A bi-objective optimisation of post-disaster relief distribution and short-term network restoration using hybrid NSGA-II algorithm. Issue 19 (2nd October 2022)

• Record Type: Journal Article
• Authors: Ransikarbum, Kasin; Mason, Scott J.
• Is Part Of: International journal of production research. Volume 60:Issue 19(2022)
• Page Start: 5769
• Page End: 5793
• Publication Date: 2022-10-02
• Subjects: Humanitarian logistics -- short-term restoration -- bi-objective programming -- metaheuristic -- non-dominated sorting genetic algorithm-II -- hypervolume; Factory management -- Periodicals; 658.57
• DOI: 10.1080/00207543.2021.1970846
• Languages: English
• ISSNs: 0020-7543

Abstract

Abstract : Humanitarian logistics research has recently received tremendous interest from researchers and practitioners due to its importance in assisting relief operations. While there is an increasing trend for mathematical models related to preparedness and response phases for disaster operations management, recovery-phase models are not as emphasised as other phases due to scarce data and model complication from NP-hard nature of the models. One particular approach that can provide a sufficiently good solution for the NP-hard problems is the metaheuristic approach. In this research, we explore the bi-criteria integrated response and recovery model for making strategic post-disaster decisions in the relief distribution and short-term network restoration. Next, with a focus on considering conflicting objectives between fairness and cost of this problem, we propose a hybrid approach with its evolutionary component based on the non-dominated sorting genetic algorithm-II (NSGA-II) called HNSGA-II. The proposed HNSGA-II is compared against the exact method using the approximate Pareto-front analysis. The proposed algorithm is verified using a case study from a risk assessment tool called Hazus to illustrate how to cope with the aftermath of an earthquake. Finally, results are evaluated using a Hypervolume-based technique and computation time to illustrate the efficiency of the proposed algorithm.