Dynamic Effectiveness Evaluation And Decision-making Of Anti-ship System Based On Probabilistic Finite-state Machine Sets

Anti-ship combat effectiveness is an important index to evaluate the capability of anti-ship missiles.To establish a system-level evaluation model for the anti-ship combat process for evaluating the real-time dynamic effectiveness changes during system con-frontation,and to support decision-making for combat capability maximization of the sys-tem based on the efficiency evaluation results,this thesis proposes a dynamic efficiency evaluation method and an auxiliary decision-making algorithm based on probabilistic fi-nite state machine sets(PFMS).Most of the existing effectiveness evaluation methods are relatively static concepts with relatively simple models.Thus,only the possibility of the system completing a given task is evaluated,while dynamic changes of the system’s internal performance param-eters are not reflected.To show the dynamic impact of combat condition changes on system effectiveness during confrontations,existing effectiveness evaluation methods are first studied.By comparing the advantages and disadvantages of the existing methods,traditional probabilistic finite state machines are modified,and a theoretical system of PFMS is proposed,including the definition,properties,and automatic solving methods.With this theory,interactions between various systems and subsystems in the confronta-tion system can be more intuitively reflected.The given system model can be established without relying on other tools,more performance-influencing factors can be described,and each subsystem can be more conveniently revised.This method greatly reduces the space and time complexity of the existing PFM algorithm when modeling complex sys-tems,so that the interaction within the system can be described in more detail on the basis of establishing an overall model.Based on the proposed PFMS theoretical system,this thesis gives specific methods and steps for the dynamic evaluation of system effectiveness to concurrently evaluate the effectiveness of the entire system and describe the dynamic changes of subsystem effec-tiveness.Meanwhile,the dynamic impact of the battlefield environment and enemy equip-ment on the system effectiveness can be analyzed to find the main influencing factors and provide an effective improvement program.For the anti-ship missile strike combat pro-cess,this thesis evaluates its dynamic effectiveness by using the proposed PFMS theory.The evaluation outputs the dynamic effectiveness curve of each subsystem according to real-time battlefield information,thereby reflecting the important impact factor for anti-ship effectiveness improvement.In the proposed PFMS-based effectiveness evaluation method,the state transition probability is an important factor that affects the accuracies of the modeling and solving,which is the key to actual evaluation results.Therefore,to meet the requirements of dif-ferent evaluation accuracies and speeds,this thesis further studies the existing calculation methods of state transition probability and analyses their main advantages,disadvantages,and applicable scenarios.Taking the interception probability of missiles in weapon system confrontation as an example,this thesis comprehensively analyzes the influencing factors of the one-to-one interception probability of missiles.Complete modeling and simulation analysis are conducted,and two methods named the Monte Carlo simulation method and neural network data learning method are presented to meet high accuracy and rapidity requirements.Based on effectiveness evaluation,to make full use of effectiveness evaluation data,this thesis proposes an integrated framework of model simulation,effectiveness evalua-tion,and auxiliary decision-making,which can improve the information-sharing ability of each module to adapt to the ever-changing battlefield situation.Since the optimization target of complex systems is often a problem of high-dimensional functions,to better in-tegrate the PFMS model with an auxiliary decision-making algorithm,and to improve the information utilization,this thesis modifies the master-slave particle swarm algorithm to a multilevel master-slave particle swarm algorithm according to the structural character-istics of PFMS models.The algorithm can make full use of the performance function of the subsystem while optimizing the overall system performance and produces the effect of dimension reduction,so it has more advantages in dealing with the decision-making prob-lems of high-dimensional complex systems described by the PFMS model.This thesis gives the specific process steps of the multilevel master-slave particle swarm algorithm,makes assisted decisions for the anti-ship missile strike combat process,and finally real-izes the integration of effectiveness evaluation and assisted decision-making.