Balancing of mixed-model two-sided assembly lines with underground workstations: A mathematical model and ant colony optimization algorithm

Mixed-model assembly lines allow the production of different product variants in mass quantities on the same assembly line. In studies addressing mixed-model assembly with two-sided lines, assembly line (work)stations are classified as left-side or right-side stations depending on the operation side to which they are allocated. However, underground stations are also utilized in industry to perform tasks that need to be done underneath the product being assembled on the line. This paper introduces and mathematically formulates a mixed-model, two-sided assembly line balancing problem considering underground stations. The precedence relationships between tasks being performed in the three types of stations are defined and considered in the model. A numerical example is solved in GAMS (with CPLEX solver) and the detailed balancing solution is provided. A new ant colony optimization algorithm, in which the parameters are optimized using response surface methodology, is also developed to solve real-world problems. A total of 78 test problems are derived from the literature and their lower bounds are calculated to test the performance of the ACO algorithm. ACO finds optimum solutions for the majority of small and medium-sized test problems. In comparing the ACO results to the lower bounds for the large-sized problems, ACO finds near-optimum solutions in majority of the test cases.