Preference aggregation and DEA: An analysis of the methods proposed to discriminate efficient candidates

There are different ways to allow the voters to express their preferences on a set of candidates. In ranked voting systems, each voter selects a subset of the candidates and ranks them in order of preference. A well-known class of these voting systems are scoring rules, where fixed scores are assigned to the different ranks and the candidates with the highest score are the winners. One of the most important issues in this context is the choice of the scoring vector, since the winning candidate can vary according to the scores used. To avoid this problem, Cook and Kress [W.D. Cook, M. Kress, A data envelopment model for aggregating preference rankings, Management Science 36 (11) (1990) 1302–1310], using a DEA/AR model, proposed to assess each candidate with the most favorable scoring vector for him/her. However, the use of this procedure often causes several candidates to be efficient, i.e., they achieve the maximum score. For this reason, several methods to discriminate among efficient candidates have been proposed. The aim of this paper is to analyze and show some drawbacks of these methods.     

Aggregating preferences rankings with variable weights

One of the most important issues for aggregating preferences rankings is the determination of the weights associated with the different ranking places. To avoid the subjectivity in determining the weights, Cook and Kress (1990) [5] suggested evaluating each candidate with the most favorable scoring vector for him/her. With this purpose, various models based on Data Envelopment Analysis have appeared in the literature. Although these methods do not require predetermine the weights subjectively, some of them have a serious drawback: the relative order between two candidates may be altered when the number of first, second, …, kth ranks obtained by other candidates changes, although there is not any variation in the number of first, second, …, kth ranks obtained by both candidates. In this paper we propose a model that allows each candidate to be evaluated with the most favorable weighting vector for him/her and avoids the previous drawback. Moreover, in some cases, we give a closed expression for the score assigned with our model to each candidate.     

Discriminating DEA efficient candidates by considering their least relative total scores

Data envelopment analysis very often identifies more than one candidate in a voting system to be DEA efficient. In order to choose a winner from among the DEA efficient candidates, this paper proposes a new method that discriminates the DEA efficient candidates by considering their least relative total scores. The proposed method is illustrated with two numerical examples and proves to be effective and practical.     

A solution method to the problem proposed by Wang in voting systems

This paper proposes a method to rank multiple efficient candidates, which often happens in DEA method, by comparing the least relative total scores for each efficient candidate with the best and the least relative total scores measured in the same range. By a numerical example, our model is used to identify efficient candidate and the model can get less efficient candidates too than that can be identified by the model given by Wang and Chin [Y.M. Wang, K.S. Chin, Discriminating DEA efficient candidates by considering their least relative total scores, J. Comput. Appl. Math. 206 (2007) 209–215]. This paper also points out that there is a drawback in the theorem about ε$\epsilon$ given by Wang and Chin [Y.M. Wang, K.S. Chin, Discriminating DEA efficient candidates by considering their least relative total scores, J. Comput. Appl. Math. 206 (2007) 209–215].     

Three new models for preference voting and aggregation

Preference voting and aggregation require the determination of the weights associated with different ranking places. This paper proposes three new models to assess the weights. Two of them are linear programming (LP) models which determine a common set of weights for all the candidates considered and the other is a nonlinear programming (NLP) model that determines the most favourable weights for each candidate. The proposed models are examined with two numerical examples and it is shown that the proposed models cannot only choose a winner, but also give a full ranking of all the candidates.     

A multiple criteria ranking method based on game cross-evaluation approach

The application of Data Envelopment Analysis (DEA) as an alternative multiple criteria decision making (MCDM) tool has been gaining more attentions in the literatures. Doyle (Organ. Behav. Hum. Decis. Process. 62(1):87?C100, 1995) presents a method of multi-attribute choice based on an application of DEA. In the first part of his method, the straightforward DEA is considered as an idealized process of self-evaluation in which each alternative weighs the attributes in order to maximize its own score (or desirability) relative to the other alternatives. Then, in the second step, each alternative applies its own DEA-derived best weights to each of the other alternatives (i.e., cross-evaluation), then the average of the cross-evaluations that get placed on an alternative is taken as an index of its overall score. In some cases of multiple criteria decision making, direct or indirect competitions exist among the alternatives, while the factor of competition is usually ignored in most of MCDM settings. This paper proposes an approach to evaluate and rank alternatives in MCDM via an extension of DEA method, namely DEA game cross-efficiency model in Liang, Wu, Cook and Zhu (Oper. Res. 56(5):1278?C1288, 2008b), in which each alternative is viewed as a player who seeks to maximize its own score (or desirability), under the condition that the cross-evaluation scores of each of other alternatives does not deteriorate. The game cross-evaluation score is obtained when the alternative??s own maximized scores are averaged. The obtained game cross-evaluation scores are unique and constitute a Nash equilibrium point. Therefore, the results and rankings based upon game cross-evaluation score analysis are more reliable and will benefit the decision makers.     

Ranking ranges in cross-efficiency evaluations

The existence of alternate optima for the DEA weights may reduce the usefulness of the cross-efficiency evaluation, since the ranking provided depends on the choice of weights that the different DMUs make. In this paper, we develop a procedure to carry out the cross-efficiency evaluation without the need to make any specific choice of DEA weights. The proposed procedure takes into consideration all the possible choices of weights that all the DMUs can make, and yields for each unit a range for its possible rankings instead of a single ranking. This range is determined by the best and the worst rankings that would result in the best and the worst scenarios of each unit across all the DEA weights of all the DMUs. This approach might identify good/bad performers, as those that rank at the top/bottom irrespective of the weights that are chosen, or units that outperform others in all the scenarios. In addition, it may be used to analyze the stability of the ranking provided by the standard cross-efficiency evaluation.     

A new method for ranking non-extreme efficient units in data envelopment analysis

Data envelopment analysis (DEA) evaluates the performance of decision making units (DMUs). When DEA models are used to calculate efficiency of DMUs, a number of them may have the equal efficiency 1. In order to choose a winner among DEA efficient candidates, some methods have been proposed. But most of these methods are not able to rank non-extreme efficient DMUs. Since, the researches performed about ranking of non-extreme efficient units are very limited, incomplete and with some difficulties, we are going to develop a new method to rank these DMUs in this paper. Therefore, we suppose that DMU _o is a non-extreme efficient under evaluating DMU. In continue, by using “Representation Theorem”, DMU _o can be represented as a convex combination of extreme efficient DMUs. So, we expect the performance of DMU _o be similar to the performance of convex combination of these extreme efficient DMUs. Consequently, the ranking score of DMU _o is calculated as a convex combination of ranking scores of these extreme efficient DMUs. So, the rank of this unit will be determined.     

Modeling the effect of campaign advertising on US presidential elections when differences across states matter

We provide a stochastic electoral model of the US Presidential election where candidates take differences across states into account when developing their policy platforms and advertising campaigns. Candidates understand the political and economic differences that exist across states and voters care about candidates’ policies relative to their ideals, about the frequency of candidates’ advertising messages relative to their ideal message frequency, their campaign tolerance level, and vote taking into account their perceptions of candidates’ traits and competencies with their vote also depending on their sociodemographic characteristics. In the local Nash equilibrium, candidates give maximal weight to undecided voters and swing states and little weight to committed voters and states. These endogenous weights pin down candidates’ campaign and depend on the probability with which voters choose each candidate which depends on candidates’ policies and advertising campaigns. Weights vary across candidates’ policy and ad campaigns, reflecting the importance voters in each state give to the two dimensions and the variation in voters’ preferences across states.     

On the DEA total weight flexibility and the aggregation in cross-efficiency evaluations

This paper discusses the DEA total weight flexibility in the context of the cross-efficiency evaluation. The DMUs in DEA are often assessed with unrealistic weighting schemes in their attempt to achieve the best ratings in their self-evaluation. We claim here that in a peer-appraisal like the cross-efficiency evaluation the cross-efficiencies provided by such weights cannot play the same role as those obtained with more reasonable weights. To address this issue, we propose to calculate the cross-efficiency scores by means of a weighted average of cross-efficiencies, instead of with the usual arithmetic mean, so the aggregation weights reflect the disequilibrium in the profiles of DEA weights that are used. Thus, the cross-efficiencies provided by profiles with large differences in their weights, especially those obtained with zero weights, would be attached lower aggregation weights (less importance) than those provided by more balanced profiles of weights.