to validate the effectiveness of the proposed method. The results demonstrate notable improvements compared to traditional single-product assembly (TSPA) line in scenarios with workstation numbers 4, 5, 6, and 7. Specifically, the production takt time is reduced by 4.19% to 9.56%, and CBR is enhanced by approximately 50%. Further comparison with three other classic algorithms confirms the superiority of ADGA. Additionally, it is observed that remanufacturability (proportion of remanufactured parts) has a significant impact on assembly performance. As remanufacturability increases, both takt time and CBR increase, reaching their maximum values when remanufacturability is around 0.5.