Heuristics for selecting machines and determining buffer capacities in assembly systems

This paper considers a design problem of assembly systems consisting of stations and buffers that connect the stations. We present a solution procedure for finding the minimum cost configuration which gives a desired throughput rate for an assembly system. The configuration is defined by the machines to be used in stations (machine configuration) and capacities of buffers. Three heuristics are proposed, which simultaneously select the machines to be used in the stations and determine the capacities of the buffers. These heuristics start from an initial configuration such as a lower configuration, which consists of less efficient machines and large size buffers, or an upper configuration, which consists of more efficient machines and small size buffers. Then, the heuristics search for a near optimal solution by repeatedly generating promising machine configurations and determining (near) optimal buffer sizes for the machine configurations. Results of computational experiments show that the proposed heuristics give relatively good configurations in a reasonable amount of time.