Multi-objective worker allocation optimisation in a multiple U-line system

Purpose This paper aims to evaluate two operational modes of the worker allocation problem (WAP) in the multiple U-line system (MULS). Five objectives are optimised simultaneously for the most complicated operational modes, i.e. machine-dominant working and fixed-station walking. Besides, the benefits of using multiline workstations (MLWs) are investigated. Design/methodology/approach The elite non-dominated sorting differential evolutionary III (ENSDE III) algorithm is developed as a solution technique. Also, the largest remaining available time heuristic is proposed as a baseline in determining the number and utilisation of workers when the use of MLWs is not allowed. Findings ENSDE III outperforms the cutting-edged multi-objective evolutionary algorithms, i.e. multi-objective evolutionary algorithm based on decomposition and non-dominated sorting differential evolutionary III, under two key Pareto metrics, i.e. generational distance and inverted generational distance, regardless of the problem size. The best-found number of workers from ENSDE III is substantially lower than the upper bound. The MULS with MLWs requires fewer workers than the one without. Research limitations/implications Although this research has extended several issues in the basic model of multiple U-line systems, some assumptions were used to facilitate mathematical computation as follows. The U-line system in this research assumed that all lines were produced only a single product. Besides, all workers were well-trained to gain the same skill. These assumptions could be extended in the future. Practical implications The implication of this research is the benefits of multiline workstations (MLWs) used in the multiple U-line system. Instead of leaving each individual line to operate independently, all lines should be working in parallel through the use of MLWs to gain benefits in terms of worker reduction, balancing worker’s workload, higher system utilisation. Originality/value This research is the first to address the WAP in the MULS with machine-dominant working and fixed-station walking modes. Worker’s fatigue due to standing and walking while working is incorporated into the model. The novel ENSDE III algorithm is developed to optimise the multi-objective WAP in a Pareto sense. The benefits of exploiting MLWs are also illustrated.