Adaptive dynamic programming for online solution of a zero-sum differential game-零和微分对策在线解决方案的自适应动态规划.pdfVIP

201l ．，函n加f砀∞哆4印f 9(3)353_360 DOI for AdaptiVe online dynamicprogramming 0IofaaZer0-SUmzer0．sum nlIIerentlaldif．ferentialgame game V＆蟠IE 2 DragIlna 1，FraIll【LEMS Rt络earch C印毗E雒t 1．United钕量lIIologies Har哟rd。CT06108，U．S．A．； 2．Au“oⅡ嘶on卸dRob嘶csResearch ofTcx硒al IIIsdtll峨UIIiVersityAdington，FortWbr山1X76118．U．S．A． Abst阻ct：Tllis uses p印er诵llpresent柚appmximate，adaptivedyn删c tlIe pm铲a删[11ing(ADP)alg硎thm，that ideaof reinforcement detenrIineonlinettIeNaSh solution integral leanling(IltL)，to for zero． equilibrium t11e铆o．player sumdiffbfentialwit|llinear andinfinite horizon cost．The isbuiltaroundaniterative g锄e dynamics quadmtic algorithm methodt|Iathasbeen intheconⅡDl tlle deVeloped continuous一血_Ile engineeringcommuni哆forsolving g锄ealgebraic 硒ccati underliesthe hereshowhowtheADP willenhancethe equaIion(CT．GARE)，which