Novel approaches for performance improvement of fruit fly optimization algorithm

The solution of optimization problems has become a subject of interest in recent years. Many meta-heuristic methods have been developed to solve these problems. Meta-heuristic methods do not guarantee an optimum solution. Meta-heuristic methods aim to find acceptable solutions in a reasonable time. Meta-heuristic methods are often not problem specific. Meta-heuristic approaches are general purpose, flexible and adaptable to problems. Meta-heuristic methods have been used extensively in the solution of optimization problems in recent years. Fruit Fly Optimization Algorithm (FOA) is a meta-heuristic algorithm introduced in 2011. Inspired by the fruit fly's foraging behavior. FOA is a simple structure, intuitive approach that is easy to understand and program, easy to adapt to optimization problems, with few design parameters. Although it has such advantages, it also has disadvantages. It has fast to the local optimum. The decision function is always positive. The update strategy is small because it is in the [-1, 1] range. In this study, it is aimed to eliminate the disadvantages of the FOA, to improve the performance of the algorithm and to provide better quality results. For this purpose, three different improves have been made in FOA. In the first development, sign parameters were added to the FOA and called SFOA. In the second, the decision-making strategy of the FOA was made in two stages and called saFOA. In the third, two different versions have been developed in which FOA is considered in the worst-case solutions during the search and is named pFOA_v1 and pFOA_v2. The performance of the newly proposed FOA versions was tested in 21 well-known numerical benchmark functions. The experimental results are compared with the meta-heuristic algorithms which are well known in the literature. Experimental results show that the proposed FOA versions produce comparable, successful and competitive results for continuous optimization problems. Keywords: Continuous optimization, fruit fly optimization algorithm, heuristic algorithms, Metaheuristic, swarm intelligence