Bat Algorithm Based On Centroid Strategy

Bat algorithm is a novel population-based swarm intelligent algorithm inspired by microbats echolocation behavior. Due to the simple concepts, easy implementation, it has been widely applied to many areas successfully. In this thesis, the centroid strategy is employed to make bat algorithm more effective.The standard bat algorithm is not a global convergent algorithm. To overcome this shortcoming, a new variant, called bat optimization algorithm with centroid strategy is designed. In this algorithm, the arithmetic centroid strategy is employed to improve the exploration capability. Furthermore, it is proved to be convergent with probability one. To test the performance, CEC2013 benchmark test suit is used to test, and compared with other four variants of bat algorithm. Simulation results show it is effective.To investigate the influence of centroid strategy, geometric centroid strategy and harmonic centroid strategy are also designed. However, all neighbors with different performance are the same influence for these three centroid strategy, it is confused with the natural phenomenon. To avoid such problem, the weighted arithmetic centroid strategy, weighted geometric centroid strategy and weighted harmonic centroid strategy are also designed. The weights are self-adjusted and proportional to the corresponding fitness values. The weight is larger with the better performance, and vice versa. Simulation results show bat algorithm with weighted harmonic centroid strategy achieves the best performance when compared with other five different centroid strategies, it also outperforms other six evolutionary algorithms.Finally, the bat algorithm with centroid strategy is applied to solve wireless sensor network LEACH protocol optimization problem. Our algorithm can further optimize the cluster head position refer to distance and remaining energy in the cluster head selection algorithm. Simulation results show proposed protocol is effective and prolongs the network lifetime.