Energy Efficient Random Search in Euclidean Space using Lévy Flight

Agents operating in unknown environments commonly rely on fundamental random search strategies to locate targets without prior knowledge of their surroundings. This research is motivated by the need to identify the most efficient movement strategy in autonomous robotics applications. Previous works often focus on specific aspects, lacking a comparative analysis of the random walk movements. Here, we investigate the comparative performance of agents utilizing random walk movements such as Random Waypoint, Brownian Motion, and Lévy Flight. The simulations were performed using the Opportunistic Network Environment (ONE) simulator. Each movement was tested under similar conditions, where targets were spread using a spatial Poisson random distribution and a spatial clustered distribution. We evaluated their performance based on the coverage over time for each movement under various target distributions and analyzed the best parameter for Lévy Flight on the defined target distributions. Results offer practical insights for agent designs during random search and validate whether Lévy Flight demonstrates superior performance as suggested in previous studies.