Problem Specific Variable Selection Rules for Constraint Programming: A Type II Mixed Model Assembly Line Balancing Problem Case

Abstract

The main idea of constraint programming (CP) is to determine a solution (or solutions) of a problem assigning values to decision variables satisfying all constraints. Two sub processes, an enumeration strategy and a consistency, run under the constraint programming main algorithm. The enumeration strategy which is managing the order of variables and values to build a search tree and possible solutions is crucial process in CP. In this study problem-based specific variable selection rules are studied on a mixed model assembly line balancing problem. The 18 variable selection rules are generated in three main categories by considering the problem input parameters. These rules are tested with benchmark problems in the literature and experimental results are compared with the results of mathematical model and standard CP algorithm. Also, benchmark problems are run with two CP rules to compare experimental results. In conclusion, experimental results are shown that the outperform rules are listed and also their specifications are defined to guide to researchers who solve optimization problems with CP.