A sequencing problem for a mixed-model assembly line with stochastic operation times

Shigenori Kotani,Zhao Xiaobo,Katsuhisa Ohno

2004-01-01

Abstract:There are many papers that discuss the sequencing problems of mixed-model assembly lines. It is well known that leveling the work loads at each work station on the assembly line is important. Therefore, the sequencing problems for leveling the work loads have been studied by many researchers and numerous algorithms have been presented. In these papers it is assumed that the operation times of a worker at each work station are deterministic, not stochastic. In fact, the operation times are variable in real production and are determined when the worker finishes his operations. Therefore, we should consider that the operation times of the sequencing problem are stochastic. In this paper, we discuss a sequencing problem for a mixed-model assembly line with stochastic operation times. Firstly, in order to express the operations of each work station on an assembly line with stochastic operation times, we derive a calculation method of the expectations of worker start positions, worker end positions after the operations are completed and utility work. These three expectations cannot be calculated analytically. Then, we propose an approximation calculation method. Though the operation times are continuous random variables, we consider the operation times as discrete random variables and approximately compute the three expectations using transition probabilities. We show that this calculation method is accurate. Using this method, we propose an algorithm, based on simulated annealing, that minimizes the total expectation of utility work. Moreover, we compare a sequencing schedule obtained by our algorithm with a sequencing schedule for minimizing the total utility work through a numerical example, and we demonstrate that these two sequencing schedules are different from each other. Therefore, it is important to determine the objective function of the sequencing problem according to the various related conditions.