銆婇浕瀛愭妧琛撴噳鐢ㄣ
鎮ㄦ墍鍦ㄧ殑浣嶇疆锛棣栭爜 > 宓屽叆寮忔妧琛 > 瑷▓鎳夌敤 > 鍩轰簬澶氬堡鎰熺煡姗熶唬鐞嗘ā寮忕殑鍦扮悆绯荤当妯″紡鐗╃悊鍙冩暩鍎寲鏂规硶
鍩轰簬澶氬堡鎰熺煡姗熶唬鐞嗘ā寮忕殑鍦扮悆绯荤当妯″紡鐗╃悊鍙冩暩鍎寲鏂规硶
2019骞撮浕瀛愭妧琛撴噳鐢ㄧ8鏈
鍚 鍒1锛岄粌 娆2锛岃枦 宸1
1.娓呰彲澶у 瑷堢畻姗熺瀛歌垏鎶琛撶郴锛屽寳浜100084锛2.鍖椾簽鍒╂閭eぇ瀛 鐢熸佺郴绲辩瀛歌垏绀炬渻涓績锛屽紬鎷夋牸鏂澶86011
鎽樿锛 鍦扮悆绯荤当妯″紡涓墿鐞嗗弮鏁哥殑涓嶇‘鎬ф渻灏嶆埃鍊欐ā鎷熺殑绮惧害鐢g敓宸ㄥぇ鐨勫奖闊匡紝鍎寲鐗╃悊鍙冩暩灏嶆彁楂樻埃鍊欓爯娓殑婧栫‘鎬ц嚦闂滈噸瑕併傞氬父鍦ㄥ湴鐞冪郴绲辨ā寮忕殑鍙冩暩鍎寲涓湁澶氬嬬洰鏍囬渶瑕佸悓鏅傚劒鍖栵紝鐒惰岀洰鍓嶅父鐢ㄧ殑閫插寲澶氱洰鏍囩畻娉曞湪鍦扮悆绯荤当妯″紡涓婁娇鐢ㄩ渶瑕佹サ楂樼殑瑷堢畻浠e児锛屽洜姝ゆ彁鍑轰簡涓绋熀浜庡灞ゆ劅鐭ユ(MLP)绁炵稉缍茶矾鐨勫鐩爣浠g悊妯″紡鍙冩暩鍎寲鏂规硶MO-ANN銆傛鏂规硶鍒╃敤澶氬堡鎰熺煡姗熷缓绔嬩唬鐞嗘ā寮忥紝鐢ㄤ唬鐞嗘ā寮忎締闋愪及鍊欓伕閲囨ǎ榛炵殑鍎姡锛屾彁楂樹簡澶氱洰鏍囧劒鍖栫殑绮惧害鍜屾敹鏂傛с傚湪瑜囬洔鏁稿鍑芥暩鍜屽柈鏌卞ぇ姘fā寮忎笂鐨勫皪姣斿椹楄〃鏄庯紝MO-ANN鍎寲绠楁硶鐩稿皪浜庨插寲澶氱洰鏍囩畻娉曞叿鏈夋槑椤劒鍕紝鐗瑰埆鏄湪鐔卞付鏆栨睜-鍦嬮殯闆插椹楃殑鍠煴澶ф埃妯″紡涓紝MO-ANN鏀舵杺閫熷害鍙浉灏峃SGAIII鎻愬崌5鍊嶄互涓娿
涓湒鍒嗛鍙凤細 TP302.1锛汿P183
鏂囩嵒鏍囪瓨纰硷細 A
DOI锛10.16157/j.issn.0258-7998.190606
涓枃寮曠敤鏍煎紡锛 鍚冲埄锛岄粌娆o紝钖涘穽. 鍩轰簬澶氬堡鎰熺煡姗熶唬鐞嗘ā寮忕殑鍦扮悆绯荤当妯″紡鐗╃悊鍙冩暩鍎寲鏂规硶[J].闆诲瓙鎶琛撴噳鐢紝2019锛45(8)锛99-103.
鑻辨枃寮曠敤鏍煎紡锛 Wu Li锛孒uang Xin锛孹ue Wei. Physical parameter optimization method for earth system model based on multi-layer perceptron surrogate model[J]. Application of Electronic Technique锛2019锛45(8)锛99-103.
Physical parameter optimization method for earth system model based on multi-layer perceptron surrogate model
Wu Li1锛孒uang Xin2锛孹ue Wei1
1.Department of Computer Science and Technology锛孴singhua University锛孊eijing 100084锛孋hina锛 2.Center for Ecosystem Science and Society锛孨orthern Arizona University锛孎lagstaff 86011锛孶SA
Abstract锛 The uncertainty of physical parameters in earth system models has a huge impact on the performance of climate simulations. Tuning physical parameters is critical to improving the accuracy of climate predictions. Usually, in the parameter optimization of earth system model, there are multiple objectives that need to be optimized simultaneously. However, the commonly used multi-objective evolutionary algorithms require a very high computational cost for tuning earth system models. Therefore, this paper proposes a multi-objective parameter optimization method MO-ANN based on multi-layer perceptron(MLP) neural network and surrogate model. This method uses a multi-layer perceptron to build a surrogate model to improve the accuracy and convergence of multi-objective optimization. Comparative experiments on complex mathematical functions and single-column atmospheric models show that the MO-ANN optimization algorithm has obvious advantages over the evolutionary multi-objective algorithms. With the warm pool-International Cloud Experiment(TWP-ICE) single column atmospheric model, the convergence rate of the proposed multi-objective optimization method can be improved by more than 5 times compared with the known NSGAIII method.
Key words : parameter optimization锛沵ultilayer perceptron锛沵ulti-objective optimization锛沞arth system model

0 寮曡█

    杩戝勾渚嗭紝闅忕潃妤电姘e欎簨浠剁殑闋荤箒鍑虹従锛屾埃鍊欓爯娓晱椤屽彈鍒拌秺渚嗚秺澶氱殑闂滄敞銆鍦扮悆绯荤当妯″紡閫氶亷妲嬪缓涓绯诲垪鐨勬暩瀛哥墿鐞嗘柟绋嬬祫渚嗚繎浼兼ā鎷熺湡瀵﹀湴鐞冪郴绲变腑瀛樺湪鐨勭墿鐞嗐佸寲瀛稿拰鍕曞姏绛夎闆滅殑閬庣▼锛屾槸闋愭脯鏈締姘e欑殑閲嶈宸ュ叿[1]銆傜劧鑰屽湴鐞冪郴绲辨ā寮忕殑鐗╃悊鍙冩暩鍖栨柟妗堜腑瀛樺湪澶ч噺鐨勪笉纭畾鍙冩暩锛岄欎簺鍙冩暩灏嶅湴鐞冪郴绲辨ā寮忕殑妯℃嫙鎬ц兘鏈夊緢澶х殑褰遍熆[2]銆傚劒鍖栫畻娉曟槸涓绋噺鍖栦笉纭畾鎬с佹牎婧栦笉纭畾鍙冩暩鐨勬湁鏁堟柟娉曘傞氬父鍦扮悆绯荤当妯″紡妯℃嫙鐨勭洰鏍囨湁鍏╁嬩互涓婏紝濡傞檷姘淬佹韩搴﹀拰婵曞害绛夊皪姘e欓爯娓潪甯搁噸瑕佺殑澶氬嬭畩閲忋傚彟澶栧湪鍦扮悆绯荤当妯″紡涓湁寰堝姘e欑従璞★紝渚嬪鐔卞付澶ф埃瀛g瘈鍐呴渿钑(MJO)銆佹澅浜炲棰(EASM)銆佸巹鐖惧凹璜/鍗楁柟婵ゅ嫊(ENSO)绛夛紝閫欎簺鍒嗗埆鏄爆浜庝笉鍚岀墿鐞嗘剰缇╃殑姘e欑従璞★紝灏嶄汉椤炵殑鐢熺敚鐢熸椿閮芥湁鐫閲嶈鐨勫奖闊匡紝鍥犳灏嶆埃鍊欓爯娓崄鍒嗛噸瑕併傝嫢瑕佷娇寰楀湴鐞冪郴绲辨ā寮忎腑鐨勫鍊嬫埃鍊欒绱犻兘鐩″彲鑳藉緱鍒板劒鍖栵紝姝ゆ檪闇瑕佸埄鐢澶氱洰鏍囧劒鍖渚嗗皪涓嶇‘瀹氬弮鏁搁茶鍎寲銆

    甯哥敤鐨勫鐩爣鍎寲绠楁硶鏄插寲澶氱洰鏍囩畻娉曪紝渚嬪NSGAIII銆丮OPSO銆丮OEA-D绛夌畻娉曪紝鍏朵腑NSGAIII鍜孧OEA-D绠楁硶鎳夌敤闈炲父寤f硾锛屼笖NSGAIII鏄洰鍓嶅叕瑾嶇殑鏈鐐洪珮鏁堢殑澶氱洰鏍囧劒鍖栨柟娉曚箣涓[3-4]銆傞插寲澶氱洰鏍囩畻娉曢氬父鍏ㄥ眬鎬ц純濂戒絾鏄渶瑕佺殑鐤婁唬姝ユ暩杓冨锛岄欏湪鍦扮悆绯荤当妯″紡涓剰鍛崇潃澶氭鐨勬ā寮忛亱琛屻傝屽湴鐞冪郴绲辨ā寮忛亱琛屼竴娆¢兘闇瑕佹サ楂樼殑瑷堢畻浠e児锛屾垚鐧句笂鍗冩鐨勬ā鎷熻▓绠楁垚鏈洿涓嶅彲鎺ュ彈[5]銆傚洜姝ゅ湪鍦扮悆绯荤当妯″紡涓ラ渶楂樻晥銆佹敹鏂傚揩鐨勫鐩爣鍎寲绠楁硶銆

    鏈枃绲愬悎鍌崇当浠g悊妯″紡鍎寲鎬濇兂鎻愬嚭浜嗕竴绋熀浜澶氬堡鎰熺煡姗绁炵稉缍茬怠浠g悊妯″紡鐨勫鐩爣鍎寲鏂规硶銆傛鏂规硶鍒╃敤澶氬堡鎰熺煡姗熶綔鐐烘浛浠g湡瀵︽ā寮忕殑浠g悊鍥炴妯″瀷锛岀敤姝や唬鐞嗘ā鍨嬩締浼拌▓鍎寲鍙冩暩鐨勯伕鍙栥傚湪寤虹珛鍒濆妯″瀷涔嬪緦锛屾瘡澶氫竴鍊嬮噰妯i粸姝ゆ柟娉曢兘鏇存柊鍎寲绛栫暐锛岀浉姣旈插寲绠楁硶鐨勭ó缇ゆ洿鏂扮瓥鐣ワ紝瀹冭兘澶犳湁鏁堝湴鎻愰珮绠楁硶鐨勬敹鏂傛с

1 澶氱洰鏍囦唬鐞嗘ā寮忓劒鍖栨柟娉

1.1 鍩轰簬澶氬堡鎰熺煡姗熶唬鐞嗘ā寮鍙冩暩鍎寲鏂规硶绺介珨鎬濊矾

    浠g悊妯″紡鍎寲鏂规硶鐨勭附楂旀濊矾濡備笅锛氶鍏堝埄鐢ㄧ暥鍓嶆墍鏈夋ǎ鏈寤轰竴鍊嬬当瑷堝洖姝告ā鍨嬬ū鐐轰唬鐞嗘ā寮忥紝鐒跺緦鍒╃敤浠g悊妯″紡鍘讳及瑷堜笅涓鍊嬫洿鍎弮鏁哥殑浣嶇疆锛涘皢姝よ檿寰楀埌鐨勬渶鍎弮鏁镐唬鍏ョ湡瀵︽ā寮忎腑閬嬭锛岀嵅寰楁柊鐨勭湡瀵︾殑閲囨ǎ榛烇紝灏嗘閲囨ǎ榛炲姞鍏ュ師鏈夋ǎ鏈紝涓璧锋寤烘柊鐨勪唬鐞嗘ā寮忋傚姝ゅ弽瑜囩枈浠o紝鐩村埌鏂伴噰妯i粸婊胯冻鍎寲姊濅欢銆

    闆栫劧鍩轰簬鍏嬭!閲(Kriging)鍜屽緫鍚戝熀鍑芥暩(RBF)浠g悊妯″紡鐨勫柈鐩爣鍎寲绠楁硶琚ぇ閲忕敤浜庡湪瑜囬洔鍟忛鐨勫劒鍖栨眰瑙[6]锛岀劧鑰岀洰鍓嶉潰鍚戝鐩爣鐨勪唬鐞嗘ā寮忓劒鍖栫畻娉曠殑鐮旂┒閭勮純鐐虹己涔忋傚彟澶栧湪鍌崇当鐨勪唬鐞嗘ā寮忓劒鍖栦腑锛屼唬鐞嗘ā鍨嬬殑閬告搰閫氬父鐐虹当瑷堝洖姝告柟娉曪紝閫欎簺鏂规硶灏嶈闆滃宄扮殑鍦扮悆绯荤当妯″紡鐨勪唬鐞嗙簿搴︿笉瓒筹紝閫欏皢鏈冨皫绶诲熀浜庝唬鐞嗘ā寮忕殑鏈鍎弮鏁镐及瑷堜笉澶犵簿婧栵紝閫蹭竴姝ュ奖闊垮劒鍖栫畻娉曠殑绮惧害銆傛湰鏂囬嚌灏嶄互涓婂晱椤屾彁鍑轰簡鍩轰簬澶氬堡鎰熺煡姗熺殑澶氱洰鏍囧劒鍖栨柟娉(MO-ANN)锛屼富瑕佹濇兂鏈変互涓嬪叐榛烇細(1)鍚屾檪绲愬悎浠g悊妯″紡鍎寲鐨勬濊矾鍜岄插寲澶氱洰鏍囧劒鍖栫畻娉曠殑闈炴敮閰嶈В鎺掑簭绛栫暐銆侻O-ANN姣忎竴娆″皨鎵剧暥鍓嶆渶鍎噰妯i粸鏅傚繀闋堝湪闈炴敮閰嶈В涓皨鎵撅紝閫欐ǎ鐨勯伕鎿囬槻姝㈤伕鍙栧埌鐨勯噰妯i粸鍑虹従涓鍊嬬洰鏍囩壒鍒ソ锛岃屽彟澶栦竴鍊嬬洰鏍囨サ宸儏娉侊紱(2)鍩轰簬澶氬堡鎰熺煡姗熺缍撶恫绲$殑浠g悊妯″紡鑳藉鐩稿皪鍌崇当浠g悊妯″紡鐨勫洖姝告柟娉曟洿濂藉湴瀵︾従澶氬嬭几鍏ュ埌澶氬嬭几鍑虹殑鍥炴鍟忛銆侻O-ANN绠楁硶娴佺▼濡傚湒1鎵绀恒傚叾涓棞閸垫椹熺殑瑙i噵濡備笅锛

jsj1-t1.gif

    (1)閲囨ǎ銆傚垵濮嬮噰妯f湁鍏╁嬮噸瑕佺殑鐩殑锛氬叾涓鏄偤浜嗗垵姝ユ帰绱㈠弮鏁哥┖闁擄紝浣垮緱鍎寲绠楁硶灏嶅弮鏁哥┖闁撴湁涓鍊嬪熀鏈殑浜嗚В锛涘叾浜屾槸鐐轰簡鍒濆鍖栦唬鐞嗘ā寮忋傛湰鏂囬伕鎿囩殑鏄媺涓佽秴绔嬫柟閲囨ǎ锛屽畠鐨勬濇兂鏄湪鍙冩暩绌洪枔涓垎灞ら殢姗熸娊妯c傚洜鐐烘湁浜嗗垎灞ょ殑绛栫暐锛岄噰妯e湪鍙冩暩绌洪枔涓純鐐哄叏闈紝鑳藉灏嗗劒鍖栫畻娉曞缓绔嬪湪涓鍊嬭壇濂界殑鍩虹涓娿

    (2)姹傚緱闈炴敮閰嶈В闆嗗苟鎺掑簭锛屾帓搴忓緦鍙栧緱鐣跺墠鏈鍎潪鏀厤瑙o紝浠ュ倷寰岀簩鐢熸垚鍊欓伕閲囨ǎ榛為泦銆傞潪鏀厤瑙g殑鎺掑簭浣垮緱澶氬嬬洰鏍囪兘琚稖鍚堣冩叜銆

    (3)妲嬪缓鍩轰簬澶氬堡鎰熺煡姗熺殑浠g悊妯″紡銆傛牴鎿氬湴鐞冪郴绲辨ā寮忕殑鍙冩暩鍒版ц兘鐨勮闆滅壒鎬э紝姝ょ畻娉曚腑閬告搰鐨勫缓妯℃柟娉曟槸鍏锋湁鏇村挤闈炵窔鎬ц〃閬旇兘鍔涚殑澶氬堡鎰熺煡姗熺缍撶恫绲°

    (4)闋愪及涓嬩竴鍊嬫渶鍎噰妯i粸銆傛湰鏂囬爯浼颁笅涓鍊嬮噰妯i粸鐨勭瓥鐣ラ噰鐢ㄧ殑鏄枃鐛籟7]涓彁鍒扮殑闅忔鎿惧嫊绛栫暐鐨勬敼閫茬増鏈紝鍏朵富瑕佹濇兂鏄寤哄叐绲勫欓伕閲囨ǎ榛為泦鍚堬紝绗竴绲勭偤鍦ㄧ暥鍓嶇湡瀵︽渶鍎噰妯i粸闄勮繎闅忔鎿惧嫊锛岀浜岀祫鐐哄湪鍏ㄥ弮鏁哥┖闁撲腑鐨勯殢姗熸摼鍕曘傜劧寰屽埄鐢ㄥ叐绋⿻鍍圭浉绲愬悎鐨勬柟娉曞皪鎵鏈夊欓伕閲囨ǎ榛為茶瑭曞児銆傜涓绋⿻鍍规柟娉曟槸鍒╃敤浠g悊妯″紡灏嶅欓伕閲囨ǎ榛為茶浼拌▓锛屼及瑷堢祼鏋滃ソ鐨勶紝琚伕鍙栫殑姗熸渻澶э紱绗簩绋⿻鍍规柟寮忔槸鏍规摎鍊欓伕闆嗕腑鐨勯噰妯i粸鑸囩暥鍓嶅凡鏈夐噰妯i粸鐨勮窛闆締琛¢噺鐨勶紝璺濋洟瓒婅繎锛屾晥鏋滆秺濂姐傚叐绋⿻鍍规柟寮忕浉绲愬悎鐨勬柟娉曚篃鏇村姞鍏ㄩ潰鍦拌 閲忎簡涓鍊嬪欓伕閲囨ǎ榛炵殑濂藉銆

    (5)鍦ㄧ湡瀵︾殑妯″瀷涓婏紙鍦扮悆绯荤当妯″紡涓婏級瑭曚及鏂伴噰妯i粸鐨勭祼鏋溿傛渶寰屽皢閬稿嚭鐨勬渶鍎ǎ鏈粸甯跺叆鐪熷妯″瀷涓亱琛岋紝寰楀埌鐪熷鐨勯噰妯g祼鏋溿傞欎竴姝ュ湪瑜囬洔鍑芥暩涓偤鍑芥暩鍊肩殑瑷堢畻锛屽湪鍦扮悆绯荤当妯″紡涓墖鐐轰竴娆℃ā鍨嬮亱琛屽拰瑭曚及鐨勯亷绋嬨

    (6)灏嗘柊閲囨ǎ榛炲姞鍏ュ凡鏈夋ǎ鏈泦锛岄噸鏂版嫙鍚堜唬鐞嗘ā寮忋傚皢鏈鏂扮殑閲囨ǎ绲愭灉鍔犲叆宸叉湁妯f湰搴紝閲嶆柊妲嬪缓浠g悊妯″紡銆備緷姝ら噸瑜囷紝鐩村埌绠楁硶鏀舵杺鎴栬呮槸閬斿埌瑕忓畾鐨勭枈浠f鏁搞

1.2 澶氬堡鎰熺煡姗熶唬鐞嗘ā寮忕殑瀵︾従

    澶氬堡鎰熺煡姗熺缍撶恫绲℃槸涓绋墠鍚戠殑浜哄伐绁炵稉缍茬怠锛圓rtificial Neural Network锛孉NN锛夛紝鍙互鐪嬫垚涓绲勮几鍏ュ悜閲忓埌涓绲勮几鍑哄悜閲忕殑鏄犲皠銆傚畠鐢辫几鍏ュ堡銆侀殣鍚堡鍜岃几鍑哄堡鍏卞悓妲嬫垚锛屽叾涓櫎浜嗚几鍏ュ堡涔嬪姣忎竴灞ら兘甯舵湁闈炵窔鎬ф縺娲诲嚱鏁革紝浣垮緱妯″瀷鏇磋兘澶犻傛噳闈炵窔鎬ц純寮风殑鐗规ц〃閬斻

    澶氬堡鎰熺煡姗熺缍撶恫绲″湪闈炵窔鎬у洖姝镐笂鍗佸垎鍙楁杩庯紝鐮旂┒琛ㄦ槑澶氬堡鎰熺煡姗熸槸閫氱敤鐨勫嚱鏁搁艰繎鍣紝鐢氳嚦閫傚悎闈炲厜婊戝拰鍒嗘閫g簩鍟忛锛屼笖鍏剁浉灏嶄簬鍌崇当鐨勬鍣ㄥ缈掔畻娉曡岃█锛屽叿鏈夋洿楂樼殑鎷熷悎绮惧害[8]

    鍊煎緱娉ㄦ剰鐨勬槸閫欒!鐨勪唬鐞嗘ā寮忕殑鎷熷悎閬庣▼骞朵笉鏄竴鑸鍣ㄥ缈掓剰缇╀笂灏嬫眰鍋忓樊鍜屾柟宸殑骞宠 鎯呮硜锛岄欒!鍍呭儏灏嗗叾浣滅偤涓鍊嬪洖姝稿櫒渚嗕娇鐢ㄣ侻O-ANN绠楁硶涓瘡澧炲姞涓鍊嬮噰妯i粸閮介噸鏂拌〒绶翠竴閬嶄唬鐞嗘ā寮忥紝姣忎竴娆′唬鐞嗘ā寮忕殑鎷熷悎閬庣▼瑕佺洝閲忔簴纭紝浠ユ湡鏈涗唬鐞嗘ā寮忚兘澶犳洿鍔犵簿纭湴闋愪及涓嬩竴鍊嬫渶鍎噰妯i粸鐨勪綅缃傛湰鏂囩殑鎷熷悎绛栫暐鏄皢澶氬嬭秴鍙冩暩鎺у埗涓嶈畩锛屽埄鐢ㄥ弽鍚戝偝鎾畻娉(BP)澶氭閲嶈瑷撶反鐣跺墠鎵鏈夋ǎ鏈紝浠ュ揩閫熼仈鍒拌姹傜殑鎷熷悎绮惧害銆傛鎷熷悎鏂规硶淇濊瓑浜嗘瘡涓娆$枈浠d腑浠g悊妯″紡妯″瀷鐨勫揩閫熺┅瀹氥

2 瀵﹂瑷▓

    鐐轰簡椹楄瓑涓婅堪澶氱洰鏍囧劒鍖栨柟娉曠殑鏈夋晥鎬э紝鏈枃鍒嗗埆鍦ㄨ闆滄暩瀛稿嚱鏁稿拰鍠煴澶ф埃妯″紡(Single Column Atmosphere Model锛孲CAM)涓婂皢姝ょ畻娉曡垏鍓嶆枃鎻愬埌鐨勬噳鐢ㄥ唬娉涚殑NSGAIII鍜孧OEA-D绠楁硶閫茶浜嗗皪姣旀脯瑭︺傚洜鐐哄湪鐪熷鐨勫湴鐞冪郴绲辨ā寮忎腑甯屾湜浠ョ洝鍙兘灏戠殑妯″紡閬嬭娆℃暩渚嗙‘瀹氭渶鍎弮鏁革紝浠ヤ笅瑭曟瘮鎵閬靛惊鐨勮瀹氭槸灏嶆暩瀛稿嚱鏁哥殑瑷堢畻娆℃暩鍦200娆′互鍐咃紝鍦⊿CAM涓婄殑妯℃嫙娆℃暩鍦600娆′互鍐咃紝鏌ョ湅鍦ㄦ绡勫湇鍐呭悇椤炲劒鍖栫畻娉曠殑琛ㄧ従鎯呮硜銆傚叾涓暩瀛稿嚱鏁哥殑閬告搰鐨勬槸甯哥敤鐨勫鐩爣娓│鍑芥暩ZDT2[9]鍜孌TLZ7[10]

    SCAM鏄敤渚嗘ā鎷熷浐瀹氬湪鏌愬嬬稉绶害鐨勫ぇ姘g墿鐞嗛亷绋嬶紝瀹冩槸鐢辩壒瀹氱殑閭婄晫鍜屽挤杩牬鎵椹呭嫊鐨勶紝鏄皥闁鐐轰簡鐮旂┒鍦扮悆绯荤当妯″紡鐨勭墿鐞嗗弮鏁稿寲鏂规鑰岄枊鐧肩殑宸ュ叿锛屽皪鍦扮悆绯荤当妯″紡鐨勭櫦灞曟湁閲嶈鎰忕京銆傛湰鏂囬伕鎿囩殑鏄啽甯舵殩姹-鍦嬮殯闆插椹(TWP-ICE)[11]鍜屾贩鍚堢浉浣-鍖楁サ闆插椹桵-PACE[12]銆俆WP-ICE瀵﹂鐨勬ā鎷熸檪闁撴槸寰2006骞1鏈18鏃ワ綖2鏈13鏃ャ侻-PACE瀵﹂鐨勬ā鎷熸檪闁撴槸寰2004骞10鏈06鏃ワ綖10鏈22鏃ャ傚叐鍊婼CAM瀵﹂鐨勮娓垎鍒締婧愪簬鍏╁嬬珯榛炵殑鐒$窔闆绘帰绌虹珯鎵鐛插緱鐨勮娓敹鎿氥

    鍏╁婼CAM瀵﹂鐨勫劒鍖栫洰鏍囬兘鏄娇寰楁ā寮忎腑鏈鍙楅棞娉ㄧ殑涓浜涜畩閲忚垏瑙娓殑璺濋洟鏇村姞鎺ヨ繎锛屽叿楂旂殑璁婇噺閬告搰濡傝〃1鎵绀恒

jsj1-b1.gif

    妯″紡妯℃嫙璁婇噺鑸囧闅涜娓窛闆㈢殑鍏紡瑷堢畻濡備笅锛

jsj1-gs1-3.gif

    涓嶇‘瀹氬弮鏁稿拰鍙栧肩瘎鍦嶆槸鏍规摎涔嬪墠鐨勭爺绌朵締纭畾鐨[5]锛屽叿楂旂殑鍙冩暩瑕嬭〃2鎵绀猴紝鍏朵腑zmconv_c0_lnd鍜寊mconv_c0_ocn鏄皪闄嶆按(PRECT)鍜岃蓟灏(FLUT銆丗SNTOA)褰遍熆寰堝ぇ鐨勫弮鏁革紝zmconv_tau鏄皪娴侀檷姘翠腑鏈鏁忔劅鐨勫弮鏁革紝cldsed_ai涔熻璀夋槑鐐烘槸灏嶈蓟灏勬湁寰堝挤褰遍熆鐨勫弮鏁搞

jsj1-b2.gif

    澶氱洰鏍囩殑瑭曞児鏍囨簴鏈夊緢澶氾紝渚嬪闆㈡暎搴︼紙Spread锛夈佷笘浠h窛闆紙GD锛夈乭pervolume鍜屽弽涓栦唬璺濋洟锛圛GD锛夌瓑銆傚叾涓環ypervolume鐐轰竴鍊嬮棞浜庨潪鏀厤瑙g殑闆㈡暎搴︺佹敹鏂傛х殑缍滃悎鎸囨爣锛屽洜鍏跺彲浠ュ悓鏅傝冩叜閫欏叐鍊嬮噸瑕佺殑鎬ц兘鑰屾垚鐐鸿繎骞翠締澶氱洰鏍囪⿻鍍规寚鏍囦腑鏈甯哥敤鐨勬寚鏍囦箣涓銆侷GD鎸囨爣琛¢噺鐨勬槸鐣跺墠鐛插緱鐨勯潪鏀厤瑙i泦鑸囩湡瀵︾殑甯曠疮鎵樺墠娌跨殑璺濋洟锛屾槸灏嶅鐩爣绠楁硶鏀舵杺鎬ф渶濂界殑琛¢噺鏂规硶涔嬩竴銆傞氬父hypervolume鍜孖GD閰嶅悎浣跨敤渚嗗垽鏂峰鐩爣鍎寲鏂规硶鐨勫劒鍔c俬ypervolume鍊艰秺澶ц秺濂斤紝鍙嶄笘浠h窛闆㈣秺灏忚秺濂姐傛湰鏂囧鐩爣鍟忛鍙婂叾灏嶆噳鐨勫弮鑰冮粸閬告搰濡傝〃3鎵绀恒

jsj1-b3.gif

3 瑜囬洔鏁稿鍑芥暩涓婂鐩爣鍎寲绲愭灉

    涓嬫枃鎵鏈夊湒涓瘡涓娆$枈浠i兘鐐10娆″嚱鏁歌▓绠楁垨10娆℃ā寮忔ā鎷熴傚湒2鍜屽湒3涓殑绺藉嚱鏁告ā鎷熸鏁哥偤200娆★紝姣10娆″嚱鏁歌▓绠椾箣寰岋紝閫插寲澶氱洰鏍囩畻娉曞拰MO-ANN绠楁硶瑷堢畻涓娆¢潪鏀厤瑙g殑hypervolume鍜孖GD銆傛湰鏂囨彁鍑虹殑MO-ANN浠g悊妯″紡澶氱洰鏍囧劒鍖栨柟娉曠浉灏嶄簬NSGAIII鍜孧OEA-D渚嗚鑳藉鏇村揩鍦版彁鍗嘮DT2鍜孌TLZ7鍑芥暩鍎寲鏁堟灉銆

jsj1-t2.gif

jsj1-t3.gif

4 鍠煴澶ф埃妯″紡涓婂鐩爣鍎寲绲愭灉

    SCAM鍎寲鍟忛鐒℃硶姹傚嚭鐪熸鐨勫笗绱墭鍓嶆部锛屽洜姝ゅ皪浜庡鐩爣鍦⊿CAM涓婄殑瑭曞児锛屾湰鏂囦互hypervolume浣滅偤鏍囨簴銆傝垏鍦╖DT2鍜孌TLZ7涓婄殑鍎寲娓│鐩稿悓锛岄欒!涔熷皢姣10娆℃ā寮忛亱琛屼綔鐐1娆$枈浠o紝瑷堢畻涓娆ypervolume銆傚緸鍦4涓彲浠ョ湅鍑哄湪TWP-ICE涓奙O-ANN鑳藉鏇村揩鏇村ソ鍦扮嵅鍙栨洿鍎殑闈炴敮閰嶈В闆嗐傚湪绗10娆$枈浠f檪宸茬稉鍙栧緱杓冨劒鐨勭祼鏋滐紝鑰孨SGAIII閫插寲澶氱洰鏍囩畻娉曞墖鍎寲閫熷害鐩稿皪绶╂參锛屽湪绗60娆$枈浠f檪渚濊垔娌掓湁瀹屽叏鏀舵杺锛孧O-ANN鏀舵杺閫熷害鏄叾鐨5鍊嶄互涓娿俆WP-ICE鐩稿皪M-PACE妯℃嫙鏅傞枔鏇撮暦锛岀墿鐞嗗弮鏁稿拰妯℃嫙鎬ц兘涔嬮枔鐨勯棞绯讳篃鐩稿皪鏇磋闆滀竴浜涖傚洜姝ゅ湪M-PACE涓婄殑MO-ANN澶氱洰鏍囩畻娉曠殑鍎嫝鏈兘鏈塗WP-ICE涓婇’钁楋紝浣嗘槸涔熸槸3鍊嬪鐩爣绠楁硶涓簿搴︽渶楂樼殑绠楁硶銆

jsj1-t4.gif

5 绲愯珫

    鏈枃棣栧厛鍒嗘瀽浜嗗湴鐞冪郴绲辨ā寮忎腑闈㈣嚚鐨勫鐩爣鍎寲鍟忛銆傜劧寰屽皪鐣跺墠鍎寲鏂规硶鍦ㄨ闆滃湴鐞冪郴绲辨ā寮忎笂浣跨敤鎵瀛樺湪鐨勫晱椤屽仛浜嗙啊瑕佸垎鏋愶紝骞舵彁鍑轰簡鍩轰簬澶氬堡鎰熺煡姗熺缍撶恫绲★紙MLP锛夌殑浠g悊妯″紡鍎寲绠楁硶MO-ANN銆傛绠楁硶鍒╃敤鍩轰簬MLP鐨勫洖姝告ā鍨嬩唬鏇跨湡瀵﹀湴鐞冪郴绲辨ā寮忛爯娓渶鍎噰妯i粸锛屾瘡澧炲姞涓娆¢噰妯i粸鏇存柊涓娆′唬鐞嗘ā鍨嬨傛渶寰屾湰鏂囧皢MO-ANN鑸囧父鐢ㄥ鐩爣鍎寲绠楁硶鍦ㄨ闆滃嚱鏁稿拰鍠煴澶ф埃妯″紡涓婄殑鍎寲鎬ц兘閫茶浜嗗皪姣旓紝绲愭灉琛ㄦ槑鏂版彁鍑虹殑澶氱洰鏍囦唬鐞嗘ā寮忓劒鍖栫畻娉昅O-ANN鍦ㄨ闆滄暩瀛稿嚱鏁稿拰鍠煴澶ф埃妯″紡涓婇兘鍏锋湁鏄庨’鐨勫劒鍕紝鍦═WP-ICE妯″紡涓婃敹鏂傞熷害鍙浉灏峃SGAIII鎻愬崌5鍊嶄互涓娿傜稖涓婃墍杩帮紝鍩轰簬澶氬堡鎰熺煡姗熶唬鐞嗘ā寮忕殑澶氱洰鏍囧劒鍖栫畻娉曡兘澶犳洿鍔犲叏闈㈡湁鏁堝湴鎳夊皪瑜囬洔鍦扮悆绯荤当妯″紡涓婄殑鍙冩暩鍎寲鍟忛銆

鍙冭冩枃鐛

[1] 鐜嬫枌锛屽懆澶╄粛锛屼繛姘稿挤锛岀瓑.鍦扮悆绯荤当妯″紡鐧煎睍灞曟湜[J].姘h薄瀛稿牨锛2008锛66(6)锛857-869.

[2] MASTRANDREA M D锛孧ACH K J锛孭LATTNER G K锛宔t al.The IPCC AR5 guidance note on consistent treatment of uncertainties锛歛 common approach across the working groups[J].Climatic Change锛2011锛108(4)锛675.

[3] DEB K锛孞AIN H.An evolutionary many-objective optimization algorithm using reference-point-based nondominated sorting approach, part i锛歴olving problems with box constraints[J].IEEE Transactions on Evolutionary Computation锛2014锛18(4)锛577-601.

[4] ZHANG Q锛孡i Hui.MOEA/D锛歛 multiobjective evolutionary algorithm based on decomposition[J].IEEE Transactions on Evolutionary Computation锛2007锛11(6)锛712-731.

[5] ZHANG T锛孡I L锛孡IN Y锛宔t al.An automatic and effective parameter optimization method for model tuning[J].Geoscientific Model Development锛2015锛8(11)锛3579-3591.

[6] XU H锛孼HANG T锛孡UO Y锛宔t al.Parameter calibration in global soil carbon models using surrogate-based optimization[J].Geoscientific Model Development锛2018锛11(7)锛3027-3044.

[7] REGIS R G锛孲HOEMAKER C A.A stochastic radial basis function method for the global optimization of expensive functions[J].Informs Journal on Computing锛2007锛19(4)锛497-509.

[8] SELMIC R R锛孡EWIS F L.Neural-network approximation of piecewise continuous functions锛歛pplication to friction compensation[J].IEEE Transactions on Neural Networks锛2002锛13(3)锛745-751.

[9] ZITZLER E锛孌EB K锛孴HIELE L.Comparison of multiobjective evolutionary algorithms锛歟mpirical results[J].Evolutionary Computation锛2000锛8(2)锛173-195.

[10] COELLO C A C锛孡AMONT G B锛孷AN VELDHUIZEN D A.Evolutionary algorithms for solving multi-objective problems[M].New York锛歋pringer锛2007.

[11] LIN Y锛孌ONNER L J锛孭ETCH J锛宔t al.TWP-ICE global atmospheric model intercomparison锛歝onvection responsiveness and resolution impact[J].Journal of Geophysical Research锛欰tmospheres锛2012锛117(D9)锛欴09111.

[12] VERLINDE J锛孒ARRINGTON J Y锛孧CFARQUHAR G M锛宔t al.The mixed-phase arctic cloud experiment[J].Bulletin of the American Meteorological Society锛2007锛88(2)锛205-222.



浣滆呬俊鎭:

鍚  鍒1锛岄粌  娆2锛岃枦  宸1

(1.娓呰彲澶у 瑷堢畻姗熺瀛歌垏鎶琛撶郴锛屽寳浜100084锛2.鍖椾簽鍒╂閭eぇ瀛 鐢熸佺郴绲辩瀛歌垏绀炬渻涓績锛屽紬鎷夋牸鏂澶86011)