Research On Improvement Of Fruit Fly Optimization Algorithm

Fruit Fly Optimization Algorithm(FOA) is a new kind of global optimization algorithms swarm intelligence. The algorithm derived from the drosophila foraging behavior simulation, it has obvious advantages, such as simple principle, a few parameters to adjust, the code is easy to achieve, and it also has a high convergence speed, many scholars pay attention on it. However, the current research and application of FOA still in its infancy, there are still some deficiencies, such as premature convergence, the last evolutionary optimization are slowly.To improve the shortcomings of FOA, the FOA algorithm theory and methods conducted are also deeply studied in this paper, this paper proposes two technical improvements to improve and optimize the original optimization algorithm. The main contents are as follows:(1) This paper describes several common intelligent algorithm theory and summarizes the group’s main intelligence operation process, then describes the basic principles of FOA, such as the main code and algorithm processes of it, make an analysis of the research status and application and the current direction of the algorithm, at last, the paper shows the advantages and disadvantages of the algorithm.(2) Draw cross-factor and metropolis rule into FOA, through the optimization algorithm to find the best individual in evolutionary process, by using the cross-operating to update individual position and to update globally optimal individual, based on incremental fitness guidelines as simulated annealing criteria used to select members of the group into the next iteration, continuing to search the global optimal solution by analyzing and comparing the experimental, the results shows the improved algorithm outperforms the original optimization algorithm.(3) Then the paper proposes a new algorithm, after the above, although the improved algorithm improves performance, but the depth of the improved search algorithm is not enough, so this paper also proposes an optimization algorithm based on cellular automata. On the one hand, when looking for the best individual, the introduction of cellular automaton evolution rules allow cellular automata to select individuals into the next generation, by comparing individuals globally, to avoid falling into local optimum, on the other hand,the introduction of random disturbances characteristic of fruit fly position disturbance, increased the diversity of the population of the latter part of the search, the test results of the common function analysis shows that the improved algorithm has good convergence effect and high performance.This paper proposes the two kinds of improved algorithm have good features, through the experiment on the standard function optimization problems, it has achieved good effect with the design and the algorithm theory, this research proposed two improved algorithm to a certain extent, to avoid the fruit fly optimization algorithm is easily trapped into local optimal and precocity, at the same time, it provides a theoretical analysis based on the algorithm for the future study.