Abstract:Air defense operation system of systems (ADOSoS) is a typical complex system with nonlinear, adaptive, uncertain, emergent properties. Traditional capability index selection methods in which the indices are selected based on expert knowledge and experience and decomposed in form of tree-diagram, can not describe the capability of the ADOSoS objectively and accurately. Here, the ADOSoS operations based on Weapon System of Systems Synthetic Simulation Test Bed are designed and simulated, and the assembled machine learning model (RF-GLR) is built, the Random Forest with Generalized Linear Regression algorithm, to farm the multidimensional and holographic operation data deeply. According to the model, one capability index of ADOSoS { \Air Defense Threat Index" has been proposed by
the emergent summation of the remarkable basic-capacity-indices of subsystems which re°ect the holistic capability of air defense system. The effectiveness of \Air Defense Threat Index" to describe holistic, dynamic and adjective ADOSoS is verifed by contradistinctive experiment, and the application for prediction and the real-time observation for the ADOSoS with the index are also illustrated. Our method supplies a feasible way to construct the capability indices scientifically and measures the system of systems objectively, and provides the guidance for the commanders to make decisions under the information condition.
姚晓毅,郭圣明,胡晓峰,杨镜宇. 基于RF-GLR组合机器学习算法的防空体系能力指标挖掘方法研究[J]. 指挥与控制学报, 2015, 1(3): 269-277.
YAO Xiao-Yi,GUO Sheng-Ming,HU Xiao-Feng,YANG Jing-Yu. Designing Evaluation Indices for Air Defense Systems Based on RF-GLR Integrated Machine Learning Algorithms. Journal of Command and Control, 2015, 1(3): 269-277.