Research On Assessment Methods And Modeling Of Target Damage Based On Bayesian Network

The Assessment on target firepower damage aims at getting the firepower damage level of the operational goals by quantifying the operational effects and comparing the standard damage degree of the target firepower. To a large extent, the accuracy of the assessment on target firepower damage affects the command's decision in the subsequent military action. However, the accuracy of the assessment on target firepower damage is challenged by the appearance of modern means, such as military deception and camouflage. Meanwhile, the low efficient design and development of the assessment has to be acquired from large amounts of calculations in the complicated and various war conditions. Therefore, geared to the assessment problems with uncertain conditions, it is significant to find out an accurate method, and raise an efficient assessment model and construction method.Based on the probability theory and graph theory, Bayesian network could express and inference the uncertain knowledge, and solve uncertain problems caused by random conditions. Hence the dissertation brings Bayesian network in assessment on target firepower damage to solve the uncertain and random problems. The dissertation emphasizes on introducing the basic Bayesian network methods and the expanding of its related methods, and building the basic model which are based on Bayesian network; then it discusses on the basic model and puts forward to weighting the Bayesian network in consideration of the relationship of network node. Weighting the Bayesian network realizes more accurate probability delivery, and meanwhile distributes the weighting of each node dynamically according to the outside conditions. Therefore, the improved model of weighting the Bayesian network is built. The dissertation conducts to the emulation realization of the model in last chapter, analyzes their influences on precision assessment by introducing the naive approach of Bayesian network and comparing with the weighting of Bayesian network, and finally gets the weighted Bayesian network. According to the actual situation, the predicted result of the weighted model is more close to the real result, and the confidence is higher than before.In the end, with compositional modeling applied, the assessment software on the firepower damage is obtained, and on the basis of the software, the accurate assessment can be preliminarily acquired under the uncertain conditions. The research of the dissertation generally applies to not only assessment on target damage under various operational situations, but also realtime, dynamic, fast and accurate assessment on warfare damage by modifying and expanding the model.