Intelligent Effectiveness Evaluation Based On Combat Simulation Data

With the development of complex system simulation techniques,computational capabilities,and data management capabilities,the simulation results tend to be big data,which pose new challenges to the effectiveness evaluation model.There are also high-dimension,high-redundancy,and high-correlation issues among indexes,and more effective methods are needed to reduce the dimensions of indexes.Simultaneously,how to optimize the simulation parameters based on the evaluation results becomes more practical than the effectiveness evaluation itself.Based on the above background,we use artificial intelligence technology to carry out research on effectiveness evaluation technology.The main contents of the research include the following aspects:Firstly,using the sparse autoencoder neural network to reduce the dimensionality of simulation indicators.The 10 indicators of the aircraft carrier simulation experiment are dimension reduced and the evaluation values are obtained.The experimental comparison shows that the effect of the autoencoder neural network is better than the traditional PCA method,and the data features can be better extracted.Secondly,the architecture of deep neural network is innovated,and TPN?Triple Path Network?neural network model is proposed.In this paper,six kinds of deep neural networks based on CNN are built to evaluate the performance.It is found that ResNet works best in the simulation data used in this paper.By considering the modular structure and the parallel operation of the network layer,the TPN network model is improved based on the ResNet and DPN networks.And the accuracy of TPN in MNIST and CIFAR-10reaches 99.28%and 94.65%respectively,which is better than DPN.In the performance evaluation experiment,the TPN network also achieves the best training effect,with the lowest regression error and the best convergence effect.According to the trained TPN network model,an intelligent performance evaluation model is established.Finally,a performance evaluation based on improved ant colony?POIAC?algorithm is proposed to achieve global optimization of simulation parameters.Two kinds of path selection rules,four pheromone update rules and three pheromone interval rules are considered in the ant colony algorithm.In the pheromone update rules,the improved elite ant colony rule AScomp is proposed.The designed 24 sets of algorithms are compared and the results show that the proposed AS_comp rule is most suitable for solving the inverse problem of efficiency evaluation.On this basis,the improved ant colony algorithm is used to achieve the global optimization of the simulation parameter input,and finally the parameter combination with the largest performance evaluation value is obtained.