เอกสารการประชุมวิชาการและเสนอผลงานวิจัย มหาวิทยาลัยทักษิณ ครั้งที่ 19 2552 - page 348

„µ¦ž¦³»¤ª·µ„µ¦Â¨³Áœ°Ÿ¨Šµœª·‹´¥ ¤®µª·š¥µ¨´¥š´„¬·– ‡¦´ÊŠš¸É 19 ž¦³‹Îµže 2552
5
(Pure Technical Efficiency) Ž¹ÉŠ®µÅ—o‹µ„„µ¦‡Îµœª–‡nµž¦³·š›·£µ¡šµŠÁš‡œ·‡ „¦–¸Ÿ¨˜°Âšœ˜n° œµ—Ÿ´œÂž¦
‹µ„„µ¦¤¸ œµ— °Š®œnª¥Ÿ¨·˜ (Unit) š¸É˜„˜nµŠ„´œ (Variable Returns to Scale) ‡nµž¦³·š›·£µ¡š¸É‡Îµœª–Å—o
ž¦³„°—oª¥ ž¦³·š›·£µ¡Á·Š„µ¦Ÿ¨·˜Ášnµœ´Êœ Ž¹ÉŠ®¤µ¥™¹Š ¦³—´ª· š¥µ„µ¦ ®¦º°Áš‡ÃœÃ¨¥¸š¸Éčo °Š‡r‡ªµ¤¦¼oĜ
°Š‡r„¦ „µ¦Ÿ¨·˜Ã—¥Äož{‹‹´¥œÎ µÁ oµÄœ¦³—´š¸ÉÁ®¤µ³¤Áž}œ˜oœ ×¥®µ„‡nµ (Pure Technical Efficiency) Ášnµ„´ 1
®¤µ¥‡ªµ¤ªnµ ®œn ª¥ (Unit) œ´Êœ¤¸„µ¦Ÿ¨·˜°¥nµŠ¤¸ž¦³· š›·£µ¡šµŠÁš‡œ·‡ ¨³®µ„‡nµ TEVRS œo °¥„ªnµ 1 ®¤µ¥‡ªµ¤
ªnµ ®œn ª¥ (Unit) œ´Êœ¤¸ ‡ªµ¤—o°¥ž¦³·š›·£µ¡šµŠÁš‡œ·‡ ®¦º°‡ªµ¤—o°¥ž¦³·š›·£µ¡°´œÁœºÉ °Š‹µ„„µ¦Äo nªœŸ¤
°Šž{‹‹´¥œÎµÁ oµš¸ÉŤnÁ®¤µ³¤
‡nµž¦³·š›·£µ¡˜n° œµ— (Scale Efficiency : SE) ‡nµœ¸Ê °„Ä®o š¦µªnµÁ¤ºÉ °¤¸„µ¦Äož{‹‹´ ¥œÎµÁ oµÁ¡·É¤ ¹Êœ
°¥n µŠÁž}œ´—nªœÂ¨oª Ÿ¨Ÿ¨· ˜‹³Á¡·É¤ ¹Êœ°¥nµŠÁž}œ´—nªœ¤µ„œo °¥Á¡¸¥ŠÄ— ×¥®œn ª¥Ÿ¨·˜Ä—š¸É ¤¸ž¦³·š›·£µ¡˜n °
œµ——Šªn µ „µ¦Á¡·É ¤ž{‹‹´ ¥œÎµÁ o µ°¥nµŠÁž}œ´—nªœ Ÿ¨Ÿ¨·˜š¸É ŗo‹³Á¡·É ¤ ¹Êœ°¥nµŠÁž}œ´—nªœÄœ œµ—‡Šš¸É ®¦º°š¸É
Á¦¸¥„„´ œªnµ ¤¸Ÿ¨˜°Âšœ˜n ° œµ—‡Šš¸É (Constant Returns to Scale) ®µ„®œnª¥Ÿ¨·˜µ¤µ¦™Ÿ¨·˜Å—oÁ¡·É¤ ¹Êœ —Šªnµ
®œnª¥Ÿ¨·˜œ´Êœ¤¸ „µ¦Ÿ¨·˜š¸É¤¸Ÿ¨˜°Âšœ˜n ° œµ—Á¡·É¤ ¹Êœ (Increasing Returns to Scale) ¨³ÄœšµŠ˜¦Š„´œ oµ¤ ®µ„
®œnª¥Ÿ¨· ˜µ¤µ¦™Ÿ¨· ˜Å—o œo °¥„ªn µ´—n ªœ °Šž{‹‹´ ¥œÎ µÁ oµ —Šªn µ„µ¦Ÿ¨·˜¤¸ Ÿ¨˜°Âšœ˜n ° œµ—¨—¨Š
(Decreasing Return to Scale) ‡nµ SE ¤¸‡nµœo °¥„ªnµ®¦º°Ášnµ„´ 1 ×¥ ®µ„‡nµ SE Ášnµ„´ 1 ®¤µ¥‡ªµ¤ªn µ ®œnª¥
(Unit) œ´Êœ¤¸Ÿ¨„µ¦Ÿ¨·˜°¥nµŠ¤¸ ž¦³·š›·£µ¡˜n° œµ— ¨³®µ„‡nµ SE œo °¥„ªnµ 1 ®¤µ¥‡ªµ¤ªnµ ®œn ª¥ (Unit) œ´Êœ¤¸
‡ªµ¤—o°¥ž¦³·š›·£µ¡˜n° œµ—®¦º°¤¸ ‡ªµ¤—o°¥ž¦³·š›·£µ¡°´œÁœºÉ°Š‹µ„¤¸„µ¦Ÿ¨·˜Äœ œµ—š¸ÉŤnÁ®¤µ³¤
Data Envelopment Analysis (DEA) Áž}œª· ›¸„µ¦š¸Éœ· ¥¤Äo„´œ°¥nµŠÂ¡¦n®¨µ¥Äœ„µ¦ª´—ž¦³·š›·£µ¡ °Š
®œnª¥Šµœ®¦º°°Š‡r„¦˜nµŠ Ç Ã—¥ÁŒ¡µ³Äœ®œnª¥Šµœ®¦º°Ã‡¦Š„µ¦ °Š£µ‡¦´“µ¨ ®¦º°®œnª¥Šµœš¸É ŤnªŠ®µ„εŦ
ÁœºÉ°Š‹µ„ DEA µ¤µ¦™šÎ µ„µ¦ª´—ž¦³·š›·£µ¡ °Š°Š‡r„¦ ×¥„µ¦¡· ‹µ¦–µž{‹‹´ ¥œÎµÁ o µ (Inputs) ¨³Ÿ¨Ÿ¨· ˜
(Outputs) š¸É Áž}œ˜´ªÂž¦Á·Š‡»–£µ¡ (Qualitative Variable) ¨³Á· Šž¦·¤µ– (Quantitative Variables) ŗo®¨µ¥ž{‹‹´¥
Ĝ‡¦µªÁ—¸ ¥ª„´œ ץčoÁš‡œ·‡„µ¦Ãž¦Â„¦¤Á·ŠÁoœšµŠ‡–·˜«µ˜¦r (Mathematical Programming) šÎµÄ®o µ¤µ¦™
ª·Á‡¦µ³®r‡ªµ¤¤¸ž¦³·š›·£µ¡®¦º°‡ªµ¤—o °¥ž¦³·š›·£µ¡Äœ„µ¦Äoš¦´¡¥µ„¦š¸É ¤¸°¥¼n¨³Ÿ¨Ÿ¨· ˜š¸Éŗo œ°„‹µ„œ¸Ê¥´Š
µ¤µ¦™®µµÁ®˜» °Š„µ¦—o°¥ž¦³·š›·£µ¡ (Inefficiency) Ž¹ÉŠµ¤µ¦™ÄoÁž}œ o°¤¼¨Äœ„µ¦ž¦´ ž¦»Šž¦³·š›·£µ¡
°Š®œn ª¥Šµœ
„µ¦ª·‹´¥Äœ‡¦´ÊŠœ¸Ê‹¹ŠÁ¨º °„čo‹Î µ¨°Š Data Envelopment Analysis (DEA) ˜µ¤Âœª‡·— °Š (Farrell,
1957) ×¥Áž}œÂ‹Îµ¨°Šš¸ÉšÎµ„µ¦¡·‹µ¦–µ—oµœž{‹‹´¥„µ¦Ÿ¨· ˜ (Input Oriented Measures) ˜nÁœºÉ°Š‹µ„ o °‹Îµ„´—Äœ
‹Îµ¨°Š °Š °Š (Farrell, 1957) œ´Êœ¤¤˜·Ä®o‹Îµ¨°Š—´Š„¨nµª¤¸¨´„¬–³Ÿ¨˜°Âšœ˜n° œµ—‡Šš¸É (Constant
Returns to Scale) Ášn µœ´Êœ œ´Éœ‡º° ¤¸o°¤¤˜·ªnµ®œnª¥Ÿ¨·˜¤¸„µ¦Ÿ¨·˜Äœ œµ—š¸É Á®¤µ³¤Â¨oª ˜nĜ‡ªµ¤Áž}œ‹¦· Š¥´Š
¤¸ž{ ‹‹´¥°ºÉœš¸É°µ‹šÎ µÄ®oŤn µ¤µ¦™Ÿ¨·˜ – œµ—„µ¦Ÿ¨·˜š¸ÉÁ®¤µ³¤ Ánœ œµ— °Š®œnª¥Ÿ¨·˜ ‹¹ ŠÅ—o¤¸„µ¦¡´•œµ
‹Îµ¨°Š£µ¥Ä˜o o°¤¤˜· Variable Return to Scale (VRS) Ž¹ÉŠÂ‹Îµ¨°Š VRS (Banker, Charnes and Cooper
1984) œ¸Ê µ¤µ¦™¨— o°‹Îµ„´—˜µ¤Â‹Îµ¨°Š °Š °Š (Farrell, 1957) šÎµÄ®o µ¤µ¦™Â¥„Â¥³Â®¨nŠš¸É¤µ ®¦º°µÁ®˜»
°Š‡ªµ¤¤¸ž¦³· š›·£µ¡Å—oªnµ¤µ‹µ„„µ¦Äo (Pure technical efficiency : TEVRS) ®¦º°¤µ‹µ„„µ¦¤¸ž¦³· š›·£µ¡˜n°
œµ— (Scale efficiency ®¦º° Scale inefficiency : SE)
„µ¦ª·Á‡¦µ³®rĜn ªœš¸É °ŠÁž}œ„µ¦ª· Á‡¦µ³®r®¨nŠš¸É¤µ °Š‡ªµ¤¤¸ž¦³·š›·£µ¡(ĜÁ·Š„µ¦¦·„µ¦) ×¥
°µ«´¥„¦°Âœª‡·—„µ¦¡´•œµÂ¨³¦´¦°Š‡»–£µ¡Ã¦Š¡¥µµ¨ „µ¦¡´•œµÂ¨³¦´¦°Š‡»–£µ¡Ã¦Š¡¥µµ¨ (Hospital