A Modified Bacterial Foraging Optimization Algorithm for Global Optimization
To improve the optimization ability of Bacterial Foraging Optimization (BFO), A Modified Bacterial Foraging Optimization algorithm is proposed, which we named MBFO. In MBFO, tumble directions of bacteria are guided by the global best of the population to
- PDF / 206,679 Bytes
- 9 Pages / 439.37 x 666.142 pts Page_size
- 84 Downloads / 286 Views
Abstract. To improve the optimization ability of Bacterial Foraging Optimization (BFO), A Modified Bacterial Foraging Optimization algorithm is proposed, which we named MBFO. In MBFO, tumble directions of bacteria are guided by the global best of the population to make bacteria search the optimization area more effectively. Then, chemotactic step size of each bacterium will change dynamically to adapt with the environment. Meanwhile, in reproduction loop, all individuals will be chosen with a probability. To test the global optimization ability of MBFO, we tested it on ten classic benchmark functions. Original BFO, PSO and GA are used for comparison. Experiment results show that MBFO algorithm has significant improvements compared with original BFO and it performs best on most functions among the compared algorithms. Keywords: Modified bacterial foraging optimization Chemotaxis Adaptive strategies
1 Introduction Swarm intelligence algorithms are a kind of powerful optimization algorithms which inspired by the social behaviors of animal swarms in nature. By interaction and cooperation of the individuals who have only simple behaviors, complex collective intelligence could emerge on the level of swarm. Recent years, many swarm intelligence algorithms have been proposed, such as Ant Colony Optimization (ACO) [1], Particle Swarm Optimization (PSO) [2] and Bacterial Foraging Optimization (BFO) [3] et al. BFO is proposed by Passino in 2002. It is inspired by the foraging and chemotactic behaviors of E. coli bacteria. Recently, BFO algorithm and its variants have been used for many numerical optimization [4] and engineering optimization problems [5, 6]. However, original BFO algorithm has some defects. First, the tumble angles are generated randomly. Useful information has not been fully utilized. Second, the chemotactic step size in BFO is a constant. It will make the population hard to converge to the optimal point at the end stage. In this paper, we proposed a Modified Bacterial Foraging Optimization (MBFO) algorithm. Several adaptive strategies are used in MBFO to improve its optimization ability. First, the tumble angles are generated towards the direction of global best, which enhances its optimization ability and convergence speed. Second, adaptive step size is employed. Chemotactic step size of © Springer International Publishing Switzerland 2016 D.-S. Huang et al. (Eds.): ICIC 2016, Part I, LNCS 9771, pp. 627–635, 2016. DOI: 10.1007/978-3-319-42291-6_62
628
X. Yan et al.
each bacterium will increase or decrease depending on whether its position becomes better or not. This could make the bacteria use different search strategies in different stages. Third, in reproduction loop, all individuals will be chosen with a probability related with their fitness. The rest of the paper is organized as followed. In Sect. 2, the original BFO algorithm is introduced. In Sect. 3, the MBFO algorithm is proposed and its strategies are described in detail. In Sect. 4, we test the proposed MBFO and other three algorithms on ten be
Data Loading...
