No abstract

No abstract

Two-sided Assembly Line Balancing (2S-ALB) is important in assembly plants that produce large-sized high-volume products, such as in automotive production. The 2S-ALB problem involves different assembly resources such as worker skills, tools, and machines required for the assembly. This research modelled and optimised the 2S-ALB with resource constraints. In the end, besides having good workload balance, the number of resources can also be optimised. For optimisation purpose, Particle Swarm Optimisation was modified to reduce the dependencies on a single best solution. This was conducted by replacing the best solution with top three solutions in the reproduction process.Computational experiment result using 12 benchmark test problems indicated that the 2S-ALB with resource constraints model was able to reduce the number of resources in an assembly line. Furthermore, the proposed modified Particle Swarm Optimisation (MPSO) was capable of searching for minimum solutions in 11 out of 12 test problems. The good performance of MPSO was attributed to its ability to maintain the particle diversity over the iteration. The proposed 2S-ALB model and MPSO algorithm were later validated using industrial case study. This research has a twofold contribution; novel 2S-ALB with resource constraints model and also modified PSO algorithm with enhanced performance.

No abstract

Mixed-model assembly line balancing problem (MMALBP) is an NP-hard problem which requires an effective algorithm for solution. In this study, an assessment of metaheuristic algorithms to optimize MMALBP is conducted using four popular metaheuristics for this problem, namely Particle Swarm Optimization (PSO), Simulated Annealing (SA), Ant Colony Optimization (ACO) and Genetic Algorithm (GA). Three categories of test problem (small, medium, and large) were used ranging from 8 to 100 number of tasks. For computational experiment, MATLAB software is used in investigate the metaheuristic algorithms performances to optimize the designated objective functions. The results reveal that ACO algorithm performed better in term of finding the best fitness functions when dealing with a large number of tasks. Averagely, it has produces better solution quality than PSO by 5.82%, GA by 9.80%, and SA by 7.66%. However, PSO more superior in term of processing time compared to ACO by 29.25%, GA by 40.54%, and SA by 73.23%. Hence, future research directions such as using the actual manufacturing assembly line data to test the algorithm performances are likely to happen.