Enhancing the accuracy of firefly algorithm by using the reproduction mechanism

Abstract

Up to now, different algorithms based on evolutionary processes have been provided, which each has its own strengths and weaknesses. A physical phenomenon has many and complex variables and assumptions, and a number of these assumptions should be ignored in order to have a simpler modeling. Generally speaking, full modeling of a natural process through evolutionary algorithms is impossible, therefore, researchers consider special requirements for the natural and biological modeling processes, and they begin modeling, considering the assumptions limits. Evolutionary algorithms for modeling, need simple conditions and assumptions, and this simplification can take away a modeling from its actual state. Firefly algorithm is one of the evolutionary algorithms that is presented on the basis of social behavior and optical pulses between the insects. Lack of sufficient breeding of fireflies is one of the drawbacks that is not considered in the algorithm, and this factor has a negative impact on the convergence of the algorithm. Unlike the firefly algorithm, the weed algorithm, is a reproduction mechanism among plants, and the number of each plant's offspring are considered in accordance with its fitness, and in this study, this mechanism is used in firefly algorithm. The results show that the proposed algorithm has better accuracy and convergence, at least in comparison with firefly and bat algorithms.