• Àüü
  • ÀüÀÚ/Àü±â
  • Åë½Å
  • ÄÄÇ»ÅÍ
´Ý±â

»çÀÌÆ®¸Ê

Loading..

Please wait....

±¹³» ÇÐȸÁö

Ȩ Ȩ > ¿¬±¸¹®Çå > ±¹³» ÇÐȸÁö > µ¥ÀÌÅͺ£À̽º ¿¬±¸È¸Áö(SIGDB)

µ¥ÀÌÅͺ£À̽º ¿¬±¸È¸Áö(SIGDB)

Current Result Document :

ÇѱÛÁ¦¸ñ(Korean Title) µµ·Î ³×Æ®¿öÅ©¿¡¼­ ·£µå¸¶Å© ´ÙÂ÷¿ø ôµµ¹ýÀ» ÀÌ¿ëÇÑ È¿À²ÀûÀÎ M-Æ®¸® ´ë·®ÀûÀç ¾Ë°í¸®Áò
¿µ¹®Á¦¸ñ(English Title) Efficient M-tree Bulk Loading Algorithm in Road Networks Using Landmark Multidimensional Scaling
ÀúÀÚ(Author) ¿ø¼ÒÇö   ±ÇÇõÀ±   Sohyun Won   Hyuk-Yoon Kwon   ³ë¿õ±â   Woong-Kee Loh  
¿ø¹®¼ö·Ïó(Citation) VOL 36 NO. 03 PP. 0091 ~ 0104 (2020. 12)
Çѱ۳»¿ë
(Korean Abstract)
º» ¿¬±¸¿¡¼­´Â µµ·Î ³×Æ®¿öÅ©¸¦ À§ÇÑ M-Æ®¸® ´ë·®ÀûÀç(bulk loading) ¾Ë°í¸®ÁòÀ» Á¦¾ÈÇÑ´Ù. µµ·Î ³×Æ®¿ö Å©´Â ¸Å¿ì µ¿ÀûÀ̸ç, Â÷·®ÀÇ À̵¿, ¿¹ÃøÇÏÁö ¸øÇÑ °ø»ç¿Í »ç°í µîÀ¸·Î ÀÎÇÏ¿© ±³Åë »óȲÀÌ ²÷ÀÓ¾øÀÌ º¯È­ÇÑ´Ù. µµ·Î ³×Æ®¿öÅ© ÀÀ¿ë¿¡¼­´Â ÀÌ·¯ÇÑ º¯È­¸¦ ¹Ý¿µÇϵµ·Ï À妽º¸¦ ÁÖ±âÀûÀ¸·Î À籸¼ºÇÏ´Â °ÍÀÌ ÇʼöÀûÀÌ´Ù. °¢ °¢ÀÇ º¯È­¿¡ ´ëÇÏ¿© ±âÁ¸ÀÇ À妽º¸¦ ¼öÁ¤Çϱ⺸´Ù´Â Àüü µµ·Î ³×Æ®¿öÅ© µ¥ÀÌÅͼ¿¡ ´ëÇÑ »õ·Î¿î À妽º¸¦ ºü¸£°Ô À籸¼ºÇÏ´Â ´ë·®ÀûÀç°¡ È¿À²ÀûÀÌ´Ù. ±âÁ¸ÀÇ M-Æ®¸® ´ë·®ÀûÀç ¾Ë°í¸®ÁòÀº ¡®ºñ½Ñ¡¯ ÃÖ´Ü°æ·Î °Å¸® °è»ê °ú µð½ºÅ© ÆäÀÌÁö ¿¢¼¼½º¸¦ °úµµÇÏ°Ô ¼öÇàÇÏ¿© ÃæºÐÇÑ ¼º´ÉÀ» °ÅµÎÁö ¸øÇÏ¿´´Ù. º» ¿¬±¸¿¡¼­ Á¦¾ÈÇÏ´Â M-Æ® ¸® ´ë·®ÀûÀç ¾Ë°í¸®ÁòÀº ·£µå¸¶Å© ´ÙÂ÷¿ø ôµµ¹ý(Landmark Multidimensional Scaling, LMDS)À» ÀÌ¿ëÇÏ ¿© ÃÖ´Ü°æ·Î °Å¸® °è»êÀ» ´ëÆø ÁÙÀδÙ. ¶ÇÇÑ, °¢ ³ëµå¸¦ Çѹø¾¿¸¸ µð½ºÅ©¿¡ ÀúÀåÇÏ°í ´õ ÀÌ»ó ÀÐÁö ¾ÊÀ½À¸·Î ½á µð½ºÅ© ¾×¼¼½º Ƚ¼ö¸¦ Å©°Ô ÁÙ¿´´Ù. ½ÇÇè °á°ú, Á¦¾ÈµÈ ¾Ë°í¸®ÁòÀº ±âÁ¸ÀÇ ¾Ë°í¸®Áò¿¡ ºñÇÏ¿© ´ÜÀÏ ¾²·¹µå ¹öÀüÀº ÃÖ´ë 21.3¹è, ´ÙÁß ¾²·¹µå ¹öÀü(µ¿½Ã ¾²·¹µå 64°³)Àº ÃÖ´ë 106¹è±îÁö M-Æ®¸® »ý¼º ¼º´ÉÀÌ Çâ»óµÇ¾ú°í, Á¦¾ÈµÈ ¾Ë°í¸®ÁòÀ¸·Î »ý¼ºÇÑ M-Æ®¸®¸¦ ÀÌ¿ëÇÑ k-ÃÖ±ÙÁ¢ °´Ã¼(k-nearest neighbor) °Ë»öÀÇ ¼º´ÉÀÌ ÃÖ´ë 1.22 ¹è±îÁö Çâ»óµÇ¾ú´Ù.
¿µ¹®³»¿ë
(English Abstract)
In this study, we propose an algorithm for M-tree bulk loading in road networks. Road networks are highly dynamic; traffic situation changes constantly owing to vehicle movements and unpredicted constructions and accidents. Thus, it is crucial for road network applications to periodically reorganize the index to fully reflect the changes. In such a circumstance, rather than modifying the existing index for each of the changes, it is more efficient to quickly bulk load a new index for the entire road network. The previous M-tree bulk loading algorithms heavily conduct ¡®expensive¡¯ shortest-path distance computations and disk page accesses, thereby degrading their performances. The algorithm proposed in this study reduces the number of shortest-path distance computations using landmark multidimensional scaling (LMDS). In addition, since our algorithm stores M-tree nodes in disk and does not access them again, the number of disk page accesses is also dramatically reduced. Experimental results demonstrated that our algorithm outperformed the previous algorithm by up to 21.3 times (single-thread version) and 106 times (multi-thread version with 64 concurrent threads) for M-tree construction and that the performance of the k-nearest neighbor (k-NN) search using the M-tree built by our algorithm was also improved by up to 1.22 times.
Å°¿öµå(Keyword) ÅؽºÆ® À¯»çµµ ÃøÁ¤   ºñÁ¤Çü µ¥ÀÌÅÍ ºÐ¼®   SNS »ç¿ëÀÚ À§Ä¡ ¿¹Ãø   Text similarity measurement   Unstructured data analysis   SNS user location prediction   M-Æ®¸®   ´ë·®ÀûÀç   µµ·Î ³×Æ®¿öÅ©   ·£µå¸¶Å© ´ÙÂ÷¿ø ôµµ¹ý   M-tree   bulk loading   road network   landmark multidimensional scaling  
ÆÄÀÏ÷ºÎ PDF ´Ù¿î·Îµå