RL之PG:基于TF利用策略梯度算法玩Cartpole游戏实现智能得高分

版权声明:本文为博主原创文章,未经博主允许不得转载。 https://blog.csdn.net/qq_41185868/article/details/83309071

RL之PG:基于TF利用策略梯度算法玩Cartpole游戏实现智能得高分

输出结果

视频观看地址:强化学习—基于TF利用策略梯度算法玩Cartpole游戏实现智能得高分

设计思路

测试过程

Episode: 1 ~ 5 Average reward: 15.000000.
Episode: 6 ~ 10 Average reward: 18.000000.
Episode: 11 ~ 15 Average reward: 19.000000.
Episode: 16 ~ 20 Average reward: 16.000000.
Episode: 21 ~ 25 Average reward: 20.000000.
Episode: 26 ~ 30 Average reward: 24.000000.
Episode: 31 ~ 35 Average reward: 17.000000.
Episode: 36 ~ 40 Average reward: 18.000000.
Episode: 41 ~ 45 Average reward: 28.000000.
Episode: 46 ~ 50 Average reward: 13.000000.
Episode: 51 ~ 55 Average reward: 14.000000.
Episode: 56 ~ 60 Average reward: 24.000000.
Episode: 61 ~ 65 Average reward: 31.000000.
Episode: 66 ~ 70 Average reward: 15.000000.
Episode: 71 ~ 75 Average reward: 29.000000.
Episode: 76 ~ 80 Average reward: 22.000000.
Episode: 81 ~ 85 Average reward: 21.000000.
Episode: 86 ~ 90 Average reward: 25.000000.
Episode: 91 ~ 95 Average reward: 24.000000.
Episode: 96 ~ 100 Average reward: 21.000000.
Episode: 101 ~ 105 Average reward: 20.000000.
Episode: 106 ~ 110 Average reward: 35.000000.
Episode: 111 ~ 115 Average reward: 34.000000.
Episode: 116 ~ 120 Average reward: 26.000000.
Episode: 121 ~ 125 Average reward: 19.000000.
Episode: 126 ~ 130 Average reward: 17.000000.
Episode: 131 ~ 135 Average reward: 39.000000.
Episode: 136 ~ 140 Average reward: 28.000000.
Episode: 141 ~ 145 Average reward: 22.000000.
Episode: 146 ~ 150 Average reward: 22.000000.
Episode: 151 ~ 155 Average reward: 21.000000.
Episode: 156 ~ 160 Average reward: 29.000000.
Episode: 161 ~ 165 Average reward: 25.000000.
Episode: 166 ~ 170 Average reward: 24.000000.
Episode: 171 ~ 175 Average reward: 27.000000.
Episode: 176 ~ 180 Average reward: 24.000000.
Episode: 181 ~ 185 Average reward: 23.000000.
Episode: 186 ~ 190 Average reward: 24.000000.
Episode: 191 ~ 195 Average reward: 25.000000.
Episode: 196 ~ 200 Average reward: 36.000000.
Episode: 201 ~ 205 Average reward: 24.000000.
Episode: 206 ~ 210 Average reward: 20.000000.
Episode: 211 ~ 215 Average reward: 27.000000.
Episode: 216 ~ 220 Average reward: 38.000000.
Episode: 221 ~ 225 Average reward: 31.000000.
Episode: 226 ~ 230 Average reward: 25.000000.
Episode: 231 ~ 235 Average reward: 31.000000.
Episode: 236 ~ 240 Average reward: 25.000000.
Episode: 241 ~ 245 Average reward: 23.000000.
Episode: 246 ~ 250 Average reward: 26.000000.
Episode: 251 ~ 255 Average reward: 21.000000.
Episode: 256 ~ 260 Average reward: 25.000000.
Episode: 261 ~ 265 Average reward: 30.000000.
Episode: 266 ~ 270 Average reward: 28.000000.
Episode: 271 ~ 275 Average reward: 30.000000.
Episode: 276 ~ 280 Average reward: 36.000000.
Episode: 281 ~ 285 Average reward: 28.000000.
Episode: 286 ~ 290 Average reward: 31.000000.
Episode: 291 ~ 295 Average reward: 26.000000.
Episode: 296 ~ 300 Average reward: 21.000000.
Episode: 301 ~ 305 Average reward: 29.000000.
Episode: 306 ~ 310 Average reward: 32.000000.
Episode: 311 ~ 315 Average reward: 22.000000.
Episode: 316 ~ 320 Average reward: 21.000000.
Episode: 321 ~ 325 Average reward: 32.000000.
Episode: 326 ~ 330 Average reward: 41.000000.
Episode: 331 ~ 335 Average reward: 24.000000.
Episode: 336 ~ 340 Average reward: 30.000000.
Episode: 341 ~ 345 Average reward: 30.000000.
Episode: 346 ~ 350 Average reward: 25.000000.
Episode: 351 ~ 355 Average reward: 35.000000.
Episode: 356 ~ 360 Average reward: 22.000000.
Episode: 361 ~ 365 Average reward: 49.000000.
Episode: 366 ~ 370 Average reward: 30.000000.
Episode: 371 ~ 375 Average reward: 37.000000.
Episode: 376 ~ 380 Average reward: 19.000000.
Episode: 381 ~ 385 Average reward: 55.000000.
Episode: 386 ~ 390 Average reward: 32.000000.
Episode: 391 ~ 395 Average reward: 52.000000.
Episode: 396 ~ 400 Average reward: 35.000000.
Episode: 401 ~ 405 Average reward: 35.000000.
Episode: 406 ~ 410 Average reward: 41.000000.
Episode: 411 ~ 415 Average reward: 49.000000.
Episode: 416 ~ 420 Average reward: 44.000000.
Episode: 421 ~ 425 Average reward: 59.000000.
Episode: 426 ~ 430 Average reward: 24.000000.
Episode: 431 ~ 435 Average reward: 37.000000.
Episode: 436 ~ 440 Average reward: 35.000000.
Episode: 441 ~ 445 Average reward: 37.000000.
Episode: 446 ~ 450 Average reward: 37.000000.
Episode: 451 ~ 455 Average reward: 32.000000.
Episode: 456 ~ 460 Average reward: 56.000000.
Episode: 461 ~ 465 Average reward: 35.000000.
Episode: 466 ~ 470 Average reward: 29.000000.
Episode: 471 ~ 475 Average reward: 40.000000.
Episode: 476 ~ 480 Average reward: 46.000000.
Episode: 481 ~ 485 Average reward: 37.000000.
Episode: 486 ~ 490 Average reward: 34.000000.
Episode: 491 ~ 495 Average reward: 47.000000.
Episode: 496 ~ 500 Average reward: 49.000000.
Episode: 501 ~ 505 Average reward: 53.000000.
Episode: 506 ~ 510 Average reward: 38.000000.
Episode: 511 ~ 515 Average reward: 42.000000.
Episode: 516 ~ 520 Average reward: 32.000000.
Episode: 521 ~ 525 Average reward: 43.000000.
Episode: 526 ~ 530 Average reward: 38.000000.
Episode: 531 ~ 535 Average reward: 46.000000.
Episode: 536 ~ 540 Average reward: 68.000000.
Episode: 541 ~ 545 Average reward: 33.000000.
Episode: 546 ~ 550 Average reward: 37.000000.
Episode: 551 ~ 555 Average reward: 36.000000.
Episode: 556 ~ 560 Average reward: 36.000000.
Episode: 561 ~ 565 Average reward: 44.000000.
Episode: 566 ~ 570 Average reward: 46.000000.
Episode: 571 ~ 575 Average reward: 55.000000.
Episode: 576 ~ 580 Average reward: 45.000000.
Episode: 581 ~ 585 Average reward: 53.000000.
Episode: 586 ~ 590 Average reward: 54.000000.
Episode: 591 ~ 595 Average reward: 37.000000.
Episode: 596 ~ 600 Average reward: 61.000000.
Episode: 601 ~ 605 Average reward: 57.000000.
Episode: 606 ~ 610 Average reward: 68.000000.
Episode: 611 ~ 615 Average reward: 60.000000.
Episode: 616 ~ 620 Average reward: 46.000000.
Episode: 621 ~ 625 Average reward: 52.000000.
Episode: 626 ~ 630 Average reward: 44.000000.
Episode: 631 ~ 635 Average reward: 64.000000.
Episode: 636 ~ 640 Average reward: 60.000000.
Episode: 641 ~ 645 Average reward: 57.000000.
Episode: 646 ~ 650 Average reward: 74.000000.
Episode: 651 ~ 655 Average reward: 65.000000.
Episode: 656 ~ 660 Average reward: 48.000000.
Episode: 661 ~ 665 Average reward: 43.000000.
Episode: 666 ~ 670 Average reward: 54.000000.
Episode: 671 ~ 675 Average reward: 68.000000.
Episode: 676 ~ 680 Average reward: 42.000000.
Episode: 681 ~ 685 Average reward: 50.000000.
Episode: 686 ~ 690 Average reward: 71.000000.
Episode: 691 ~ 695 Average reward: 61.000000.
Episode: 696 ~ 700 Average reward: 74.000000.
Episode: 701 ~ 705 Average reward: 88.000000.
Episode: 706 ~ 710 Average reward: 50.000000.
Episode: 711 ~ 715 Average reward: 51.000000.
Episode: 716 ~ 720 Average reward: 67.000000.
Episode: 721 ~ 725 Average reward: 71.000000.
Episode: 726 ~ 730 Average reward: 74.000000.
Episode: 731 ~ 735 Average reward: 74.000000.
Episode: 736 ~ 740 Average reward: 68.000000.
Episode: 741 ~ 745 Average reward: 71.000000.
Episode: 746 ~ 750 Average reward: 67.000000.
Episode: 751 ~ 755 Average reward: 63.000000.
Episode: 756 ~ 760 Average reward: 61.000000.
Episode: 761 ~ 765 Average reward: 46.000000.
Episode: 766 ~ 770 Average reward: 64.000000.
Episode: 771 ~ 775 Average reward: 48.000000.
Episode: 776 ~ 780 Average reward: 71.000000.
Episode: 781 ~ 785 Average reward: 67.000000.
Episode: 786 ~ 790 Average reward: 93.000000.
Episode: 791 ~ 795 Average reward: 74.000000.
Episode: 796 ~ 800 Average reward: 57.000000.
Episode: 801 ~ 805 Average reward: 74.000000.
Episode: 806 ~ 810 Average reward: 57.000000.
Episode: 811 ~ 815 Average reward: 72.000000.
Episode: 816 ~ 820 Average reward: 70.000000.
Episode: 821 ~ 825 Average reward: 53.000000.
Episode: 826 ~ 830 Average reward: 40.000000.
Episode: 831 ~ 835 Average reward: 79.000000.
Episode: 836 ~ 840 Average reward: 62.000000.
Episode: 841 ~ 845 Average reward: 65.000000.
Episode: 846 ~ 850 Average reward: 91.000000.
Episode: 851 ~ 855 Average reward: 50.000000.
Episode: 856 ~ 860 Average reward: 74.000000.
Episode: 861 ~ 865 Average reward: 84.000000.
Episode: 866 ~ 870 Average reward: 77.000000.
Episode: 871 ~ 875 Average reward: 66.000000.
Episode: 876 ~ 880 Average reward: 79.000000.
Episode: 881 ~ 885 Average reward: 69.000000.
Episode: 886 ~ 890 Average reward: 93.000000.
Episode: 891 ~ 895 Average reward: 66.000000.
Episode: 896 ~ 900 Average reward: 81.000000.
Episode: 901 ~ 905 Average reward: 61.000000.
Episode: 906 ~ 910 Average reward: 70.000000.
Episode: 911 ~ 915 Average reward: 73.000000.
Episode: 916 ~ 920 Average reward: 68.000000.
Episode: 921 ~ 925 Average reward: 62.000000.
Episode: 926 ~ 930 Average reward: 89.000000.
Episode: 931 ~ 935 Average reward: 87.000000.
Episode: 936 ~ 940 Average reward: 63.000000.
Episode: 941 ~ 945 Average reward: 96.000000.
Episode: 946 ~ 950 Average reward: 100.000000.
Episode: 951 ~ 955 Average reward: 92.000000.
Episode: 956 ~ 960 Average reward: 88.000000.
Episode: 961 ~ 965 Average reward: 85.000000.
Episode: 966 ~ 970 Average reward: 82.000000.
Episode: 971 ~ 975 Average reward: 78.000000.
Episode: 976 ~ 980 Average reward: 68.000000.
Episode: 981 ~ 985 Average reward: 62.000000.
Episode: 986 ~ 990 Average reward: 92.000000.
Episode: 991 ~ 995 Average reward: 78.000000.
Episode: 996 ~ 1000 Average reward: 109.000000.
Episode: 1001 ~ 1005 Average reward: 54.000000.
Episode: 1006 ~ 1010 Average reward: 61.000000.
Episode: 1011 ~ 1015 Average reward: 99.000000.
Episode: 1016 ~ 1020 Average reward: 89.000000.
Episode: 1021 ~ 1025 Average reward: 99.000000.
Episode: 1026 ~ 1030 Average reward: 72.000000.
Episode: 1031 ~ 1035 Average reward: 92.000000.
Episode: 1036 ~ 1040 Average reward: 103.000000.
Episode: 1041 ~ 1045 Average reward: 111.000000.
Episode: 1046 ~ 1050 Average reward: 142.000000.
Episode: 1051 ~ 1055 Average reward: 137.000000.
Episode: 1056 ~ 1060 Average reward: 153.000000.
Episode: 1061 ~ 1065 Average reward: 149.000000.
Episode: 1066 ~ 1070 Average reward: 81.000000.
Episode: 1071 ~ 1075 Average reward: 127.000000.
Episode: 1076 ~ 1080 Average reward: 72.000000.
Episode: 1081 ~ 1085 Average reward: 149.000000.
Episode: 1086 ~ 1090 Average reward: 89.000000.
Episode: 1091 ~ 1095 Average reward: 111.000000.
Episode: 1096 ~ 1100 Average reward: 155.000000.
Episode: 1101 ~ 1105 Average reward: 126.000000.
Episode: 1106 ~ 1110 Average reward: 139.000000.
Episode: 1111 ~ 1115 Average reward: 115.000000.
Episode: 1116 ~ 1120 Average reward: 145.000000.
Episode: 1121 ~ 1125 Average reward: 178.000000.
Episode: 1126 ~ 1130 Average reward: 150.000000.
Episode: 1131 ~ 1135 Average reward: 168.000000.
Episode: 1136 ~ 1140 Average reward: 128.000000.
Episode: 1141 ~ 1145 Average reward: 152.000000.
Episode: 1146 ~ 1150 Average reward: 153.000000.
Episode: 1151 ~ 1155 Average reward: 117.000000.
Episode: 1156 ~ 1160 Average reward: 134.000000.
Episode: 1161 ~ 1165 Average reward: 104.000000.
Episode: 1166 ~ 1170 Average reward: 141.000000.
Episode: 1171 ~ 1175 Average reward: 154.000000.
Episode: 1176 ~ 1180 Average reward: 157.000000.
Episode: 1181 ~ 1185 Average reward: 143.000000.
Episode: 1186 ~ 1190 Average reward: 152.000000.
Episode: 1191 ~ 1195 Average reward: 154.000000.
Episode: 1196 ~ 1200 Average reward: 154.000000.
Episode: 1201 ~ 1205 Average reward: 186.000000.
Episode: 1206 ~ 1210 Average reward: 181.000000.
Episode: 1211 ~ 1215 Average reward: 162.000000.
Episode: 1216 ~ 1220 Average reward: 172.000000.
Episode: 1221 ~ 1225 Average reward: 200.000000.

猜你喜欢

转载自blog.csdn.net/qq_41185868/article/details/83309071