The mixed and multi model line balancing problem: a comparison

This paper gives a review of the literature on the mixed and the multi model line balancing problem. Compared with the single model line little attention has been paid to the mixed and the multi model line balancing problem. The mixed model line balancing problem can be subdivided into two related problems. The balancing problem deals with the assignment of work elements to work stations. The sequencing problem deals with the determination of the sequence in which the models are produced. It appears that the literature on the mixed model line balancing problem consists of two different approaches that transform the mixed model line into a single model line balancing problem. To accomplish this transformation, the first procedure defines a combined precedence diagram and the second approach uses average task processing times. An experiment was carried out in order to compare several heuristics that are based on the combined precedence diagram, on their performance. An optimisation method was added. The results indicate that the position of common work elements in the precedence diagram of the different models influences both the CPU time of a heuristic and the way in which the total work content of the single models is spread over the work stations. The results also show that good solutions with regard to the number of required stations go together with long CPU times. For a number of instances that took a long CPU time, we can however decrease these CPU times considerably (about factor 20) without deteriorating the performance of the methods by applying the heuristics to the reversed combined precedence diagram. Eindhoven University of Technology, Department of Operations Planning and Control, Paviljoen F14, P.O. Box 513, 5600 MB Eindhoven, The Netherlands