The Research On The Method Of Combat Effectiveness Evaluation Based On Neural Network

At the present time, the combat effectiveness evaluation has become one of the hottest subjects in the areas of military academia and weapons and equipment development. As the diversity of warfare conditions and variety of weapons increase, It could not meet the requirements of the modern information warfare only to evaluate the combat effectiveness with the military experts' experience and intuition. So the objective, rapid and effective technologies are necessary for the development of evaluations.In the study of combat effectiveness evaluation, combat simulation always plays a key role. However, with the information age coming, combat simulation researches are facing many problems. For example, to dealt with the "curse of dimensionality", the "curse of complexity" and nonlinear characteristics and so on. While neural network is very useful to solve these problems, because it is a cosmically system which can work in parallel, and it can establish the model according to the internal relations. So this paper mainly studied the method of combat effectiveness evaluation based on the neural network by using the neural network technology.Firstly, the concept and methods of combat effectiveness evaluation are summarized in this paper. Then the paper introduces the usual methods of combat effectiveness evaluation and the theory of neural network, and mainly studies the BP and RBF neural network models.Secondly, by analyzing the characteristics of combat effectiveness evaluation, the paper demonstrates the feasibility of the method of neural network for the combat effectiveness evaluation, and points out the advantages and disadvantages of the neural network. Based on the above aspects, an advanced RBF neural network arithmetic, which can enhance the operational precision and the capability of application, was put forward. Then established the model of combat effectiveness evaluation based on this advanced arithmetic.Finally, according to the model, a system of combat effectiveness evaluation was designed with VC++ and Matlab, and with this system the effectiveness of the red agent in EINSTein was evaluated. The simulation results proved that this system is usable and available, and it is also feasible to use the neural network in the field of combat effectiveness evaluation.