A genetic algorithm for proactive project scheduling with resource transfer times

In this paper, we investigate the proactive resource-constrained project scheduling problem with resource transfer times under uncertain environment, aiming to generate robust baseline schedules that will be as stable as possible during project execution. The contribution of this paper is threefold. Firstly, the proactive project scheduling problem with resource transfer times is proposed and modeled as an integrated (one-phase) robust project scheduling problem where decisions regarding activity scheduling and resource transfers are simulta-neously considered. Secondly, the computational complexity of the proposed problem is proved to be NP-hard in the strong sense, and a genetic algorithm (GA) is developed to solve this problem, in which two measures are proposed to respectively improve the efficiency and the effectiveness of the GA. Thirdly, through analyzing the results from the designed computational experiment, one interesting phenomenon is found that the consideration of breakable flows as well as a local search of resource transfer priority rules in the decoding procedure does not improve schedule robustness so much but costs much more computing time. Another finding is that the resource transfer priority rule we propose, resource transfer efficiency (RTE), is more likely to obtain the optimal solutions for the tested PSPLIB instances.

Automated summary

This paper investigates the proactive resource-constrained project scheduling problem with resource transfer times, proposes a novel robust project scheduling model, and solves it using a genetic algorithm. Experimental results show that considering breakable flows and local search in the decoding process does not significantly improve schedule robustness, but increases computational time.