包絡分析法と遺伝的アルゴリズムによる事例ベース意思決定支援モデル  [in Japanese] A New Scheme of Case - based Decision Support Systems by Using DEA and GA Techniques  [in Japanese]

Access this Article

Search this Article

Author(s)

Abstract

本稿では,包絡分析法(DEA)と遺伝的アルゴリズム(GA)を用いた事例に基づく意思決定支援モデルを提案する.本モデルは,複数の要因を入力とし,そこから評価の対象となる結果が生み出されるような一連の活動を対象とし,多くの活動事例の中でユーザが目標とする効率のレベルと過去の活動の特徴を反映した新たな活動方針を導くものである.本モデルは次のような特徴を持つ.1.過去の活動事例がDEAにより求められる効率レベルと各活動の参照関係などの定性的な特徴によって分類され,それをもとに任意の条件に対して将来の活動計画を作成することができる.2. 一元的に判断できない問題に対してもさまざまな側面から効率的であるか否かが評価され,それによって作成された活動計画に対しては,与えられた条件のもとで効率的であることが保証されている.たとえば,マーケティング分野では,広告を複数のメディアに投入し,その広告効果を評価する.活動事例の内容は,複数メディアへの広告投入量と,投入した期間と商品認知率などの評価値で特徴づけられる各種宣伝活動である.これを効率レベルと活動の特徴により分類し,それらを参考にした任意の評価値を満足するような将来の活動案を効率レベルに基づいて意思決定が行えることになる.We propose a new scheme of case-based decision support systems (DSS) by using Data Envelopment Analysis (DEA) and Genetic Algorithms (GA). The application fields for our scheme are the cases of multiple input/output activity in which the efficiency of the outputs is evaluated. The case-based DSS of our scheme offers activity-policy which reflects any level both of the efficiency and features in the activity referring to a lot of cases of past activity in the same area. Our scheme is constructed by two procedures, i.e., analysis procedure and estimation procedure. In analysis procedure, all cases are recursively evaluated by solving a DEA model. The remainder cases except for the cases belonging to efficiency frontier are also evaluated by the DEA model, and the similar processing is repeated. After analysis procedure, it is possible to classify the whole cases into multiple hierarchies by the level of the efficiency, and also into the groups with common features between inputs and output, which cover multiple hierarchies. In estimation procedure, according to some past cases, features, frontier levels, and required conditions, any future activity plan is searched by GA with fitness function using considered factors. After estimation procedure, users control the variety of required conditions about past activities, and finally decide the plan according to their requests.

We propose a new scheme of case-based decision support systems (DSS) by using Data Envelopment Analysis (DEA) and Genetic Algorithms (GA). The application fields for our scheme are the cases of multiple input/output activity in which the efficiency of the outputs is evaluated. The case-based DSS of our scheme offers activity-policy which reflects any level both of the efficiency and features in the activity referring to a lot of cases of past activity in the same area. Our scheme is constructed by two procedures, i.e., analysis procedure and estimation procedure. In analysis procedure, all cases are recursively evaluated by solving a DEA model. The remainder cases except for the cases belonging to efficiency frontier are also evaluated by the DEA model, and the similar processing is repeated. After analysis procedure, it is possible to classify the whole cases into multiple hierarchies by the level of the efficiency, and also into the groups with common features between inputs and output, which cover multiple hierarchies. In estimation procedure, according to some past cases, features, frontier levels, and required conditions, any future activity plan is searched by GA with fitness function using considered factors. After estimation procedure, users control the variety of required conditions about past activities, and finally decide the plan according to their requests.

Journal

  • 情報処理学会論文誌数理モデル化と応用(TOM)

    情報処理学会論文誌数理モデル化と応用(TOM) 42(SIG05(TOM4)), 89-98, 2001-05-15

    Information Processing Society of Japan (IPSJ)

References:  11

Codes

  • NII Article ID (NAID)
    110002725841
  • NII NACSIS-CAT ID (NCID)
    AA11464803
  • Text Lang
    JPN
  • Article Type
    Article
  • ISSN
    1882-7780
  • NDL Article ID
    5760275
  • NDL Call No.
    Z74-C192
  • Data Source
    CJP  NDL  NII-ELS  IPSJ 
Page Top