Premier League Championship Algorithm: A Multi-population-Based Algorithm and Its Application on Structural Design Optim
The League Championship Algorithm (LCA) is a population-based algorithm motivated by competitions for the championship in sports leagues, in which each solution in the population is considered as the team formation adopted by a sport team. These artificia
- PDF / 1,116,440 Bytes
- 26 Pages / 439.37 x 666.142 pts Page_size
- 68 Downloads / 284 Views
Abstract The League Championship Algorithm (LCA) is a population-based algorithm motivated by competitions for the championship in sports leagues, in which each solution in the population is considered as the team formation adopted by a sport team. These artificial teams compete according to a given schedule generated based on a single round-robin logic. Using a stochastic method, the result of the game between pair of teams is determined based on the fitness value criterion in such a way that the fitter individual has more chance to win. Given the result of the games in the current iteration, each team preserves changes in its formation to generate a new solution following a SWOT-type analysis and the championship continues for several iterations. In this chapter, a Premier League Championship Algorithm (PLCA), which is an extended version of the LCA, is proposed for structural optimization based on the concept of post championship. The PLCA is a multi-population algorithm wherein each subpopulation forms a local league in which different individuals compete and produce new solutions. The performance of the PLCA method is investigated on three structural design test problems under displacement and stress constraints. Numerical results demonstrate that the PLCA seems to be a promising alternative approach for structural optimization problems.
A. Husseinzadeh Kashan (B) Faculty of Industrial and Systems Engineering, Tarbiat Modares University, Tehran, Iran e-mail: [email protected] S. Jalili Afagh Higher Education Institute, Urmia, Iran e-mail: [email protected] S. Karimiyan Department of Civil Engineering, Islamshahr Branch, Islamic Azad University, Islamshahr, Iran e-mail: [email protected] © Springer Nature Singapore Pte Ltd. 2019 A. J. Kulkarni et al. (eds.), Socio-cultural Inspired Metaheuristics, Studies in Computational Intelligence 828, https://doi.org/10.1007/978-981-13-6569-0_11
215
216
A. Husseinzadeh Kashan et al.
1 Introduction Optimum design of engineering structures is more vital now than ever before. The growing demand for economically designed engineering structures and rising cost of building materials have forced structural engineers to optimize structures using various optimization techniques. In dealing with a structural optimization problem, two types of optimization techniques are available: (i) traditional gradient-based optimization techniques, (ii) meta-heuristic search techniques. The traditional gradient-based optimization methods are not very much suitable for real-world structural design problems and their efficiency depends on the differentiability and continuity of the objective function, as well as the selection of adequate initial solution vector. For some of the practical structural optimization problems, in which the cross sections are assigned from a given list of available profiles, the implementation of traditional methods is not an easy task. As an attractive alternative to the traditional methods, meta-heuristic optimization techniques, e.g., Genetic
Data Loading...
