Reciprocal Altruism-based Path Planning Optimization for Multi-Agents

This paper investigates solutions for the fundamental yet challenging problem of path planning of autonomous multi-agents. The novelty of the proposed algorithm, reciprocal altruism-based particle swarm optimization (RAPSO), lies in the introduction of information sharing among the agents. The RAPSO utilizes kinship relatedness among the agents during the optimization process to reciprocate the significant data. The concept of reciprocation is introduced to ensure that all agents remain in close contact through information exchange. The amount of exchange depends upon their physical location in the search space and their associated health indicator. Agents are classified as donors, recipients, or un-active concerning their health indicator and their positions. The ability of RAPSO to keep all agents closer to local optima through reciprocal altruism is evaluated for path planning optimization problem. Simulation results show that the RAPSO is very competitive when compared with the canonical PSO. The results of the generalized simulation scenario also prove its potential in solving path planning problems in robotics.