A decomposition approach for the project scheduling problem (network analysis)

The problem addressed in this research is the project scheduling problem. Two types of solution methodologies commonly discussed in the literature are optimization procedures and heuristic ranking procedures. Optimization procedures are only capable of solving problems of moderate size. Heuristic ranking procedures, on the other hand, can handle problems of virtually any size. However, results reported in the literature demonstrate that no single priority rule is a consistent best performer when applied to problems with different characteristics. Based on this research, a heuristic decomposition procedure which shows promise as a more consistent performer than heuristic ranking procedures is developed. In this procedure, a sequence of subproblems is generated and resource conflicts among activities in each subproblem are optimally resolved. Some limited computational experience with the method is reported.