基于蒙特卡洛树搜索的全域指挥控制协同

All-Domain Command and Control Coordination Based on Monte Carlo Tree Search

面向未来全域联合作战概念下的智能化指挥控制协同需求,重点研究针对具体作战任务的作战单元协同,通过利用复杂网络建模方法,构建指挥控制协同网络模型,并设计3种指挥控制协同模式.在此基础上,借鉴强化学习思想,提出基于蒙特卡洛树搜索的指挥控制协同算法,来实现指挥控制协同决策过程的最佳策略选择,最终生成面向具体任务的最优作战簇,降低指挥控制协同网络的搜索代价,提高全域指挥控制协同的智能化水平.以联合反舰作战为背景,通过仿真实验,构建不同指挥控制协同模式下的指挥控制协同网络,并验证基于蒙特卡洛树搜索的指挥控制协同算法的有效性.

In order to meet the requirements of intelligent command and control coordination under the concept of future all-domain joint operations,the paper focuses on the combat unit coordination for specific combat tasks.By using the complex network modeling method,a command and control coordination network model is constructed,and three command and control coordination modes are designed.On this basis,learning from the idea of reinforcement learning,a command and control coordination algorithm based on Monte Carlo tree search is proposed to realize the optimal strategy selection in the decision process of command and control coordination,and finally generate the optimal combat cluster for specific tasks,reduce the search cost of command and control coordination networks,and improve the intelligent level of all-domain command and control coordination.Based on the background of joint anti-ship operations,command and control coordination networks under different command and control coordination modes are constructed through simulation experiments,and the effectiveness of the command and control coordination algorithm based on Monte Carlo tree search is verified.