Optimal scheduling of probabilistic repetitive projects using completed unit and genetic algorithms

In this paper we introduce the completed unit algorithm (CU-AL), a probabilistic scheduling methodology for repetitive projects. The algorithm has two main advantages, simplicity and short computational time, that facilitate and expedite its use in simulation modeling and optimization. An integration between CU-AL and genetic algorithm (GA) is established to optimize the problem of maximizing profit for repetitive projects with probabilistic activity durations. This integration between CU-AL and GA is explained in detail through an example project with 5 activities and 10 repetitive units. A simulation model for this project is developed in Stroboscope, an activity-based simulation system. The optimization is performed by ChaStrobeGA, a Stroboscope add-on using genetic algorithm to optimize the overall objective function of project profit. Discussion of the results provides insight into the tradeoff between maintaining and relaxing resource continuity constraints in order to maximize expected project profit.