A procedure for dealing with multiple objectives in cell formation decisions

A procedure for dealing with multiple objectives in cell formation decisions is proposed. Theoretical foundations from multiobjective decision making have been utilized to provide a framework for the study. The purpose of the study is to develop a procedure that provides managers with a means of developing multiple objective cell formation decisions as an alternative to intractable integer programming models. The authors use a best random seeds heuristic to generate a large number of alternatives. Nondominated solution theory is used to develop a list of potentially useful alternatives. Finally, preference cone theory is used to aid decision maker selection from these nondominated alternatives. Each of these three technologies have been developed elsewhere. This study considers aspects of applying these technologies to cell formation decisions. The procedure and its theoretical underpinnings are explained step-by-step and illustrated in a four-objective, four-cell, 24-machine, and 50-part cell formation problem. The example contains 12 nondominated solutions and requires only four pairwise comparisons of nondominated solutions by a decision maker before convergence to the preferred solution. Computer processing requirements are not burdensome, and decision maker participation is expected to be efficient.