A computerized search methodology for the design of mixed model assembly systems

Abstract

The majority of the line balancing and model sequencing techniques for the design of mixed model assembly systems are of the sequential search type. Assembly stations are sequentially optimized under some objective function and constraints. This thesis project presents a backtracking exhaustive search algorithm, implemented in Prolog language, for line balancing and model sequencing on mixed model assembly systems. The developed software permits the utilization of different objective functions. For line balancing, either the minimization of difference in work content or the smoothing of stations can be investigated. For model sequencing, either the penalty cost method or the minimization of assembly line length can be employed. A real world case problem was utilized to assess the developed methodology against the sequential search techniques and in all cases, better optimums were found.