对任何仿真系统,只有在保证一定可信度的基础才具有实用价值.在文(Ⅰ)战损仿真原型系统的基础上,结合经济性和可行性的权衡分析,提出了装甲装备战损仿真系统的修正策略.进行了典型条件下的杀伤力参数试验,利用神经网络技术修正BAD参数经验值和试验值的误差,然后将训练完成的神经网络模块嵌入在PiercingAgent中,实现对BAD模型参数的动态修正.最后,进行了仿真试验,并利用实战数据和全尺寸物理实验数据进行了仿真系统的有效性验证.研究成果为装甲装备战损或具有类似特征系统的仿真提供了一种研究途径.%For any simulating system, only its credibility being confirmed can it take on practical worthiness. When economical and feasibility are concerned, the verifying and validating tactic is put forward to battle damage simulation prototype system in paper (I). Then BAD parameter experiment is done under typical condition. Neural network is used to verify error between experiment data and experiential value of BAD parameters, the trained neural network is embedded into Piercing _ Agent. At last system simulating result is validated through real war data and experiment data of real equipment. So a fresh approach on complexity theory is explored to battle damage simulating or analogous problem.
展开▼