Solving Constrained Multiple Resource Networks Both Forward and Backward Using Brooks Algorithm

Abstract

Many Brooks algorithm (BAG)-based heuristics have been proposed to solve the constrained multiple resource problem. BAG starts at the beginning of a project network, determines a list of available tasks, and then uses a heuristic to select a task to schedule. This research modifies BAG to solve the project network in both a forward and backward (F&B) direction. This modified BAG approach was used on Patterson's 110 test problems with several selection heuristics, such as ACTIM, latest finish time (LFT), resource over time (ROT), and ACTRES. Results of this research show that solving the networks in both the F&B directions led to improvements of more than 30% for the average error, the maximum error and the percentage of optimum solutions found when compared to the traditional BAG approach.