سال انتشار: ۱۳۸۶

محل انتشار: اولین کنفرانس بین المللی تحقیق در عملیات ایران

تعداد صفحات: ۱

نویسنده(ها):

Josef Jablonsky – Department of EconometricsFaculty of Informatics and Statistics, University of Economics PragueW. Churchill Sq. 4, 130 67 Praha 3

چکیده:

Data envelopment analysis (DEA) models are widely used analytical tools for evaluation of relative efficiency of transformation of multiple inputs into multiple outputs within the set of homogenous decision making units. They allow estimating relative efficiency of the set of units, split them into two groups – efficient and inefficient – and recommend improving of inputs and/or outputs of inefficient units for reaching the efficient frontier.The main aim of the paper is to verify alternative approaches for allocation of resources among university departments at the Faculty of Informatics and Statistics of the University of Economics Prague. The current practice is based on multiple criteria decision making techniques. The main criteria influencing the amount of resources (mainly for salaries and remunerations for the department) are given by real outputs of the department, i.e. the number of hours of courses taught within the last two teaching periods, qualification structure of the department, publications and other indirect outputs (exams, PhD students, number of master students, etc.). Allocation of recourses is a very sensitive task and changing of current practice is conditioned by a consensus of all departments. That is why our approach is just a contribution to the discussion how to solve this allocation problem better. The paper introduces the approach based on standard DEA models. The main input is the amount of resources allocated to the department by the current system. The outputs are the criteria mentioned above. Our approach is based on solving standard DEA model with constant or variable returns to scale. By the model the departments are split into two groups – efficient and inefficient. Our approach offers how to change inputs of efficient (increase of inputs, i.e. resources) and inefficient (decrease of inputs) under the assumption that the outputs remain unchanged in order to move all the originally inefficient departments to the efficient frontier. The DEA results are compared to the current practice and to the results of several other multicriteria decision making methods.Numerical experiments with DEA models are realized on our own MS Excel based application that allows solving problems up to 200 decision making units and 50 inputs and outputs. The paper contains brief information about our DEA solver with covers basic DEA models including super-efficiency models, models with uncontrollable inputs and outputs and models with undesirable inputs and outputs. The system can be simply downloaded from my personal web pages and used without any restrictions