| Historically, corporations have been evaluated on traditional financial analysis. Internally, firms have also considered operational performance measures. Inclusion of quality measurements is also appropriate today with the adoption of the total quality management (TQM) philosophy by most companies.; Evaluating performance is then essentially a multiple objective decision analysis problem. There are a number of Management Science techniques to solve this type of problem, but almost all of these require the introduction of subjective weights and preferences.; Data envelopment analysis (DEA) is a technique that does not require the introduction of subjective preferences. However, there have been three problems in using DEA. First, DEA frequently has low discriminating power between the units being evaluated. Secondly, there is a lack of discrimination of the effect of the various performance measures employed. Finally, DEA can assign inappropriate weights to common or similar variables in more than one objective. Here, a variant model is developed to address these issues. The discriminating power is enhanced by employing more differentiating criteria as the objective functions, rather than just the traditional efficiency measure. The impact of the various performance measures is also discerned as well as additional control of the assignment of weights by using another multiple objective extension. In addition, quality measurements are added to the model.; Beyond providing a more effective and comprehensive evaluation of historical performance, DEA can be utilized dynamically to assess the potential impact of contemplated decisions on future efficiency. Heretofore, DEA has only been applied statically. This dynamic utilization of DEA can provide the basis of a decision support system (DSS). To illustrate this, forecasting is combined with data envelopment analysis to form the basis of such a decision support system. Simulation is then utilized to provide a sensitivity analysis with regard to the management-controlled inputs relative to the forecasted variable(s). This provides management with an indication of the potential impact of their decisions and the tractability of the resulting process by considering a range rather than a point estimate.; Finally, the two models, the variant DEA model, which addresses the main inadequacies of traditional DEA, and the dynamic application of DEA are illustrated by examples that apply to the freight service business. |
