A Relative Position-based Bacterial Foraging Optimization for Numerical Optimization

Bacterial foraging optimization (BFO) algorithm has been widely applied to various optimization problems. However, BFO often suffers from premature convergence and lacking of population information exchanging. To overcome these shortcomings, a relative position-based bacterial foraging optimization (RPBFO) is proposed. The three-layer circulation structure is replaced by a single-layer circulation structure in this algorithm. The relative position-based updating method is used to replace the absolute position-based updating method in the chemotactic operation. The reproduction step of BFO is eliminated. And the escape strategy is employed in the elimination-dispersal operation. Then the optimization results of the RPBFO algorithm are tested on 11 benchmark functions. The results show that the optimization ability of the RPBFO algorithm is significantly better than the original BFO and GA algorithms. On most benchmark functions, it also shows a better performance in convergence speed and accuracy than the PSO algorithm.