Intelligent Decision Support Technology Driven By Real-time Battlefield Situation

The "state" in the battlefield situation refers to the real-time state of all information on the current battlefield,which includes the attributes and relationships of each entity,and the "state" refers to the development trend of the entire battlefield.The battlefield situation reflects the target positioning of the enemy,us,and neighbors.Comprehensive information about the state of action and the battlefield environment.In combat command,it is necessary to quickly analyze the situation to form an accurate understanding of the battlefield situation,and this is the prerequisite for combat command.In the context of current informationized operations,commanders need to quickly analyze the real-time battlefield data connected to the command post,and then form a situational understanding of the current battlefield environment,and on this basis,draw a decision-making plan for the next moment.With the development of information technology,the combat environment of the modern battlefield has become more complex and changeable,and situation information is diverse and huge in quantity.In the face of future wars,how to quickly analyze the situation,form an accurate situation understanding,and efficiently and reasonably arrive at a decision-making plan has become the key to winning the war.As the artificial intelligence technology dominated by deep learning and reinforcement learning matures,decision-making systems based on intelligence and supplemented by humans are becoming the focus of research.This paper mainly studies the intelligent decision support technology driven by the real-time situation of the battlefield,and proposes the end-to-end framework construction technology from the acquisition of situation data to the intelligent scheme.First,for the understanding of the battlefield situation,a data-driven situation knowledge system construction model is proposed to generate the key situation characteristics that affect the situation change.Secondly,intent prediction and group recognition are carried out according to the key situation characteristics,and the result of intelligent decision-making based on the current situation understanding is output.Then,on the basis of situation understanding,a WPK-BCQ(war-knowledge-prior Batch-constrained deep Q-learning)deep chemical model based on prior knowledge is proposed to effectively solve the initial cold start problem faced by the initial stage of intelligent decision-making.Finally,an end-to-end intelligent decision-making support framework is constructed.Under the developed intelligent decision-making system,real-time battlefield data is analyzed to obtain an understanding of the battlefield situation,and then parallel deductions are performed to generate intelligent decisions.