学科分类
/ 1
1 个结果
  • 简介:Thispaperwillpresentanapproximate/adaptivedynamicprogramming(ADP)algorithm,thatusestheideaofintegralreinforcementlearning(IRL),todetermineonlinetheNashequilibriumsolutionforthetwo-playerzerosumdifferentialgamewithlineardynamicsandinfinitehorizonquadraticcost.Thealgorithmisbuiltaroundaniterativemethodthathasbeendevelopedinthecontrolengineeringcommunityforsolvingthecontinuous-timegamealgebraicRiccatiequation(CT-GARE),whichunderliesthegameproblem.W...

  • 标签: 近似 / 适应的动态编程 游戏代数学的 Riccati 方程 零和的微分游戏 纳什平衡