Bacterial foraging optimization based on improved chemotaxis process and novel swarming strategy

摘要

Bacterial foraging optimization (BFO), a biological-inspired optimization algorithm, has been applied in various fields, such as complex function optimization, robot path planning. However, there still exist several insufficiencies in BFO algorithm due to the fixed chemotaxis step-size, the less-efficient search direction for tumbling and the swarming strategy with lower convergence rate. In order to deal with these issues, based on the Lévy flight step-size and particle swarm optimization (PSO) operator, this paper proposes the improved BFO algorithm (LPBFO). To reduce the mutual interference among different dimensions, each bacterium selects one dimension for tumbling randomly during the chemotactic process in LPBFO. The step-size of each bacterium is determined by the stochastic flight lengths of the improved Lévy flight which can generate small step-size with high frequency and big step-size occasionally; moreover, the stochastic step-size is also reduced adaptively based on the evolutionary generations, which makes the bacteria transform from global search to local search. Furthermore, inspired by the social information term in PSO, this paper employed the global best solution to improve the swarming performance. Several experiments on benchmark functions are carried out with the purpose of evaluating the performance of the proposed method. Experimental results show that the proposed algorithm achieves noticeable improvement compared with other competitive algorithms.