Adaptive pursuit-evasion under adversarial confrontations-part I: Performance-measure statistics in linear-quadratic differential games

摘要

The research proposes a new method of assessing performance risk and uncertainty with high confidence for a linear-quadratic class of stochastic pursuit-evasion on the basis of multi-player confrontations and competitions. The proposed method to characterize higher-order performance uncertainties for this class of differential games is enabled by three prerequisites: 1) the Know-How related to competence; 2) the Know-How-to-Cooperate for coordination between activities; and 3) the performance-measure statistics for coalitional performance robustness capturing adversarial elements from competitive teams and stochastic surprises from stationary environments.