ÉϺ£ÖÎÀíÂÛ̳µÚ399ÆÚ£¨ÂÞ¿¡¸±½ÌÊÚ£¬ÉϺ£½»Í¨´óѧ°²Ì©¾­¼ÃÓëÖÎÀíѧԺ£©

Ìâ    Ä¿£º»ùÓÚ×ÔÊÊÓ¦²ÉÑù¹æ¶¨µÄÈ«Ðò¹á·¨A Fully Sequential Procedure with Adaptive Sampling Rules

ÑÝ ½² ÈË£ºÂÞ¿¡£¬ÉϺ£½»Í¨´óѧ¸±½ÌÊÚ£¨³¤Æ¸£©

Ö÷ ³Ö ÈË£ºÕòè´£¬Ð±¦GGÖÎÀíѧԺ½ÌÊÚ

ʱ    ¼ä£º2019Äê9ÔÂ25ÈÕ£¨ÖÜÈý£©£¬ÉÏÎç10:30

µØ    µã£ºÐ£±¾²¿¶«ÇøÖÎÀíѧԺʵ477ÊÒ

Ö÷°ìµ¥Ôª£ºÐ±¦GGÖÎÀíѧԺ¡¢Ð±¦GGÖÎÀíѧԺÇàÀÏ´óʦÁªÒê»á

Ñݽ²È˼ò½é£º

ÂÞ¿¡£¬ÉϺ£½»Í¨´óѧ°²Ì©¾­¼ÃÓëÖÎÀíѧԺ¸±½ÌÊÚ£¨³¤Æ¸£©£¬²©Ê¿Éúµ¼Ê¦
¡£2013Äê±ÏÒµÓÚÏã¸Û¿Æ¼¼´óѧ£¬»ñµÃ¹¤Òµ¹¤³ÌÓëÎïÁ÷ÖÎÀí²©Ê¿Ñ§Î»£»2009Äê±ÏÒµÓÚÄϾ©´óѧÊýѧϵ£¬»ñµÃͳ¼Æѧѧʿѧλ
¡£ÖØÒª×êÑз½ÏòÔ̺¬Ëæ»ú½¨Ä£¡¢·ÂÕæÓÅ»¯¡¢Êý¾Ý·ÖÎö£¬ÒÔ¼°ËüÃÇÔÚ´óÐÍ·þÎñϵͳ¡¢½¡È«Ò½ÁÆÖÎÀíºÍ½ðÈÚ·çÏÕÖÎÀíµÈ·½ÃæµÄÀûÓÃ
¡£Ö÷³Ö¶àÏî¿ÆÑлù½ðÏîÄ¿£¬Ô̺¬¹ú¶ÈÓÅÁ¼ÇàÄê¿Æѧ»ù½ðÏîÄ¿£¬¹ú¶ÈÇàÄê¿Æѧ»ù½ðÏîÄ¿£¬ÉϺ£ÊнÌί¡°³¿êØ´òË㡱ÏîÄ¿µÈ
¡£ÔÚ¡¶Operations Research¡·£¬¡¶INFORMS Journal on Computing¡·ºÍ¡¶Naval Research Logistics¡·µÈ¹ú¼ÊÆÚ¿¯ÉÏ°ä·¢¶àƪѧÊõÂÛÎÄ
¡£

Ñݽ²ÄÚÈݼò½é£º

Selecting the best system design from a finite set of alternatives is known as ranking-and-selection (R&S) in the simulation literature. Many procedures, from either frequentist or Bayesian approaches, has been designed in order to solve R&S problems more effectively or efficiently. Typically, frequentist procedures emphasize more on the effectiveness of a statistical guarantee while Bayesian procedures focus more on the efficiency of using a small number of total samples. In this paper, we aim to take both the effectiveness and efficiency into consideration, from the frequentist point of view. In particular, we design a fully sequential procedure with an adaptive sampling rule, which provide a probabilistic guarantee of correct selection in an asymptotic sense. We demonstrated both the effectiveness and efficiency of our proposed procedure by comparing with KN and OCBA, two classical procedures in frequentist and Bayesian frameworks, through extensive numerical experiments.

Ó­½Ó¿í´óʦÉú²ÎÓ룡