LAUSR

Abstract The resource leveling problem (RLP) involves the development of a project schedule that specifies the starting time of each activity constrained by the project deadline and precedence, aimed at the minimizing the variation of the resource utilization. To shorten project duration and to find more effective and flexible object functions, normalized entropy and relative entropy are proposed. The former considers shortening the project duration, improved based on the entropy model; the latter measures the similarity of two distributions, that is the actual versus the theoretical resource allocation based on the resource assignments per activity, and which covers not only entropy itself, but also cross entropy. Finally, discrete particle swarm optimization (DPSO) is encoded to investigate the performance of the proposed normalized entropy and relative entropy with other objective functions.