An efficient chemical reaction algorithm for multi-objective combinatorial bi-level optimization. Issue 4 (3rd April 2022)

• Record Type: Journal Article
• Authors: Abbassi, Malek; Chaabani, Abir; Said, Lamjed Ben
• Is Part Of: Engineering optimization. Volume 54:Issue 4(2022)
• Page Start: 665
• Page End: 686
• Publication Date: 2022-04-03
• Subjects: Multi-objective bi-level optimization -- chemical reaction optimization -- evolutionary algorithms -- combinatorial optimization; Engineering design -- Periodicals; Mathematical optimization -- Periodicals; 620.0042
• DOI: 10.1080/0305215X.2021.1895991
• Languages: English
• ISSNs: 0305-215X

Abstract

ABSTRACT: The Bi-Level Optimization Problem (BLOP) is defined as a mathematical program with two nested optimization tasks. Although many applications fit the bi-level framework, however, existing resolution methods were most proposed to solve single-objective bi-level problems. Regarding Multi-objective BLOPs (MBLOPs), there do not exist too many previous studies because of the difficulties associated with solving these complex problems. Additionally, a recently proposed metaheuristic, called Non-dominated sorting Chemical Reaction Optimization (NCRO), has been successfully applied to solve single-level Multi-Objective Problems (MOPs). NCRO applies a quick-non-dominated sorting technique that makes it one of the most powerful search algorithms in solving MOPs. Based on these observations, a new Bi-level Multi-objective CRO method, called BMCRO, is proposed in this article for solving MBLOPs. The main idea behind BMCRO is to come up with good solutions in an acceptable execution time within the bi-level framework. Experimental results on well-established benchmarks reveal the outperformance of the proposed algorithm against a bi-level variant of the Non-dominated Sorting Genetic Algorithm (NSGA-II) which is developed for this purpose.