Chaotic lightning search algorithm
- PDF / 1,831,946 Bytes
- 17 Pages / 595.276 x 790.866 pts Page_size
- 35 Downloads / 249 Views
METHODOLOGIES AND APPLICATION
Chaotic lightning search algorithm Mohamed Wajdi Ouertani1,2 · Ghaith Manita1,3 · Ouajdi Korbaa1,4
© Springer-Verlag GmbH Germany, part of Springer Nature 2020
Abstract Metaheuristics have proven their efficiency in treating complex optimization problems. Generally, they produce good results quite close to optimal despite some weaknesses such as premature convergence and stagnation in the local optima. However, some techniques are used to improve the obtained results, one of them is the adoption of chaos theory. Including chaotic sequences in metaheuristics has proven its efficiency in previous studies by improving the performance and quality of the results obtained. In this study, we propose an improvement of the metaheuristic lightning search algorithm (LSA) by using chaos theory. In fact, the idea is to replace the values of random variables with a chaotic sequences generator. To prove the success of the metaheuristic—chaos theory association, we tested five chaotic version of lightning search algorithm on a benchmark of seven functions. Experimental results show that sine or singer map are the best choices to improve the efficiency of LSA, in particular with the lead projectile update. Keywords Metaheuristic · Lightning search algorithm · Chaos · Chaotic maps
1 Introduction Recent studies in the field of optimization have shown a growing interest in problem-solving methods using metaheuristics. In fact, metaheuristics have advantages over other deterministic resolution approaches. Among its advantages are: simplicity, the possibility to solve large scale and nonlinear problems and their flexibility. Metaheuristics deal with complex optimization problems such as manufacturing systems design (Gen and Cheng 1996), mechanical engineering (He et al. 2004), flowshop scheduling (Murata and Ishibuchi 1994), image enhancement and segmentation (Paulinas and Ušinskas 2007), transport problem solving (Vignaux and Michalewicz 1991), calibration of fractional fuzzy controllers (Zhou et al. 2019a) and data clustering (Pacheco et al. 2018; Zhou et al. 2019b, 2017). Communicated by V. Loia.
B
Mohamed Wajdi Ouertani [email protected]
1
Laboratory MARS, LR17ES05, ISITCom, University of Sousse, Sousse, Tunisia
2
ENSI, University of Manouba, Manouba, Tunisia
3
ESEN, University of Manouba, Manouba, Tunisia
4
ISITCom, University of Sousse, Sousse, Tunisia
In general, metaheuristics try to imitate the behaviour of living beings or reproduce biological and physical phenomena. They can be classified into categories, population-based [Particle swarm optimization (PSO) (Eberhart and Kennedy 1995), Genetic algorithm (GA) (Holland 1992)] and singlepoint search [Itarated local search (ILS) (Lourenço et al. 2003), Simulated annealing (SA) (Van Laarhoven and Aarts 1987)], memoryless (SA) and memory usage methods [Ant colony optimization (ACO) (Dorigo and Di Caro 1999)], nature-inspired [GA, PSO, SA, ACO, Harris hawks optimizer (HHO) (Heidari et al. 2019)] and non-nature inspiration me
Data Loading...
