|
Abstract
|
In recent years, various optimization algorithms have been used to solve engineering problems and provide optimal solutions for them. Unlike classical algorithms, new optimization techniques are implemented to evaluate complex optimization problems. Most meta-heuristic methods have been obtained by observing and studying the behavior of animals living in nature. Exploration and exploitation capabilities are two main features that make meta-heuristic algorithms powerful for finding global optima. In fact, the exploration phase of an algorithm helps to search the feasible space for new solutions and find the global optimum. While, in the exploitation phase, to select the best solution, the algorithm searches the neighborhood of the current best solution. Therefore, the balance of these two phases is very important and an efficient algorithm provides a good exploration phase at the beginning of the process and a good exploitation ability at the final stages.
However, like other methods, FHO has disadvantages such as difficulties in finding the best global minimum or getting trapped in local minima in some specific problems. For this purpose, in this research we want to improve the performance of the FHO algorithm with two important changes: linear population reduction (LPSR) and search space reduction (SSR). In LPSR, the population size is continuously reduced. In this way, the population size in the initial iterations of the algorithm is considered large enough to explore large areas of the search space. As the algorithm advances, the size of the population is gradually reduced to increase the speed of convergence. In SSR, the domain of variables is constantly reduced to limit the search space, and so the possibility of finding optimal solutions is increased.
|