Lévy walk-based search strategy: Application to destructive foraging

Abstract

In this paper a Search strategy based on Lévy Walk is proposed. Lévy Walk increases the diversity of solutions, and constitutes good strategies to move away from local to global search. The amount of exploration and exploitation and the fine balance between them determine the efficiency of search algorithm. The highly diffusive behavior of the original Lévy Walk algorithm results in more global search. Thus, in the proposed algorithm, the time spent in local search is increased according to the fluctuation of the searched region. Moreover, a case study on destructive foraging is presented in this paper, with the aim to apply it to several swarm robotics problems. In order to present simulations in a finite two dimensional landscape with a limited number of clustered and scattered targets, the ArGOS simulator has been used.