Á¤º¸Ã³¸®ÇÐȸ ³í¹®Áö B
Current Result Document :
ÇѱÛÁ¦¸ñ(Korean Title) |
Ant-Q ÇнÀÀ» ÀÌ¿ëÇÑ Gale-Shapley ¹®Á¦ ÇØ°á¿¡ °üÇÑ ¿¬±¸ |
¿µ¹®Á¦¸ñ(English Title) |
Solving the Gale-Shapley Problem by Ant-Q learning |
ÀúÀÚ(Author) |
񊀔
Á¤ÅÂÃæ
Hyun Kim
TaeChoong Chung
|
¿ø¹®¼ö·Ïó(Citation) |
VOL 18-B NO. 03 PP. 0165 ~ 0172 (2011. 06) |
Çѱ۳»¿ë (Korean Abstract) |
º» ³í¹®¿¡¼´Â »ý¹°ÇÐÀÇ °³¹ÌµéÀÌ ÇнÀÀ» ÅëÇØ ¸ñÇ¥¸¦ ȹµæÇÏ´Â ¹æ¹ýÀ» ÀÀ¿ëÇÑ Ant-Q ¾Ë°í¸®Áò(Ant Q learning System)[1]À» Gale-Shapley[2]¾Ë°í¸®ÁòÀ» ÅëÇØ Á¦½ÃµÇ¾ú´ø ¾ÈÁ¤µÈ °áÈ¥¹®Á¦(SMP: Stable Marriage Problem)[3]ÀÇ »õ·Î¿î ÇعýÀ» ã±â À§ÇØ Àû¿ë ÇÏ¿´´Ù. SMP´Â ³²¼º(mi)µé°ú ¿©¼º(wj)µéÀº °¢ÀÚ ÀÚ½ÅÀÌ ÁÁ¾ÆÇÏ´Â ÀÌ»óÇü¿¡ ´ëÇÑ ¼±È£µµ(PL: preference list)¸¦ ¹ÙÅÁÀ¸·Î ¾ÈÁ¤ÀûÀ̸鼵µ ÃÖ¼±ÀÇ Â¦À» ã´Â °ÍÀ» ¸ñÇ¥·Î ÇÏ°í ÀÖ´Ù. Gale-Shapley ¾Ë°í¸®ÁòÀº ³²¼º(ȤÀº ¿©¼º) À§ÁÖ·Î ¾ÈÁ¤Àû(stability)ÀΠ¦(Matching)À» ¼º»ç½ÃÅ°¹Ç·Î ´Ù¾çÇÑ Á¶°ÇÀ» ¼ö¿ëÇÏÁö ¸øÇÑ´Ù.
º» ³í¹®¿¡ Àû¿ëµÈ Ant-Q´Â °³¹Ì(Ant)ÀÇ Æä·Î¸óÀ» È°¿ëÇÑ ÇнÀÀÎ ACS(Ant colony system)¿¡ °ÈÇнÀÀÇ ÀÏÁ¾ÀÎ Q-ÇнÀ[9]À» Ãß°¡ÇÑ ¹æ¹ýÀ¸·Î, SMPÀÇ »õ·Î¿î ÇعýÀ» ãÀ» ¼ö ÀÖ¾ú´Ù. |
¿µ¹®³»¿ë (English Abstract) |
In this paper, we propose Ant-Q learning Algorithm[1], which uses the habits of biological ants, to find a new way to solve Stable Marriage Problem(SMP)[3] presented by Gale-Shapley[2]. The issue of SMP is to find optimum matching for a stable marriage based on their preference lists (PL). The problem of Gale-Shapley algorithm is to get a stable matching for only male (or female). We propose other way to satisfy various requirements for SMP.
ACS(Ant colony system) is an swarm intelligence method to find optimal solution by using phermone of ants. We try to improve ACS technique by adding Q learning[9] concept. This Ant-Q method can solve SMP problem for various requirements. The experiment results shows the proposed method is good for the problem. |
Å°¿öµå(Keyword) |
¾ÈÁ¤µÈ °áÈ¥¹®Á¦
°³¹Ì¾Ë°í¸®Áò
Q-ÇнÀ
Ant-Q
TSP
Stable Marriage Problem
Ant Colony System
Q-learning
|
ÆÄÀÏ÷ºÎ |
PDF ´Ù¿î·Îµå
|