Design and implementation of generalized frameworks for population-based metaheuristics

The success of population-based meta-heuristic algorithms in problem solving has increased the interest in these algorithms. Artificial Bee Colony and Particle Swarm Optimization algorithms are two of population-based meta-heuristic algorithms. Researchers have introduced new variants that have been developed with minor changes in the structure of these algorithms. These newly proposed variants generally depend on the set of problems used by the researchers and the experience of the researchers. This leads to insufficient algorithms for the types of problems they do not encounter. In order to overcome this problem, generalized algorithm frameworks for ABC and PSO algorithms are proposed in this thesis. A generalized algorithm framework called ABC-X, which incorporates the properties of various ABC variants and enables it to generate a problem-specific ABC algorithm with an automatic configuration tool, has been proposed. The success of this algorithm was tested with a set of criteria including fifty problems. It has also been used to solve the problem of power dispatch problem with a real-world problem. Another study is to propose an algorithm called Template PSO for the PSO algorithm. The performance of the algorithm was analyzed using the CEC'17 criterion set and SOCO high dimension criterion set. Finally, a self-adaptive search equation-based ABC (SSEABC) algorithm has been proposed, considering that the component that has the greatest effect on the performance of the algorithms is the search equation. The performance of SSEABC has been tested with CEC'17 and SOCO benchmark sets. In addition, SSEABC has been used to solve an engineering problem, the filter identification problem. All three proposed algorithms have achieved better and more competitive results than the algorithms they compare.