{"id":27160,"date":"2026-01-02T09:00:42","date_gmt":"2026-01-02T00:00:42","guid":{"rendered":"https:\/\/blog.agentsoft.co.kr\/index.php\/2026\/01\/02\/27160\/"},"modified":"2026-01-02T09:00:42","modified_gmt":"2026-01-02T00:00:42","slug":"dqn%ea%b3%bc-a2c-network%eb%a5%bc-%ed%99%9c%ec%9a%a9%ed%95%9c-cartpole-%ea%b0%95%ed%99%94%ed%95%99%ec%8a%b5-%ed%9b%88%eb%a0%a8%ea%b3%bc%ec%a0%95-%eb%b0%8f-code","status":"publish","type":"post","link":"https:\/\/blog.agentsoft.co.kr\/index.php\/2026\/01\/02\/27160\/","title":{"rendered":"DQN\uacfc A2C network\ub97c \ud65c\uc6a9\ud55c CartPole \uac15\ud654\ud559\uc2b5 \ud6c8\ub828\uacfc\uc815 \ubc0f code"},"content":{"rendered":"<p><img decoding=\"async\" src=\"https:\/\/image4.happycampus.com\/Production\/thumb212\/2024\/04\/06\/data29662205-0001.jpg\"><img decoding=\"async\" src=\"https:\/\/image4.happycampus.com\/Production\/thumb212\/2024\/04\/06\/data29662205-0002.jpg\"><\/p>\n<p><strong>\ubaa9\ucc28<\/strong><\/p>\n<p>01 Cartpole environment<br \/>\n02 DQN algorithm &amp; code<br \/>\n03 A2C algorithm &amp; code<\/p>\n<p><strong>\ubcf8\ubb38\ub0b4\uc6a9<\/strong><\/p>\n<p>OpenAI gym\uc758 CartPole\uc740 \uce74\ud2b8 \uc704\uc5d0 \ub9c9\ub300\uae30\uac00 \uace0\uc815\ub418\uc5b4 \uc788\uace0 \ub9c9\ub300\uae30\ub294 \uc911\ub825\uc5d0 \uc758\ud574 \ubc14\ub2e5\uc744 \ud5a5\ud574 \uc790\uc5f0\uc801\uc73c\ub85c \uae30\uc6b8\uac8c \ub418\ub294 \ud658\uacbd\uc744 \uc81c\uacf5\ud55c\ub2e4. CartPole\uc758 \ubaa9\uc801\uc740 \uce74\ud2b8\ub97c \uc88c, \uc6b0\ub85c \uc6c0\uc9c1\uc774\uba70 \ub9c9\ub300\uae30\uac00 \uae30\uc6b8\uc9c0 \uc54a\uace0 \uc11c \uc788\uc744 \uc218 \uc788\ub3c4\ub85d \uc720\uc9c0\uc2dc\ucf1c \uc8fc\ub294 \uac83\uc774 \ubaa9\uc801\uc778\ub370, \uac15\ud654 \ud559\uc2b5 \uc54c\uace0\ub9ac\uc998\uc744 \uc774\uc6a9\ud558\uc5ec \ub9c9\ub300\uae30\ub97c \uc138\uc6b8 \uc218 \uc788\ub294 \ubc29\ubc95\uc744 \uc18c\ud504\ud2b8\uc6e8\uc5b4 \uc5d0\uc774\uc804\ud2b8\uac00 \uc2a4\uc2a4\ub85c \ud559\uc2b5\ud560 \uc218 \uc788\ub3c4\ub85d \ud55c\ub2e4. \ub2e4\uc74c\uc740 CartPole \ud658\uacbd\uc5d0\uc11c \uc0ac\uc6a9\ub418\ub294 observation, action, reward, episode\uc758 \uc2dc\uc791\uacfc \uc885\ub8cc\uc5d0 \ub300\ud55c \uc124\uba85\uc774\ub2e4.<br \/>\nObservation: cart\uc758 \ud604\uc7ac \uc704\uce58, cart\uc758 \uc18d\ub3c4, pole\uc758 \uae30\uc6b8\uae30, pole\uc758 \uc18d\ub3c4\ub97c \uc758\ubbf8\ud55c\ub2e4. Action: \uc624\ub978\ucabd(1) \ub610\ub294 \uc67c\ucabd(0) reward: \ub9e4 \ud0c0\uc784\uc2a4\ud15d\ub9c8\ub2e4 +1\uc529 \ubcf4\uc0c1\uc744 \ubc1b\ub294\ub2e4. Episode Termination: \ub9c9\ub300\uac00 \uc911\uc2ec\uc5d0\uc11c 2.4deg\uc774\uc0c1 \uae30\uc6b8\uc5b4\uc9c0\uac70\ub098, \uba40\ub9ac \ub5a8\uc5b4\uc9c0\uba74 \uc885\ub8cc\ub41c\ub2e4.<br \/>\nstep function\uc744 \ud1b5\ud574 \ub79c\ub364\ud55c \uc6c0\uc9c1\uc784\uc5d0 \ub300\ud55c action\uc744 \ud55c\ubc88 \uc218\ud589\ud558\uace0, action\uc774 \uc2e4\ud589\ub41c \uc774\ud6c4\uc758 \uc0c1\ud0dc(observation)\uc640, \ubcf4\uc0c1(reward), \ub9c9\ub300\uac00 \uc4f0\ub7ec\uc84c\ub294\uc9c0\uc758 \uc5ec\ubd80(done) \ub4f1\uc758 \uc815\ubcf4\uac00 \ubc18\ud658\ub41c\ub2e4.<br \/>\nCode:<br \/>\n\ucf54\ub4dc\ub294 \uc81c\uac00 \uc9c1\uc811 \uc791\uc131\ud55c \uac83\uc774 \uc544\ub2d8\uc744 \ubc1d\ud799\ub2c8\ub2e4. \uc0b0\uc5c5\uc815\ubcf4\uc2dc\uc2a4\ud15c\uc804\uacf5 \ub525\ub7ec\ub2dd \uc218\uc5c5\uc744 \ub4e3\uace0 \ud504\ub85c\uc81d\ud2b8\ub3c4 \uc218\ud589\ud588\ub358\ud130\ub77c \ub525\ub7ec\ub2dd\uacfc \uac15\ud654\ud559\uc2b5\uc744 \uc870\ud569\ud55c \uc54c\uace0\ub9ac\uc998\uc5d0 \uc790\uc5f0\uc2a4\ub7fd\uac8c \uad00\uc2ec\uc774 \uc0dd\uacbc\uc2b5\ub2c8\ub2e4. \ub530\ub77c\uc11c \uc2e4\uc81c \uad6c\ud604\ub41c DQN network\uc640 A2C(Advantage Actor-Critic) network\uc758 \ucf54\ub4dc\ub97c \uc2e4\ud589\uc2dc\ud0a4\uace0 \ubd84\uc11d\ud588\uc2b5\ub2c8\ub2e4.<\/p>\n<p>\ucd9c\ucc98 : <a href=\"https:\/\/www.happycampus.com\/report-doc\/29662205\/\" target=\"_blank\">\ud574\ud53c\ucea0\ud37c\uc2a4<\/a><\/p>\n","protected":false},"excerpt":{"rendered":"<p>\ubaa9\ucc28 01 Cartpole environment 02 DQN algorithm &amp; code 03 A2C algorithm &amp; code \ubcf8\ubb38\ub0b4\uc6a9 OpenAI gym\uc758 CartPole\uc740 \uce74\ud2b8 \uc704\uc5d0 \ub9c9\ub300\uae30\uac00 \uace0\uc815\ub418\uc5b4 \uc788\uace0 \ub9c9\ub300\uae30\ub294 \uc911\ub825\uc5d0 \uc758\ud574 \ubc14\ub2e5\uc744 \ud5a5\ud574 \uc790\uc5f0\uc801\uc73c\ub85c \uae30\uc6b8\uac8c \ub418\ub294 \ud658\uacbd\uc744 \uc81c\uacf5\ud55c\ub2e4. CartPole\uc758 \ubaa9\uc801\uc740 \uce74\ud2b8\ub97c \uc88c, \uc6b0\ub85c \uc6c0\uc9c1\uc774\uba70 \ub9c9\ub300\uae30\uac00 \uae30\uc6b8\uc9c0 \uc54a\uace0 \uc11c \uc788\uc744 \uc218 \uc788\ub3c4\ub85d \uc720\uc9c0\uc2dc\ucf1c \uc8fc\ub294 \uac83\uc774 \ubaa9\uc801\uc778\ub370, \uac15\ud654 \ud559\uc2b5 \uc54c\uace0\ub9ac\uc998\uc744 \uc774\uc6a9\ud558\uc5ec \ub9c9\ub300\uae30\ub97c [&hellip;]<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[1],"tags":[33910,33911,13623,33909,33912],"class_list":["post-27160","post","type-post","status-publish","format-standard","hentry","category-uncategorized","tag-a2c","tag-cartpole","tag-code","tag-dqn","tag-33912"],"_links":{"self":[{"href":"https:\/\/blog.agentsoft.co.kr\/index.php\/wp-json\/wp\/v2\/posts\/27160","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/blog.agentsoft.co.kr\/index.php\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/blog.agentsoft.co.kr\/index.php\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/blog.agentsoft.co.kr\/index.php\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/blog.agentsoft.co.kr\/index.php\/wp-json\/wp\/v2\/comments?post=27160"}],"version-history":[{"count":0,"href":"https:\/\/blog.agentsoft.co.kr\/index.php\/wp-json\/wp\/v2\/posts\/27160\/revisions"}],"wp:attachment":[{"href":"https:\/\/blog.agentsoft.co.kr\/index.php\/wp-json\/wp\/v2\/media?parent=27160"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/blog.agentsoft.co.kr\/index.php\/wp-json\/wp\/v2\/categories?post=27160"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/blog.agentsoft.co.kr\/index.php\/wp-json\/wp\/v2\/tags?post=27160"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}