{"height":"190","width":"100%","title":"MAPPO\u3092\u4f7f\u3063\u305f\u5354\u8abf\u5b66\u7fd2","categories":["AI"],"image_url":"https://cdn-ak.f.st-hatena.com/images/fotolife/y/yoshishinnze/20260110/20260110153715.gif","provider_url":"https://hatena.blog","type":"rich","blog_url":"https://yoshishinnze.hatenablog.com/","author_name":"yoshishinnze","author_url":"https://blog.hatena.ne.jp/yoshishinnze/","blog_title":"\u6687\u3055\u3048\u3042\u308c\u3070\u30a2\u30eb\u30b4\u30ea\u30ba\u30e0\u3044\u3058\u308a","description":"\u76ee\u6b21 MAPPO (Multi-Agent PPO) \u306f\u3001\u30b7\u30f3\u30b0\u30eb\u30a8\u30fc\u30b8\u30a7\u30f3\u30c8\u5411\u3051\u306b\u975e\u5e38\u306b\u9ad8\u3044\u5b9f\u7e3e\u3092\u6301\u3064 PPO (Proximal Policy Optimization) \u3092\u3001\u30de\u30eb\u30c1\u30a8\u30fc\u30b8\u30a7\u30f3\u30c8\u74b0\u5883\uff08\u8907\u6570\u306e\u30a8\u30fc\u30b8\u30a7\u30f3\u30c8\u304c\u5354\u529b\u30fb\u7af6\u5408\u3059\u308b\u74b0\u5883\uff09\u306b\u62e1\u5f35\u3057\u305f\u30a2\u30eb\u30b4\u30ea\u30ba\u30e0\u3067\u3059\u3002 \u73fe\u5728\u3001\u30de\u30eb\u30c1\u30a8\u30fc\u30b8\u30a7\u30f3\u30c8\u5f37\u5316\u5b66\u7fd2\uff08MARL\uff09\u306b\u304a\u3044\u3066\u3001\u6700\u3082\u6a19\u6e96\u7684\u304b\u3064\u5f37\u529b\u306a\u624b\u6cd5\u306e\u4e00\u3064\u3068\u3057\u3066\u77e5\u3089\u308c\u3066\u3044\u307e\u3059\u3002 MAPPO\u306e\u57fa\u672c\u69cb\u9020\uff1aCTDE MAPPO\u306f\u3001 CTDE (Centralized Training, Decentralized Execution) \u3068\u3044\u3046\u67a0\u7d44\u307f\u3092\u63a1\u7528\u3057\u3066\u3044\u307e\u3059\u3002 \u96c6\u4e2d\u5b66\u7fd2 (Centralize\u2026","version":"1.0","provider_name":"Hatena Blog","published":"2026-01-10 15:38:24","url":"https://yoshishinnze.hatenablog.com/entry/2026/01/10/153824","html":"<iframe src=\"https://hatenablog-parts.com/embed?url=https%3A%2F%2Fyoshishinnze.hatenablog.com%2Fentry%2F2026%2F01%2F10%2F153824\" title=\"MAPPO\u3092\u4f7f\u3063\u305f\u5354\u8abf\u5b66\u7fd2 - \u6687\u3055\u3048\u3042\u308c\u3070\u30a2\u30eb\u30b4\u30ea\u30ba\u30e0\u3044\u3058\u308a\" class=\"embed-card embed-blogcard\" scrolling=\"no\" frameborder=\"0\" style=\"display: block; width: 100%; height: 190px; max-width: 500px; margin: 10px 0px;\"></iframe>"}