½üÈÕ£¬ÊýѧϵÐì×˽ÌÊÚÍŶӼ°ÆäºÏ×÷ÕßÔÚ·Ç͹¼«Ó×¼«´óÓÅ»¯ÎÊÌâµÄÓÅ»¯Ëã·¨ÁìÓò»ñµÃ³ÁÒª½øÕ¹£¬ÓйØ×êÑгɾÍÒÔ¡°A unified single-loop alternating gradient projection algorithm for nonconvex¨Cconcave and convex¨Cnonconcave minimax problems¡±ÎªÌâ°ä·¢ÓÚ¹ú¼ÊÔ˳ïÓÅ»¯¶¥¼¶ÆÚ¿¯¡¶Mathematical Programming¡·(ÖпÆÔºÒ»ÇøTop£¬ÖйúÊýѧ»áT1ÆÚ¿¯£¬Ô˳ïÓÅ»¯ÁìÓò¹ú¼ÊÈý´ó¶¥¼¶ÆÚ¿¯Ö®Ò») ¡£¸ÃÂÛÎÄÐì×˽ÌÊÚΪµÚÒ»×÷Õߣ¬Ð±¦GGΪµÚÒ»ÊðÃûµ¥Ôª ¡£
2021¼¶²©Ê¿ÉúÕÅ»ÛÁéΪ¸ÃÂÛÎĵڶþ×÷Õߣ¬²¢Æ¾´Ë¹¤×÷ÓÚ2023Äê4ÔÂ9ÈÕ»ñµÃ2023ÄêÏæÓ廦Ô˳ïѧ½áºÏÄê»áôßѧÊõ»¥»»´ó»áÓÅÁ¼×êÑÐÉúÂÛÎÄ£¨²©Ê¿×êÑÐÉú×飩һµÈ½± ¡£
·Ç͹¼«Ó×¼«´óÎÊÌâµÄÀíÂÛ¡¢Ëã·¨ºÍÀûÓÃÊôÓÚÊý¾Ý¿ÆѧºÍÈËΪÖÇÄÜÖеÄÓÅ»¯Ä£ÐÍ¡¢Ëã·¨Éè¼ÆÓë·ÖÎö×êÑÐÁìÓò£¬ÕâÊÇ 2022 Äê¹ú¶ÈÌìÈ»¿Æѧ»ù½ð¡°Ê®ËÄÎå·¢Õ¹¹æ»®Öеġ±ÓÅÏÈ·¢Õ¹ÁìÓò ¡£·Ç͹¼«Ó×¼«´óÎÊÌâͨ³£ÊÇNP-ÄÑµÄ ¡£½üÆÚ£¬¸ÃÀàµÄÓÅ»¯Ëã·¨¼°¸´ÔӶȷÖÎö£¬³ÉΪÓÅ»¯ºÍ»úе½ø½¨¡¢ÈËΪÖÇÄܵȽ»²æÁìÓò¹ú¼Ê×êÑеÄÇ°ÑغÍÈȵãÎÊÌâ ¡£±¾Ï×÷Ìá³öÁËÇó½â·Ç͹-°¼ºÍ͹-·Ç°¼¼«Ó×¼«´óÓÅ»¯ÎÊÌâµÄÒ»ÖÖÒ»Öµĵ¥Ñ»·½»ÌæÌݶÈͶӰ£¨AGP£©Ëã·¨£¬Ã¿Ò»²½µü´ú½ö±ØÒªÍÆËãÒ»¸öÌݶÈͶӰ²½ ¡£±¾¹¤×÷Ö¤ÁËÈ»¸ÃËã·¨Çó½â·Ç͹-Ç¿°¼»òÕßÊÇǿ͹-·Ç°¼µÄ¼«Ó×¼«´óÓÅ»¯ÎÊÌâµÃµ½Ö¸±êº¯ÊýÒ»½×½üËƲ»±äµãµÄµü´ú¸´ÔÓ¶ÈÊÇ-2½×µÄ£¬¶øÇó½âͨ³£»¯µÄ·Ç͹-°¼»òÕß͹-·Ç°¼¼«Ó×¼«´óÎÊÌâµÄµü´ú¸´ÔÓ¶ÈÊÇ-4½×µÄ ¡£Ä¿Ç°£¬¶ÔÓÚͨ³£»¯µÄ£¨Ç¿£©Í¹-·Ç°¼¼«Ó×¼«´óÎÊÌ⣬ÕâÊǵÚÒ»¸öÓµÓеü´ú¸´ÔӶȱ£ÕϵÄËã·¨ ¡£±¾¹¤×÷»¹Ìá³öÁËÇó½â¸üͨ³£»¯µÄ¶à¿é·Ç¹â»¬·Ç͹-£¨Ç¿£©°¼ºÍ£¨Ç¿£©Í¹-·Ç͹¼«Ó×¼«´óÎÊÌâµÄ¿é½»Ìæ½ü¶ËÌݶȣ¨BAPG£©Ëã·¨£¬ÇÒÖ¤ÁËÈ»ËÄÖÖÇ龰ϸÃËã·¨µÄÀàËƵü´ú¸´ÔÓ¶È ¡£ÊýÖµ³¢ÊÔÁ˾ÖÒ²½²ÁËÈ»Ëã·¨µÄÓÐЧÐÔ ¡£¸Ã×êÑÐÎÞÂÛ´ÓËã·¨»¹ÊÇÀíÂ۵ĽǶÈÔ´´ÐÔ¶¼ºÜÇ¿£¬Íƶ¯ÁËÓÅ»¯Ëã·¨ÔÚ»úе½ø½¨µÈÁìÓòµÄ·¢Õ¹ ¡£
½üЩÄêÀ´£¬Ðì×˽ÌÊÚÍŶÓÔÚ×îÓÅÀíÂÛÓë²½Öè¼°ÆäÔÚ»úе½ø½¨µÈÁìÓòµÄÀûÓ÷½Ãæ×ö³öÁ˺öഴÐÂÐԵŤ×÷£¬ÔÚSIAM Journal on Optimization¡¢IEEE Journal on Selected Areas in Communications¡¢Journal of Global Optimization¡¢Journal of Optimization Theory and Applications¡¢Computational Optimization and ApplicationsµÈ¹ú¼ÊȨÍþÆÚ¿¯ÉÏ°ä·¢ÂÛÎÄ30Óàƪ£¬×êÑгɾ͵õ½Ó¢¹ú»Ê¼Ò¹¤³ÌԺԺʿ¡¢ÐÙÑÀÀû¿ÆѧԺ±í¼®ÔºÊ¿L.Hanzo½ÌÊÚ£¬¼ÓÄôó»Ê¼Ò¿ÆѧԺ¡¢¹¤³ÌÔºÁ½ÔºÔºÊ¿J. Pei½ÌÊÚ£¬¼ÓÄôó»Ê¼Ò¿ÆѧԺԺʿZ.-Q. Luo½ÌÊÚ£¬INFORMsÖ÷ϯ¡¢¹ú¼Ê¶¥¼âÆÚ¿¯Math. Prog. A.¹²Í¬Ö÷±àA.Atamturk½ÌÊڵȹú¼Ê³ÛÃûר¼ÒµÄ¹«¿ªÒýÓúÍÕýÃæÆÀ¼Û ¡£ÒòÔÚÔ˳ïÓÅ»¯ÁìÓòµÄ͹Æð¿ÆÑгɾͣ¬2020ÄêÐì×˽ÌÊÚÈÙ»ñÖйúÔ˳ïѧ»áÇàÄê¿Æ¼¼½±(È«¹ú5ÈË) ¡£
±¾¹¤×÷µÃµ½¹ú¶ÈÌìÈ»¿Æѧ»ù½ðºÍÉϺ£ÊÐÌìÈ»¿Æѧ»ù½ðÖ§³Ö ¡£±¾Ï×÷ÓÉб¦GGÐì×˽ÌÊÚ¡¢²©Ê¿ÉúÕÅ»ÛÁ顢˶ʿÉúÐìÑó¡¢ÃÀ¹ú×ôÖÎÑÇÀí¹¤Ñ§ÔºGuanghui Lan½ÌÊÚºÏ×÷ʵÏÖ£¬ÓйØÂÛÎļû£ºZi Xu, Huiling Zhang, Yang Xu, Guanghui Lan. A unified single-loop alternating gradient projection algorithm for nonconvex¨Cconcave and convex¨Cnonconcave minimax problems. Mathematical Programming, (2023).https://doi.org/10.1007/s10107-022-01919-z
ÎÄÕÂÁ´½Ó£ºhttps://doi.org/10.1007/s10107-022-01919-z.
