Ȩ > Ç¥ÁØÈ °³¿ä > TTAÀÇ Ç¥ÁØÇöȲ
Ç¥ÁعøÈ£ | TTAK.KO-10.1340 | ±¸Ç¥ÁعøÈ£ | |
---|---|---|---|
Á¦°³Á¤ÀÏ | 2021-12-08 | ÃÑÆäÀÌÁö | 25 |
ÇѱÛÇ¥Áظí | ¸ðÀÇ È¯°æ ±â¹Ý ÀΰøÁö´É °ÔÀÓ ¿¡ÀÌÀüÆ® »ý¼º ¿ä±¸»çÇ× | ||
¿µ¹®Ç¥Áظí | Game AI agent Generation Guidelines based on Simulation Environment | ||
Çѱ۳»¿ë¿ä¾à | ÃÖ±Ù ±â°èÇнÀÀ» ÀÌ¿ëÇÑ ÀΰøÁö´É ¿¡ÀÌÀüÆ® »ý¼º ±â¼ú¿¡ ´ëÇÑ °ü½ÉÀÌ ³ô¾ÆÁö°í ÀÖÀ¸¸ç, ƯÈ÷ 4Â÷ »ê¾÷Çõ¸í°ú °ü·ÃµÈ ´Ù¾çÇÑ ºÐ¾ß¿¡¼ È°¹ßÈ÷ ¿¬±¸°¡ ÁøÇàµÇ°í ÀÖ´Ù. ÇÏÁö¸¸ Çö½Ç ¼¼°è¿¡¼ ÇнÀÀ» ÁøÇàÇÏ´Â µ¥ ÀÖ¾î ºñ¿ë°ú ½Ã°£ µî ¸¹Àº ÇÑ°èÁ¡ÀÌ Á¸ÀçÇϱ⠶§¹®¿¡ ½Ã¹Ä·¹ÀÌÅ͸¦ È°¿ëÇÏ´Â Ãß¼¼µµ Áõ°¡ÇÏ°í ÀÖ´Ù.
ÀÌ¿Í À¯»çÇÏ°Ô °ÔÀÓÀº ºñ±³Àû ȯ°æÀ¸·ÎºÎÅÍ Á¤º¸ ¼öÁýÀÌ ¿ëÀÌÇÏ°í ¿¡ÀÌÀüÆ® Çൿ¿¡ µû¸¥ º¸»óÀ» ÀûÀýÇÏ°Ô Á¶ÀýÇÒ ¼ö ÀÖ¾î¼, Çö½Ç ¼¼°èÀÇ º¹ÀâÇÑ ¹®Á¦µéÀ» ÇØ°áÇϱâ À§ÇÑ ½Ã¹Ä·¹ÀÌÅͷμ È°¿ëµµ°¡ Å©´Ù. ´Ù¸¸ ±Ù·¡ °ÔÀÓÀÇ °æ¿ì »ç¿ëÀÚ ¸¸Á·µµ Çâ»óÀ» À§ÇØ ´Ù¾çÇÑ ½Ã³ª¸®¿À ¹× ±â´É ±¸ÇöÀÌ ¿ä±¸µÇ±â ¶§¹®¿¡ ±× ÀÚü·Î¼ÀÇ º¹Àâµµ ¶ÇÇÑ ³ª³¯ÀÌ Áõ°¡ÇÏ´Â Ãß¼¼ÀÌ´Ù. µû¶ó¼ °ÔÀÓ °³¹ß»ç´Â ½Ã°£, ºñ¿ë, Àη µî Á¦ÇÑµÈ ÀÚ¿ø ¿©°ÇÀ¸·Î ÀÎÇØ ÀΰøÁö´É ¿¡ÀÌÀüÆ® °³¹ßÀ» À§ÇÑ ÅõÀÚ¸¦ ÁøÇàÇÏ´Â µ¥ ¾î·Á¿òÀ» °Þ°í ÀÖ´Ù. µû¶ó¼ º» Ç¥ÁØ¿¡¼´Â ÀÌ·¯ÇÑ ¹®Á¦¸¦ ÇØ°áÇÏ°í ÀÌÁ¾ ȯ°æ ¹× ´Ù¾çÇÑ °ÔÀÓ À帣¿¡¼ ¹ü¿ëÀûÀ¸·Î »ç¿ëµÉ ¼ö ÀÖ´Â ¸ðÀÇ È¯°æ ±â¹ÝÀÇ ÀΰøÁö´É °ÔÀÓ ¿¡ÀÌÀüÆ® »ý¼º ¿ä±¸»çÇ×À» Á¦°øÇÏ°í ´ëÇ¥ÀûÀÎ À¯½ºÄÉÀ̽º¸¦ ÅëÇØ È°¿ë ¹× È®Àå °¡´É¼ºÀ» º¸ÀδÙ. |
||
¿µ¹®³»¿ë¿ä¾à | Recently, interest in artificial intelligence agent generation technology using machine learning is increasing, and in particular, research is being actively conducted in various fields related to the 4th industrial revolution. However, since there are many limitations such as cost and time in progressing learning in the real world, the trend of using simulators is also increasing.
Similarly, the game is relatively easy to collect information from the environment and can appropriately adjust the reward according to the agent's actions, so it is highly utilized as a simulator for solving complex problems in the real world. However, in the case of recent games, since various scenarios and functions are required to be implemented to improve user satisfaction, the complexity of the game itself is also increasing day by day. Therefore, game developers are struggling to make investments for AI agent development due to limited resource conditions such as time, cost, and manpower. Therefore, this standard solves these problems and provide requirements for generating artificial intelligence game agents based on simulated environments that can be used universally in heterogeneous environments and various game genres, and shows the possibility of utilization and expansion through representative use cases. |
||
±¹Á¦Ç¥ÁØ | |||
°ü·ÃÆÄÀÏ | TTAK.KO-10.1340.pdf |