The new suggested binary rat swarm algorithm

Abstract

Recently, researchers have started to be interested in population-based swarm-based algorithms in optimization due to their simple structure, high optimization performance, and ease of adaptation. Although swarm-based algorithms solve continuous optimization problems, they can also be used to solve binary optimization problems. In continuous optimization, the search space variables try to approach the optimal value, while in discrete optimization, the search space variables are fixed and expressed with real values. In binary optimization, the decision variables take values of 0 and 1 and are basically in the discrete optimization class. In this paper, the proposed Rat Swarm Algorithm (RSA) to solve continuous optimization problems is examined. RSA is an algorithm based on swarm intelligence. RSA was developed by imitating the chasing and attacking behaviors of rats. In this study, the original RSA was updated again to solve binary optimization problems and Binary RSA (BinRSA) was proposed. In BinRSA, four U-shaped and four T-shaped transfer functions are used while converting the continuous search field values to binary values. Thus, eight variants of BinRSA were obtained. These are named as BinRSA1, BinRSA2, BinRSA3, BinRSA4, BinRSA5, BinRSA6, BinRSA7 and BinRSA8. Among these variants, the most successful variant of BinRSA was determined as BinRSA6. Then the BinRSA6 variant was developed by adding crossover and mutation operators and was named GBinRSA. GBinRSA's performance has been tested in knapsack problems. In addition, the success of GBinRSA was compared with different heuristic algorithms selected from the literature. According to the results obtained, it has been seen that the solution quality of the proposed algorithm is effective and comparable. The results showed that GBinRSA is a preferred heuristic for binary optimization problems.