Resource leveling in projects with stochastic minimum time lags

In project management, resources and time are two critical aspects influencing the success of a project. On the one hand, resource leveling, an effective resource optimization technique, is widely adopted to guarantee the efficient use of resources. On the other hand, to deliver a project as soon as possible, it can typically be accelerated by overlapping some activities. In a real-life project environment, uncertainty is inevitable and further complicates resource leveling and activity overlapping. However, existing research tends to study resource leveling and activity overlapping separately and little attention has been paid to level resource usage with uncertain activity overlapping. Therefore, the authors model activity overlaps as minimum time lags and study the resource leveling problem with stochastic minimum time lags (RLP-SMTL), where both the time lags and the activity durations are uncertain. This study aims to obtain a scheduling strategy such that the usage of renewable resources is as smooth as possible over time. The tuple represented by a random key vector, a strategy dynamically schedules activities at each decision point. A simulation-based solution framework for the RLP-SMTL is proposed. Built upon the proposed solution framework, two metaheuristics, an evolutionary algorithm (EA) and a bat algorithm (BA), are designed. Based on 1,080 randomly generated 100-activity instances, extensive computational experiments are performed to evaluate the effectiveness of the proposed algorithms. The results reveal that the EA outperforms the BA in terms of both the objective function's value and the timely project completion probability. Although the strategies generated by the BA are slightly weaker than the EA, the BA is much faster than the EA. The results obtained by an additional comparison experiment further show that the proposed algorithms outperform the existing best-performing metaheuristic. Additionally, an example project is adopted to illustrate how the proposed approach can be applied to practical resource leveling in construction projects. In conclusion, this paper contributes to the body of knowledge in construction engineering and management by developing effective metaheuristics that equip the project manager with an automated tool to make effective resource leveling decisions under uncertainties.