Bi-objective robust project scheduling with resource constraints and flexible activity execution lists

Abstract

This paper analyses a type of extension of the resource-constrained project scheduling problem (RCPSP) whose activities have multiple execution lists. This sort of problem is titled resource-constrained project scheduling problem with flexible activity execution lists (abbreviated as RCPSP-FAL). The goal is to minimize makespan and maximize robustness simultaneously to improve the ability of the project to handle emergencies within the design completion time. Six robustness measures are proposed depending on the free slack determined by start time and renewable resource surplus per unit time. Moreover, a bi-objective mixed-integer programming model is developed to solve the considered problem. Besides, a simple data structure, namely selection-tree, is designed to detect available activity execution lists. Two algorithms are proposed, whose names are variant-genetic algorithm (VGA) and variant-simulated-annealing algorithm (VSA), to solve the developed model. Also, an aggregation function is established for assessing the fitness consisting of makespan and robustness. Several computation experiments are designed to evaluate the proposed algorithms and the six robustness measures. Results show the characteristics of different robust measures. Meanwhile, they also show the effectiveness of VSA and VGA, as well as the relations among robustness, makespan, and resource-usage.