Multi-objective lichtenberg algorithm: A hybrid physics-based meta-heuristic for solving engineering problems. (January 2022)

• Record Type: Journal Article
• Authors: Luiz Junho Pereira, João; Antônio Oliver, Guilherme; Brendon Francisco, Matheus; Simões Cunha Jr, Sebastião; Ferreira Gomes, Guilherme
• Is Part Of: Expert systems with applications. Volume 187(2022)
• Publication Date: 2022-01
• Subjects: Multi-objective optimization -- Multi-objective lichtenberg algorithm -- Lichtenberg algorithm -- Meta-heuristics -- Lightening -- Limited diffusion aggregation -- Design; Expert systems (Computer science) -- Periodicals; Systèmes experts (Informatique) -- Périodiques; Electronic journals; 006.33
• DOI: 10.1016/j.eswa.2021.115939
• Languages: English
• ISSNs: 0957-4174

Abstract

Highlights: A Brief Review of Multi-Objective Optimization with Algorithms and Trends. A new multi-objective algorithm inspired by lightning and Lichtenberg Figures. The first hybrid multi-objective meta-heuristic: trajectory and population - based. Lichtenberg patterns presents advantageous convergence-coverage balance. Tested on complex functions-problems and compared with renowned MO meta-heuristics. Abstract: With the advancement of computing and inspired by optimal phenomena found in nature, several algorithms capable of solving complex engineering problems have been developed. This work details the development of the Multi-objective Lichtenberg Algorithm, the version capable of dealing with more than one objective of a newly created meta-heuristic inspired by the propagation of radial intra-cloud lightning and Lichtenberg figures. The algorithm considers in its optimization routine a hybrid system based on both the population and the trajectory, demonstrating a great capacity for exploration and exploitation since it distributes points to be evaluated in the objective function through a Lichtenberg figure that is shot in sizes and different rotations at each iteration. The Multi-objective Lichtenberg Algorithm (MOLA) is the first hybrid multi-objective meta-heuristic and was tested against traditional and recent meta-heuristics using famous and complex test function groups and also constrained complex engineering problems. Regarding important metrics for convergence and coverage assessment, the Multi-objective Lichtenberg Algorithm proved to be a promising multi-objective algorithm surpassing others traditional and recent algorithms such as NSGA-II, MOPSO, MOEA/D, MOGOA and MOGWO with expressive values of convergence and maximum spread.