Future unmanned battles desperately require intelli-gent combat policies,and multi-agent reinforcement learning offers a promising solution.However,due to the complexity of combat operations and large size of the comb...Future unmanned battles desperately require intelli-gent combat policies,and multi-agent reinforcement learning offers a promising solution.However,due to the complexity of combat operations and large size of the combat group,this task suffers from credit assignment problem more than other rein-forcement learning tasks.This study uses reward shaping to relieve the credit assignment problem and improve policy train-ing for the new generation of large-scale unmanned combat operations.We first prove that multiple reward shaping func-tions would not change the Nash Equilibrium in stochastic games,providing theoretical support for their use.According to the characteristics of combat operations,we propose tactical reward shaping(TRS)that comprises maneuver shaping advice and threat assessment-based attack shaping advice.Then,we investigate the effects of different types and combinations of shaping advice on combat policies through experiments.The results show that TRS improves both the efficiency and attack accuracy of combat policies,with the combination of maneuver reward shaping advice and ally-focused attack shaping advice achieving the best performance compared with that of the base-line strategy.展开更多
To solve the path following control problem for unmanned surface vehicles(USVs),a control method based on deep reinforcement learning(DRL)with long short-term memory(LSTM)networks is proposed.A distributed proximal po...To solve the path following control problem for unmanned surface vehicles(USVs),a control method based on deep reinforcement learning(DRL)with long short-term memory(LSTM)networks is proposed.A distributed proximal policy opti-mization(DPPO)algorithm,which is a modified actor-critic-based type of reinforcement learning algorithm,is adapted to improve the controller performance in repeated trials.The LSTM network structure is introduced to solve the strong temporal cor-relation USV control problem.In addition,a specially designed path dataset,including straight and curved paths,is established to simulate various sailing scenarios so that the reinforcement learning controller can obtain as much handling experience as possible.Extensive numerical simulation results demonstrate that the proposed method has better control performance under missions involving complex maneuvers than trained with limited scenarios and can potentially be applied in practice.展开更多
The dwell scheduling problem for a multifunctional radar system is led to the formation of corresponding optimiza-tion problem.In order to solve the resulting optimization prob-lem,the dwell scheduling process in a sc...The dwell scheduling problem for a multifunctional radar system is led to the formation of corresponding optimiza-tion problem.In order to solve the resulting optimization prob-lem,the dwell scheduling process in a scheduling interval(SI)is formulated as a Markov decision process(MDP),where the state,action,and reward are specified for this dwell scheduling problem.Specially,the action is defined as scheduling the task on the left side,right side or in the middle of the radar idle time-line,which reduces the action space effectively and accelerates the convergence of the training.Through the above process,a model-free reinforcement learning framework is established.Then,an adaptive dwell scheduling method based on Q-learn-ing is proposed,where the converged Q value table after train-ing is utilized to instruct the scheduling process.Simulation results demonstrate that compared with existing dwell schedul-ing algorithms,the proposed one can achieve better scheduling performance considering the urgency criterion,the importance criterion and the desired execution time criterion comprehen-sively.The average running time shows the proposed algorithm has real-time performance.展开更多
文摘Future unmanned battles desperately require intelli-gent combat policies,and multi-agent reinforcement learning offers a promising solution.However,due to the complexity of combat operations and large size of the combat group,this task suffers from credit assignment problem more than other rein-forcement learning tasks.This study uses reward shaping to relieve the credit assignment problem and improve policy train-ing for the new generation of large-scale unmanned combat operations.We first prove that multiple reward shaping func-tions would not change the Nash Equilibrium in stochastic games,providing theoretical support for their use.According to the characteristics of combat operations,we propose tactical reward shaping(TRS)that comprises maneuver shaping advice and threat assessment-based attack shaping advice.Then,we investigate the effects of different types and combinations of shaping advice on combat policies through experiments.The results show that TRS improves both the efficiency and attack accuracy of combat policies,with the combination of maneuver reward shaping advice and ally-focused attack shaping advice achieving the best performance compared with that of the base-line strategy.
基金supported by the National Natural Science Foundation(61601491)the Natural Science Foundation of Hubei Province(2018CFC865)the China Postdoctoral Science Foundation Funded Project(2016T45686).
文摘To solve the path following control problem for unmanned surface vehicles(USVs),a control method based on deep reinforcement learning(DRL)with long short-term memory(LSTM)networks is proposed.A distributed proximal policy opti-mization(DPPO)algorithm,which is a modified actor-critic-based type of reinforcement learning algorithm,is adapted to improve the controller performance in repeated trials.The LSTM network structure is introduced to solve the strong temporal cor-relation USV control problem.In addition,a specially designed path dataset,including straight and curved paths,is established to simulate various sailing scenarios so that the reinforcement learning controller can obtain as much handling experience as possible.Extensive numerical simulation results demonstrate that the proposed method has better control performance under missions involving complex maneuvers than trained with limited scenarios and can potentially be applied in practice.
基金supported by the National Natural Science Foundation of China(6177109562031007).
文摘The dwell scheduling problem for a multifunctional radar system is led to the formation of corresponding optimiza-tion problem.In order to solve the resulting optimization prob-lem,the dwell scheduling process in a scheduling interval(SI)is formulated as a Markov decision process(MDP),where the state,action,and reward are specified for this dwell scheduling problem.Specially,the action is defined as scheduling the task on the left side,right side or in the middle of the radar idle time-line,which reduces the action space effectively and accelerates the convergence of the training.Through the above process,a model-free reinforcement learning framework is established.Then,an adaptive dwell scheduling method based on Q-learn-ing is proposed,where the converged Q value table after train-ing is utilized to instruct the scheduling process.Simulation results demonstrate that compared with existing dwell schedul-ing algorithms,the proposed one can achieve better scheduling performance considering the urgency criterion,the importance criterion and the desired execution time criterion comprehen-sively.The average running time shows the proposed algorithm has real-time performance.
