Stochastic optimization of the cost of automatic assembly systems

This paper extends a study of Liu and Sanders in which they evaluated the performance of asynchronous automatic assembly systems based on system throughput or production rate. In contrast, this paper approaches the design problem by attempting to minimize the manufacturing cost. Similar to Liu and Sanders, we apply the stochastic quasigradient method in order to search for the optimal solutions, consisting of the total number of pallets and the number of buffer spaces between stations. A comparison between the ‘maximum throughput’ problem and the ‘minimum cost’ problem is made.