Synchronous Bacterial Foraging optimization for multimodal and high dimensional functions

In this paper, the authors propose a new evolutionary optimization i.e. synchronous bacterial foraging optimization (SBFO). The SBFO can be used for optimization of multimodal and high dimensional functions. It also enhances computational throughput and global search capability. The convergence of original BFO to the optimum value is very slow and its performance is also heavily affected with increased number of dimensions as well as for multimodal functions. In proposed technique, Jbest (best optimum value) is updated synchronously after fitness function evaluations of all bacteria. The fitness function evaluations of all bacteria can be done serially or parallel. In SBFO, the optimization follows Chemotaxis, swimming, tumbling, and reproduction steps to reach optimum value until computational limitations are exceeded. The local search is performed through chemotactic, swimming and tumbling steps. The step size equation used in SBFO explores the search space globally. The proposed technique is validated on three benchmark functions i.e. rosenbrock, rastrigin and griewank. The results of the suggested technique is compared with different existiong methods, which includes multimodal and high dimensional functions. The SBFO giving promising results interms of accuracy and drastically reducing the computional time. The proposed method is new and computationally efficient and accurate.