full2010.pdf - page 457

419
£µ¡š¸É 1 ˜´ ª°¥nµŠŸ¨„µ¦šÎµœµ¥ „¦–¸š¸É˜o°Š„µ¦ž¦´Â˜nŠ‡nµ Neural Network
Áª¨µÄœ„µ¦š—°Áž}œ°´ œ—´°Š¦°Š‹µ„ Fuzzy-PSO —´ Šœ´ÊœÁœºÉ°Š‹µ„ǦŠŠµœœ¸ÊÁœoœÄœž¦³· š›·£µ¡ °Š„µ¦
¡¥µ„¦–r¦µ‡µžd—Áž}œ®¨´ „ šÎµÄ®oǦŠŠµœ‡¦´ÊŠœ¸Ê Á¨º°„ª·›¸ Back Propagation Neural Network Áž}œ˜´ªÂš¸ÉčoĜ„µ¦
¡¥µ„¦–r ×¥‹´—šÎµÄœ¦¼žÂ °ŠÃž¦Â„¦¤ž¦³¥»„˜r εÁ¦È‹¦¼žœÁªÈ
ץĜ¦³œ¸Ê Ÿ¼očoŠµœµ¤µ¦™š¸É ‹³Á¨º°„ž¦´Â˜nŠ‡n µ Neural network ŗoÁ°Š ¨³Á¨º°„‹Îµœªœª´œš¸É˜o°Š„µ¦šÎµœµ¥
Ÿ¨Å—o ˜n ‹³¤¸o°‹Îµ„´—Á¡¸¥ŠÅ¤nÁ„·œ 30 ª´œ ×¥Ÿ¼očo˜o°ŠšÎ µ„µ¦Á¨º°„®»oœš¸É‹³ÄošÎµœµ¥, ‹Î µœªœª´œš¸É˜o°Š„µ¦šÎ µœµ¥,
‹Î µœªœ Input Node, ‹Îµœªœ¦°š¸É‹³ Train ŗo˜µ¤š¸É ˜o°Š„µ¦Äœ„¦–¸š¸É Ÿ¼očo¤¸‡ªµ¤¦¼o¡ºÊœ“µœÄœ„µ¦Äo ǦŠ n µ¥
ž¦³µšÁš¸¥¤Á¡ºÉ °‹³Å—o ‡ªµ¤™¼„˜o°ŠÂ¤n œ¥ÎµÄœ„µ¦šÎ µœµ¥ Á¤ºÉ°Ÿ¼o čošÎ µ„µ¦Á¨º°„¨³„—ž»i¤šÎ µœµ¥¦µ‡µ®»oœ „È‹³
ž¦µ„’ o °¤¼¨Â¨³„¦µ¢š¸É¦³Å—ošÎµ„µ¦ª· Á‡¦µ³®r°°„¤µ Ÿ¨š¸Éŗo‹³Â—Š°°„¤µÁž}œ´¨´„¬–r ˜µ¦r —´Š£µ¡š¸É 1
¦»žŸ¨Â¨³ª·‹µ¦–r
„µ¦¡¥µ„¦–r ˜¨µ—®»oœÃ—¥ÄoǦŠ nµ¥ž¦³µšÁš¸¥¤œ´ Áž}œª·›¸®œ¹É Š °Š„µ¦ª·Á‡¦µ³®ršµŠÁš‡œ·‡š¸Éŗo¦´
‡ªµ¤œ· ¥¤Â¡¦n®¨µ¥¤µ„ ÁœºÉ°Š‹µ„Áž}œ˜´ªÂš¸É¤¸‡ªµ¤¥º—®¥»n œ¼Š µ¤µ¦™ž¦´Â˜nŠ¡µ¦µ¤·Á˜°¦rŗo ®¨µ„®¨µ¥Ä®o
Á®¤µ³¤„´ „µ¦¡¥µ„¦–rĜ¦¼žÂ˜n µŠÇ œ°„‹µ„œ¸ÊǦŠ nµ¥ž¦³µšÁš¸¥¤¥´Š¤¸ „µ¦œÎµ o°¤¼¨š¸ÉÁ„·— ¹Ê œÄœ°—¸˜¤µ
Á¦¸ ¥œ¦¼o Á¡ºÉ°Äo 宦´ „µ¦¡¥µ„¦–r‡nµÄœ°œµ‡˜Å—o°¸„—oª¥ Šµœª·‹´¥œ¸Ê ‹¹ŠÅ—oÁœ°˜´ªÂ 宦´¡¥µ„¦–r®» oœ 1, 5, 7
¨³ 30 ª´ œ¨nªŠ®œoµ ×¥˜´ªÂœ¸ÊÁž}œ˜´ªÂÃ‡¦Š nµ¥ž¦³µšÁš¸¥¤ÂÁ¦¸ ¥œ¦¼o¡¦n„¨´š¸É¤¸‹Îµœªœ´ÊœŽn°œÁ¡¸¥Š
´ÊœÁ—¸¥ª ¨³‹ÎµœªœÃ®œ—Äœ´ÊœŽn°œµ¤µ¦™„ε®œ—Á°ŠÅ—o Á¡ºÉ °š¸É‹³šÎµÄ®o„µ¦¡¥µ„¦–r Ä®o‡nµ‡¨µ—Á‡¨ºÉ°œœo°¥»—
ץčoo°¤¼¨ °Š®»oœš¸É¤¸„µ¦ŽºÊ ° µ¥Äœ˜¨µ—®¨´„š¦´¡¥r®n Šž¦³Áš«Åš¥ ĜnªŠže¡» š›«´„¦µ 2551-2552 宦´
„µ¦ª´—ž¦³· š›·£µ¡ °Š˜´ªÂª´ —‹µ„‡nµ°´˜¦µ„µ¦šÎµœµ¥™¼„¨‡nµ‡ªµ¤Á¦È ªš¸ÉčoĜ„µ¦Á¦¸¥œ¦¼o ×¥Áž¦¸ ¥Áš¸¥„µ¦
¡¥µ„¦–r°Šš´ÊŠµ¤˜´ªÂ