Tackling Demand Fluctuation through Novel Flexible Assembly Line Balancing Algorithm

This paper studied the situation faced by an upstream textile manufacturer company supplying seat cover fabrics to major automotive producers which need to dynamically adjust task distribution on their assembly line in response to demand fluctuations, while improving load balance among workstations. The company itself has implemented lean manufacturing concept, specifically the takt-time rules to organize the pace of its assembly line. Problems occur when demand increases significantly and one (or several) workstations in the assembly line suffers capacity insufficiency. To rectify the insufficiency, parallel processing becomes necessary. The company relies heavily on manual labor and has limited number of workers; therefore, parallel processing can only be implemented by utilizing the available workers. To do this, a new algorithm was designed. The algorithm works to identify which workstations suffer insufficiency, which workstations are available to perform parallel processing and how long the parallel processing should be performed. The algorithm is numerically tested. Four cases which cover various situations commonly found in real world assembly line are designed as test cases. The experiment showed that the new algorithm managed to reallocate tasks among workstations in such a way that the targeted takt-time was achieved while the line smoothness index was improved.