A data envelopment analysis model for selecting material handling system designs

Abstract

The material handling system under design is an unmanned job shop with an automated guided vehicle that transport loads within the processing machines. The engineering task is to select the design alternatives that are the combinations of the four design factors: the ratio of production time to transportation time, mean job arrival rate to the system, input/output buffer capacities at each processing machine, and the vehicle control strategies. Each of the design alternatives is simulated to collect the upper and lower bounds of the five performance indices. We develop a Data Envelopment Analysis (DEA) model to assess the 180 designs with imprecise data of the five indices. The three-ways factorial experiment analysis for the assessment results indicates the buffer capacity and the interaction of job arrival rate and buffer capacity affect the performance significantly.