SBFSS: Simplified Binary Fish School Search

Swarm Intelligence has been extensively used to tackle binary and continuous optimization problems. Although there are several optimization algorithms in this field, the discovery of new complex problems continues to challenge and motivate researchers to create or enhance swarm-based algorithms. We present a new version of the Fish School Search algorithm, designed for binary optimization problems, named Simplified Binary Fish School Search (SBFSS). The reduction in the number of parameters required to run the algorithm improved not only the computational cost but also the accuracy. We assessed the performance of the SBFSS in the 0/1 Knapsack problem with 50, 100, 500 and 1000 dimensions. We also compared it with two other binary FSS versions and other well-known binary optimization algorithms: BPSO, BGA, and BABC. The results indicate that our proposal was able to achieve satisfactory results even when the number of dimensions of the problem increases, which makes the SBFSS a good candidate for problems with high dimensionality in the decision space.