Complexity Modeling And Optimization Technology On Information Network Systems

With the development of information networks in recent years,many new types of networks appear,such as satellite networks,in-vehicle networks and novel combat networks.These networks are generally hierarchical,heterogeneous and complex.For depicting characteristics of these networks,the concept of “network system” is used and becomes a hot research topic in the study of network science.Information network system is defined as a large complex system with the information as its main object,the network as its center,and the system as its supporting framework.Current studies focus on the analysis of union and systematics of combat systems,emphasizing the transformation from information advantages to decision and behavior advantages.However,there is lack of systematically analysis of combat systems using network science-based methods.In this thesis,we analyze and study an information network system in a systemic way,including complexity modeling,model optimization,routing optimization and system evaluation.Following these steps is because 1)modeling and optimization of an information network system is the basis for studying a combat system;2)the routing strategy in a system is directly related to the time-efficiency of information transmission;and 3)there is lack of unified metrics to evaluate the modeling of information network systems.First,a combat network system is modeled based on the OODA(Observe-Orient-Decide-Act)theory.Second,we propose a closeness centrality-based optimization method for assigning decision nodes in an OODA combat system.We also propose a k-shortest paths-based(KSP)routing strategy for recovering information transmission under network attacking.Finally,the network system is evaluated by comparing the influence of removing nodes or not on each combat unit to identify key nodes in the OODA network system.A procedure in an OODA combat network system from the information detection to the command action is defined as a mission link.The invulnerability and robustness of the network system are evaluated by mission link efficiency and mission link entropy.The advantages of these metrics are validated by different attacking simulation and also by comparison with other metrics.We implemented the modeling of the OODA network system and the routing algorithm by building an OPNET simulation platform and a MATLAB data analysis platform.KSP routing recovery strategy was validated to have better reliability and time-efficiency with respect to information transmission under network attacking.We optimized the assignment of decision nodes using a greedy search algorithm and a fast Top-k closeness computation method,improving the ability of decision nodes to control the entire network system and the time-efficiency of information transmission.Compared to traditional metrics,mission link efficiency was demonstrated to be more suitable for evaluating the invulnerability of the OODA combat system.Moreover,for evaluation of OODA robustness,mission link efficiency is more sensitive compared to degree distribution entropy,connectivity coefficient and betweenness distribution entropy.