Value Efficiency Analysis
A basic assumption in multiple criteria decision-making research is that there is no objectively best solution for the problem. The best solution depends on a rational DM’s preferences. The term “rational” means that the DM wants to choose the solution fo
- PDF / 374,446 Bytes
- 15 Pages / 439.37 x 666.142 pts Page_size
- 71 Downloads / 181 Views
Value Efficiency Analysis Most Preferred Unit-Based Approach
7.1
Background
A basic assumption in multiple criteria decision-making research is that there is no objectively best solution for the problem. The best solution depends on a rational DM’s preferences. The term “rational” means that the DM wants to choose the solution for which there is no other solution that is equally good on all given criteria and better at least on one criterion. As we have defined in Chap. 4, such solutions are called nondominated. An original DEA problem (Charnes et al. 1978, 1979) is value free. If a unit is on the efficient frontier, it is regarded as good as other efficient units. If a unit is inefficient, it is projected onto the efficient frontier radially or using some other prespecified feature. Even the ranking of the units is based on the use of an “efficiency” measure such as an efficiency score. When the DM is willing to insert his/her preference information into the analysis, the problem turns into a typical multiple criteria problem: there is a need not to consider efficient units equally good. A typical approach in DEA (see Chap. 6) is to use weights for inputs and/or outputs or to operate with target values. The use of weights often is based on the intuitively appealing notion “the greater the importance, the larger the weight.” Unfortunately, it does not always work. If the weights have a straightforward interpretation, such as prices, their definition and use is also straightforward. However, this is not always the case. When inputs are some raw materials and outputs produced goods, it is relatively easy to see the weights as prices. But when outputs are, for example, Ph.D. degrees and refereed publications or lives saved, it may be practically or politically impossible for the DM to give price estimates. Identifying a target or ideal point is a natural way to give preference information, but a DM needs guidance to set up targets which are realistic and please him/her. Often the ultimate goal is to find the solution that the DM prefers most at the © Springer Science+Business Media New York 2015 T. Joro, P.J. Korhonen, Extension of Data Envelopment Analysis with Preference Information, International Series in Operations Research & Management Science 218, DOI 10.1007/978-1-4899-7528-7_7
95
96
7 Value Efficiency Analysis
Fig. 7.1 Illustration on the use of weights and targets
moment of the final choice.1 Such solution is called most preferred unit (MPU)/ most preferred solution (MPS). When identifying targets, it is also important that the method provides a DM with a possibility to learn and to have a holistic view over possible solutions. Interactive approaches are suitable to fulfill these requirements. The following example describes challenges of using weights and targets. Example 7.1 Let us assume that three candidates have applied for a research assistant vacancy. We evaluate them with two criteria (outputs) on the scale from 1 to 10: competence in teaching and competence in research. We assume a constant
Data Loading...
