Game theoretic self-organization in multi-satellite distributed task allocation

Abstract This paper addresses self-organized task allocation in multi-satellite systems by formulating it as a potential game and proposing a distributed task allocation algorithm (DT2A) via learning in games. We show that each Nash equilibrium of the game constitutes a task cover that maximizes the number of executed tasks. Moreover, we also prove that the proposed DT2A converges with probability 1 to near optimal assignments in finite time. Numerical experiments demonstrate the algorithm robustness against incidental disturbances and reveal the positive effect of the memory length on solution efficiency refinement. Comparison experiments in both small-scale and large-scale scenarios highlight the superiority of the presented methodology over existing typical methods that are distributed or centralized.