Dynamic Optimization for Analyzing Effects of Multiple Resource Failures on Project Schedule Robustness

This study develops hybrid modeling and algorithmic frameworks for analyzing the mutual effects of multiple sources of uncertainty on the quality and the robustness of construction schedules. To cope with multiple sources of disruptions, i.e., random resource failures and severe weather conditions, this paper develops a simulation-optimization model that aims to generate delay resistant project schedules. The Variable Neighborhood Search (VNS) is hybridized with an event-driven simulation framework to generate efficient and robust solutions for computationally expensive resource-constrained project scheduling problems (RCPSP). The simulation experiments have been carried out by a flexible modeling framework that can be adopted by project experts to design construction schedules subject to the uncertainty associated with the multiple resource failure. The problem is mathematically formulated as a bi-objective optimization model aiming to minimize the project makespan and maximize a novel surrogate robustness function simultaneously. The computational results of the proposed VNS method have been compared with those obtained from the commercial optimization solvers. The simulation-optimization model's application is demonstrated through a case study of the hydropower plant construction project with multiple renewable and non-renewable resources. Based on an extensive statistical analysis of real-life scenarios, this study contributes to a trade-off analysis of project makespan and robustness in construction projects. The t-test statistical analysis results indicate the significance of the project's average delay reduction by implementing the robust project schedule. The outcomes confirm that the designed framework can generate a more efficient project schedule with a higher rate of protection compared with the existing robust approaches.