Experimental Evaluation of Overlapping Strategy for the Multimode Resource-Constrained Project Scheduling Problem

Abstract

Reduction in project completion time is important in today’s competitive environment. This paper presents a solution to the realistic problem of scheduling activities to minimize the project duration, in which the activities (a) require units of multiple renewable and non-renewable resources for which a limited availability is imposed, (b) are performed in one of several modes, and (c) can be executed in parallel by allowing the downstream activity to start before the end of the upstream activity based on preliminary information. To solve the above-mentioned problem, a mixed-integer nonlinear programming model is built and an efficient genetic algorithm combined with design structure matrix is proposed. A revised serial schedule generation scheme, which considers resource constraints, precedence relation constraints and overlapping relation constraints, is designed to decode the chromosome. An example illustrates that the implementation of overlapping strategy can benefit from multimode and thus accelerate the project process. Results based on a large set of generated instances demonstrate the effectiveness of the proposed algorithm and reveal the impacts of significant factors on the time reduction in project makespan.