An analysis of six greedy selection rules on a class of zero-one integer programming models

Abstract

Six greedy primal selection rules are evaluated on a class of generalized set packing models. The evaluation is conducted in accordance with experimental design methodologies proposed by Lin and Rardin. Results indicate that the simplest of rules performs best, except when the model constraints exhibit “mixed” slackness. In this case, the rule proposed earlier by Fox and Scudder performs best. The results clarify and add detail to previous work by Fox and Scudder.