Aiming at the problem of autonomous coordination of networked UAV swarm and the advantages and disadvantages of market auction method, the idea of “computation-for-communication” and its corresponding distributed task scheduling method were proposed. By analyzing explicit and implicit conflicting tasks, a set of task-related agents was established. A local optimization method based on task suppression was proposed to resolve some task conflicts in advance, so as to reduce the number of algorithm iterations. An agent position inference method based on historical bidding information was designed to provide necessary information input for local optimization. Monte Carlo simulation experiments were carried out based on the networking simulation platform and the swarm rescue scenario. The results show that compared with the representative consensus-based bundle algorithm and performance impact algorithm in the market auction method, the proposed method can obtain fewer iterations, shorter convergence time and better scheduling performance.

Multi-Target Tracking Guidance (MTTG) in unknown environments has great potential values in applications for Unmanned Aerial Vehicle (UAV) swarms. Although Multi-Agent Deep Reinforcement Learning (MADRL) is a promising technique for learning cooperation, most of the existing methods cannot scale well to decentralized UAV swarms due to their computational complexity or global information requirement. This paper proposes a decentralized MADRL method using the maximum reciprocal reward to learn cooperative tracking policies for UAV swarms. This method reshapes each UAV's reward with a regularization term that is defined as the dot product of the reward vector of all neighbor UAVs and the corresponding dependency vector between the UAV and the neighbors. And the dependence between UAVs can be directly captured by the Pointwise Mutual Information (PMI) neural network without complicated aggregation statistics. Then, the experience sharing Reciprocal Reward Multi-Agent Actor-Critic (MAAC-R) algorithm is proposed to learn the cooperative sharing policy for all homogeneous UAVs. Experiments demonstrate that the proposed algorithm can improve the UAVs’ cooperation more effectively than the baseline algorithms, and can stimulate a rich form of cooperative tracking behaviors of UAV swarms. Besides, the learned policy can better scale to other scenarios with more UAVs and targets.