Reactive heuristics for disrupted multi-mode resource-constrained project scheduling problem

Abstract

This paper accounts for the mode change disruption in the Multi-mode Resource-Constrained Project Scheduling Problem (MRCPSP). In fact, during its execution, the project may face some unexpected events, which can lead to the schedule deterioration or even unfeasibility. A reactive mathematical modeling is proposed. Moreover, three heuristics with several variants are suggested to repair the initial disrupted schedule. These heuristics aim to perform a very quick response to provide the project manager with, hopefully, a feasible schedule. An extensive computational study is performed on PSPLIB benchmark instances with complementary generated data. Finally, the experimental results have shown that the three heuristics outperform Cplex in terms of the CPU running time. Moreover, Lexicographic Regret-Based Heuristic (LRBH) outranks the other heuristics in finding feasible solutions and in terms of the makespan deviation.