Improved Social Spider Algorithm via Differential Evolution

Abstract

In this study, Social Spider Algorithm (SSA) is improved by combining it with the mutation strategies of Differential Evolution (DE) algorithm. Without altering the main operation of the SSA, we utilized the DE mutation strategies to diversify the individuals in the population. We selected two different mutation strategies to obtain two different hybrids and evaluated them. These mutation strategies are the most basic strategy and currently the best strategy of the DE algorithm proposed in the literature. Two different hybrids are evaluated via several tests performed on five different real-world problems. According to our results, the SSA is improved successfully. The hybrid utilizing the most basic strategy of the DE algorithm has given better results than the currently best strategy with an improvement 15.56% on average in all tests. It has been found that the currently best strategy resulted in worse performance probably since it provided more diversity on the population.