Üniversitelerin Etkinlik Ölçümünde Bulanık Veri Zarflama Analizi Uygulaması
Yükleniyor...
Tarih
2009
Dergi Başlığı
Dergi ISSN
Cilt Başlığı
Yayıncı
Selçuk Üniversitesi
Erişim Hakkı
info:eu-repo/semantics/openAccess
Özet
Birden çok girdi-çıktının olduğu ve girdi-çıktıların farklı ölçü birimlerine sahip olduğu durumlarda, işletmelerin (karar verme birimleri-KVB) etkinlik değerleri genellikle Veri Zarflama Analizi (VZA) ile hesaplanmaktadır. Geleneksel VZA modelleri, yalnızca kullanılan girdi ve üretilen çıktıların kesin olarak bilindiği durumlarda uygulanabilmektedir. Verilerin kesin olarak bilinmediği durumlarda etkinlik ölçümlerinin yapılabilmesi için ise Bulanık Veri Zarflama Analizi (BVZA) modelleri geliştirilmiştir. VZA ile etkinlik ölçümünde; girdi ve çıktıları ağırlıklandırmada serbestlik tanınmakta, bu da VZA’nın ayrım yapma gücünün azalmasına, çok fazla işletmenin etkin çıkmasına sebep olabilmektedir. Ayrıca, girdi ve çıktılara verilecek ağırlıkları seçmede tanınan bu serbestlik, aynı veri seti kullanılmasına rağmen, bazen çok farklı ağırlık değerlerinin verilmesine de sebep olmaktadır. Saati ve Memariani (2005) tarafından ağırlıklardaki bu esnekliğin kontrol edilebildiği ve tüm KVB’ler için aynı ağırlık kümesinin kullanıldığı bir model önerilmiştir. Bu çalışmada Saati ve Memariani tarafından önerilen model kullanılarak Türkiye’deki 24 devlet üniversitesinin 2006 yılı etkinlik ölçümleri yapılmıştır. Etkinlik ölçümü için 6 adet girdi (öğretim üyesi sayısı, öğretim görevlisi ve okutman sayısı, araştırma görevlisi sayısı, toplam personel giderleri, mal ve hizmet alım giderleri, kapalı kullanım alanı), 7 adet çıktı (önlisans ve lisans öğrenci sayısı, lisansüstü öğrenci sayısı, proje sayısı, proje bütçeleri, uluslararası yayın sayısı, ulusal yayın sayısı, öz gelirler) belirlenmiştir. Uygulama sonunda; Sakarya, Afyon Kocatepe, Yıldız Teknik ve Çanakkale Onsekiz Mart Üniversitelerinin etkinlik değerleri %95-100, Süleyman Demirel ve Mustafa Kemal Üniversitelerinin %85-90, Gaziosmanpaşa, Dumlupınar, Kocaeli, Pamukkale, Muğla, Mersin ve Akdeniz Üniversitelerinin %80-85, Kafkas ve Yüzüncü Yıl Üniversitelerinin %70-75, Eskişehir Osmangazi ve Zonguldak Karaelmas Üniversitelerinin %65-70, Niğde Üniversitesi’nin %60-65, Kırıkkale ve Abant İzzet Baysal Üniversitelerinin %55-60, Balıkesir, Adnan Menderes, Trakya Üniversitelerinin %50-55 aralığında, Gaziantep Üniversitesi’nin %45 çıkmıştır. Elde edilen sonuçlara göre üniversitelere önerilerde bulunulmuştur.
Purpose: The purpose of this study is to measure the efficiencies of the universities by using Fuzzy Data Envelopment Analysis (FDEA) of which purpose is the measurement of relative efficiencies of enterprises in the situations where there are more than one input that produces more than one output; where the inputs and outputs have different measurement units and where the data is fuzzy. Furthermore, calculation and comparison of which proportion the universities use their inputs unproductively and produce their output inefficiently, is aimed. Introduction: Efficiency, productivity like concepts have always been important and will maintain their importance in our world where the resources are limited. Efficiency and productivity analyses are very important management tools in order to find the relation between the produced output and the used source input by enterprises. Enterprises, in the production process obtain several output with different measurement units while using several input having different measurement units. For managers it is a very difficult process to determine the less effective enterprises by comparing the input – output relationships simultaneously. DEA: When there are several input – output and they have different measurement units, DEA provides an important help tool to managers. DEA aims to determine the relative efficiency of enterprises and developed by Charnes, Cooper and Rhodes (CCR) in 1978. Base assumption in DEA, which is a linear programming based technique, is to make enterprises have similar strategic objectives and to produce the same outputs by using the same inputs. In DEA; by investigating the inputs and outputs of decision making units (DMUs), the ones with the best performance are selected and by using these DMUs efficient production frontier is created. Efficiency values of the DMUs which are not on the efficient frontier are also defined according to that efficient frontier. Since the efficient frontier obtained as a result of the analysis wrap all DMUs, name of the method is called to DEA and the set formed by DMUs are called the reference set. Efficient units in the reference set are used in order for inefficient DMUs to get active. Selection of DMUs, selection of input and output, measurement of relative activity, determination of reference sets and evaluation of results are the steps in the application of DEA: FDEA: It is very important to choose the input and output data carefully and the reliability of these data in order to get accurate results from DEA which is a data based efficiency measuring method. However, in many real world applications input – output datas cannot be added up correctly and exactly. Traditional DEA models can only be used if used inputs and produced outputs are known exactly. When data are not known, a method called FDEA models are developed to be able to perform efficiency measurement. Data of FDEA models can be classified as: 1) İnterval data (Fuzzy number data that the lower and upper values or the membership function are known), 2) Ordinal data (Data where the verbal sequental relation like large-small-equal or very important-important-unimportant is known between any i. input or r. output of DMUs), 3) Missing data, and 4) Exact data. FDEA models suggested depending on the used data type can be classified as under the following headings: 1)Data models where ordinal and exact datas, 2)Data models where ordinal, interval and exact datas and 3)Data models where interval and exact datas. In DEA there is flexibility of weighting input and output by setting up separate models for each DMU. DEA assumes that each DMU choose its input and output weightings to maximise its efficiency score. However, this flexibility on the selection of input – output weightings may result in the reduction of distinguishing power of DEA and too much efficient DMU when the numbers of input – output increases while the number of DMU remains constant. This provided flexibility sometimes cause different weighting values to be given even though the same data set is used. In this study, a model proposed by Saati and Memariani (2005) which allows the flexibility of weightings to be controlled and in which the same data set is used for every DMU. Saati ve Memariani Model: In the first step, the upper bounds of output and input weights are determined. A common set of weights is determined in the second step by compacting the weight intervals. In the last step, the fuzzy efficiency of each decision making unit evaluated. Application: From the last quarter of the twentieth century began the process of transition to information society and knowledge economy, called a global economic structure has been formed. In this new structure, individuals' economic power, knowledge and education levels, the competitiveness of countries with the human and social capital has begun to be measured. This process increased expectations from universities which are primarily responsible of knowledge generation and sharing. In addition to expectations, especially with a high percentage of young population in developing countries the increase in demand for higher education, forces universities to use their resources effectively. In this study, efficiency measurement of 24 state universities in Turkey in 2006 has been carried out. 6 inputs (the number of faculty members, instructional assistants and the number of lecturers, research assistants count, total personnel expenses, procurement of goods and services costs, use of indoor space) and 7 outputs (the number of associate degree and undergraduate students, the number of graduate students, the number of projects, project budgets, the number of international publications, the number of national publications, self incomes) were used. Results: At the end of the application; Sakarya University’s, Afyon Kocatepe University’s, Yıldız Technical University’s and Çanakkale Onsekiz Mart University’s efficiency percentages have been in the interval 95% and 100%; Süleyman Demirel University’s and Mustafa Kemal University’s 85%-90%; Gaziosmanpaşa University’s, Dumlupınar University’s, Kocaeli University’s, Pamukkale University’s, Muğla University’s, Mersin University’s and Akdeniz University’s 80%-85%, Kafkas University’s and Yüzüncü Yıl University’s 70%-75%, Eskişehir Osmangazi University’s and Zonguldak Karaelmas University’s 65%-70%, Niğde University’s 60%-65%, Kırıkkale University’s and Abant İzzet Baysal University’s 55%-60%, Balıkesir University’s, Adnan Menderes University’s, Trakya University’s percentages have been in the interval 50%-55% and Gaziantep University’s efficiency percentage has been 45%.
Purpose: The purpose of this study is to measure the efficiencies of the universities by using Fuzzy Data Envelopment Analysis (FDEA) of which purpose is the measurement of relative efficiencies of enterprises in the situations where there are more than one input that produces more than one output; where the inputs and outputs have different measurement units and where the data is fuzzy. Furthermore, calculation and comparison of which proportion the universities use their inputs unproductively and produce their output inefficiently, is aimed. Introduction: Efficiency, productivity like concepts have always been important and will maintain their importance in our world where the resources are limited. Efficiency and productivity analyses are very important management tools in order to find the relation between the produced output and the used source input by enterprises. Enterprises, in the production process obtain several output with different measurement units while using several input having different measurement units. For managers it is a very difficult process to determine the less effective enterprises by comparing the input – output relationships simultaneously. DEA: When there are several input – output and they have different measurement units, DEA provides an important help tool to managers. DEA aims to determine the relative efficiency of enterprises and developed by Charnes, Cooper and Rhodes (CCR) in 1978. Base assumption in DEA, which is a linear programming based technique, is to make enterprises have similar strategic objectives and to produce the same outputs by using the same inputs. In DEA; by investigating the inputs and outputs of decision making units (DMUs), the ones with the best performance are selected and by using these DMUs efficient production frontier is created. Efficiency values of the DMUs which are not on the efficient frontier are also defined according to that efficient frontier. Since the efficient frontier obtained as a result of the analysis wrap all DMUs, name of the method is called to DEA and the set formed by DMUs are called the reference set. Efficient units in the reference set are used in order for inefficient DMUs to get active. Selection of DMUs, selection of input and output, measurement of relative activity, determination of reference sets and evaluation of results are the steps in the application of DEA: FDEA: It is very important to choose the input and output data carefully and the reliability of these data in order to get accurate results from DEA which is a data based efficiency measuring method. However, in many real world applications input – output datas cannot be added up correctly and exactly. Traditional DEA models can only be used if used inputs and produced outputs are known exactly. When data are not known, a method called FDEA models are developed to be able to perform efficiency measurement. Data of FDEA models can be classified as: 1) İnterval data (Fuzzy number data that the lower and upper values or the membership function are known), 2) Ordinal data (Data where the verbal sequental relation like large-small-equal or very important-important-unimportant is known between any i. input or r. output of DMUs), 3) Missing data, and 4) Exact data. FDEA models suggested depending on the used data type can be classified as under the following headings: 1)Data models where ordinal and exact datas, 2)Data models where ordinal, interval and exact datas and 3)Data models where interval and exact datas. In DEA there is flexibility of weighting input and output by setting up separate models for each DMU. DEA assumes that each DMU choose its input and output weightings to maximise its efficiency score. However, this flexibility on the selection of input – output weightings may result in the reduction of distinguishing power of DEA and too much efficient DMU when the numbers of input – output increases while the number of DMU remains constant. This provided flexibility sometimes cause different weighting values to be given even though the same data set is used. In this study, a model proposed by Saati and Memariani (2005) which allows the flexibility of weightings to be controlled and in which the same data set is used for every DMU. Saati ve Memariani Model: In the first step, the upper bounds of output and input weights are determined. A common set of weights is determined in the second step by compacting the weight intervals. In the last step, the fuzzy efficiency of each decision making unit evaluated. Application: From the last quarter of the twentieth century began the process of transition to information society and knowledge economy, called a global economic structure has been formed. In this new structure, individuals' economic power, knowledge and education levels, the competitiveness of countries with the human and social capital has begun to be measured. This process increased expectations from universities which are primarily responsible of knowledge generation and sharing. In addition to expectations, especially with a high percentage of young population in developing countries the increase in demand for higher education, forces universities to use their resources effectively. In this study, efficiency measurement of 24 state universities in Turkey in 2006 has been carried out. 6 inputs (the number of faculty members, instructional assistants and the number of lecturers, research assistants count, total personnel expenses, procurement of goods and services costs, use of indoor space) and 7 outputs (the number of associate degree and undergraduate students, the number of graduate students, the number of projects, project budgets, the number of international publications, the number of national publications, self incomes) were used. Results: At the end of the application; Sakarya University’s, Afyon Kocatepe University’s, Yıldız Technical University’s and Çanakkale Onsekiz Mart University’s efficiency percentages have been in the interval 95% and 100%; Süleyman Demirel University’s and Mustafa Kemal University’s 85%-90%; Gaziosmanpaşa University’s, Dumlupınar University’s, Kocaeli University’s, Pamukkale University’s, Muğla University’s, Mersin University’s and Akdeniz University’s 80%-85%, Kafkas University’s and Yüzüncü Yıl University’s 70%-75%, Eskişehir Osmangazi University’s and Zonguldak Karaelmas University’s 65%-70%, Niğde University’s 60%-65%, Kırıkkale University’s and Abant İzzet Baysal University’s 55%-60%, Balıkesir University’s, Adnan Menderes University’s, Trakya University’s percentages have been in the interval 50%-55% and Gaziantep University’s efficiency percentage has been 45%.
Açıklama
Anahtar Kelimeler
Üniversite, Bulanık, Veri Zarflama Analizi, Etkinlik, University, Fuzzy, Data Envelopment Analysis, Efficiency
Kaynak
Selçuk Üniversitesi Sosyal Bilimler Enstitüsü Dergisi
WoS Q Değeri
Scopus Q Değeri
Cilt
Sayı
22
Künye
Oruç, K. O., Güngör, İ., Demiral, M. F., (2009). Üniversitelerin Etkinlik Ölçümünde Bulanık Veri Zarflama Analizi Uygulaması. Selçuk Üniversitesi Sosyal Bilimler Enstitüsü Dergisi, 22, 279-294.
