Key Node Identification Technology Based On Target System Effectiveness Analysis

Intelligent systematized warfare has become the main form of modern battlefield.In the context of scarce resources,how to accurately and quickly identify the key nodes in the system for attack or protection has become an important task that both sides must face.The field of key node recognition has achieved certain results in the complex network science,but the effect of the method will be influenced by the topology discovery algorithm,which has certain limitations and relatively single perspective.In response to this problem,this thesis proposes two methods: firstly,a universal key node recognition method based on target system performance analysis;secondly,the key nodes are comprehensively evaluated by integrating the system efficiency with specific tasks.The main work and innovation of this thesis are as follows.:1.Based on the effectiveness of the target system,a general key node identification method is proposed in this paper.First,the clustering algorithm is used to divide the system through the spatial distribution of all target nodes,and then an efficiency evaluation method with the system as the unit is proposed,and the system with the highest efficiency value is obtained through multi-attribute decision-making,and finally a singleobjective The importance evaluation method of the method also obtains the node with the highest importance through multi-attribute decision-making,as the key node.In this method,system division and system effectiveness evaluation methods are proposed for the popular intelligent system confrontation,and the key node identification based on system effectiveness is realized.At the same time,on the basis of complex network science,the topological characteristics and capabilities Attributes are integrated to comprehensively and accurately evaluate the importance of nodes to meet the needs of actual application scenarios.2.In this thesis,a key node identification method that combines effectiveness analysis and path planning tasks is proposed.Starting from the results of the target system effectiveness analysis,combined with the path planning task scenario,the influence of interference nodes on the effectiveness of the system and the performance of path planning is evaluated.Firstly,it is necessary to model the threat sources of the planning scenario and construct a threat probability map.Secondly,two defects of the traditional artificial potential field path planning algorithm,including unreachable targets and local minima,are addressed.An artificial potential field optimization method based on deep reinforcement learning is proposed as a path planning algorithm.Through simulation experiments,this algorithm not only solves the above defects but also effectively improves the convergence speed of the deep reinforcement learning framework.Then,a Double-Dueling Double Deep Q Network(Double-D3QN)is proposed for key node identification,which enables the path planning and key node identification processes to be performed simultaneously.Finally,by continuously modifying the reward function coefficient factors,different key node identification results can be obtained,and the practical application significance and value of each can be analyzed to obtain the optimal solution under the best coefficient.3.A visualization platform for identifying key nodes in the target system effectiveness has been built.Through simulation cases,the system architecture,functions,operation procedures,and other aspects of the visualization platform were demonstrated,and multiple user interactive interface functions were demonstrated to show the key node identification process.