Combat effectiveness evaluation of real combat exercise based on data-driven
DAI Yao-zong, Jia Guo, Yisen Wang, Xiaofeng Guo · 2018
War is a typical complex giant system. In the field of real combat exercise, the experience-driven effectiveness evaluation method has the problems of subjectivity and inefficiency in evaluation, and has been unable to meet the requirements of informationized warfare. In this paper, combat effectiveness is divided into three components: win-lose effectiveness, military casualty effectiveness and weapon damage effectiveness. Data-driven algorithm-least square support vector machine (LSSVM) is introduced to build combat effectiveness evaluation model of real combat exercise. The historical data train the model and establish the nonlinear mapping complex relationship between efficiency component and operational effectiveness. Experiment shows that the LS-SVM-based tactical effectiveness assessment method has higher efficiency and accuracy.