The latest release of version 14 is 14a, with a new feature SBM Bounded Model as an extention to SBM Max of version 13, which replaced SBM Variation model of version 12. In the real world, there are cases where input resources and/od output expansion are restricted by external constraints. SBM Bounded Model takes care of such situations, so that the outcome of SBM Bounded Model becomes more realistic than before. Note that these SBM models essentially represent KAIZEN improvement. For more details, go to newsletter 17.
With Version 11.0, we add a new feature, called Resampling DEA, which deals with data variations and gives confidence intervals of DEA scores. Input/output values are subject to change for such reasons as measurement errors and arbitrariness. DEA efficiency scores need to be examined by considering such factors, and resampling calculates the confidence interval by four new innovative models. Take a look at Newsletter 12 for more details.
Version 10.0, we have added a new feature, called Non-Convex model, which deals with S-shaped curve frontiers in production. Such S-shaped frontiers, often observed in many real data sets, are non-convex whereas traditional DEA models assume convex frontiers. This model should solve the large difference between the constant-returns-to-scale and variable-returns-to-scale scores that have puzzled many practitioners. Take a look at Newsletter 11 for more details.
The new version 9.0 includes a new feature called dynamic SBM with network structure model, which basically combines two popular functions, dynamic DEA and network DEA. It allows evaluation of networked DMU's in a multi period framework. For more details, see Latest Newsletter. The latest release is 9.0f, which solves Dynamic Network SBM models with more than 5,000 DMUs, includes modified output for WeightedSBM-C,-V ("Score") and revised samples. Note there is no change in output calculation.
Since the version 7.1, we have separate versions for 32 bit OS and 64 bit OS. We provide both versions so use the right one depending on your OS setting (32bit: DEA-Solver-Pro80n_32.xls, 64bit: DEA-Solver-Pro80n_64.xls, where n is a minor version a,b,c,..). No 64 bit version is available for Versions 6.0. Version 8.0 or earlier users can purchase version 9.0 at a half price.
The last release of version 8.0 is 8.0e, which includes a completed version of Non-Oriented SBM model. Version 8.0 included a new model called EBA (epsilon-based measure of efficiency) model, which incorporates principal component analysis in the DEA framework. This feature takes into account diversity of input/output data and their relative importance for measuring technical efficiency by unifying radial and non-radial models. Using EBM model, DEA-Solver-Pro becomes quite effective even for cases with dependent and correlated data. Read more on this in DEA Solver Newsletter No. 9.
Version 7.0 released in 2009 came with a new feature called Dynamic DEA, which evaluates DMUs in a multi-period environment. Specifically in Dynamic DEA, period to period activities are modeled as carry-over links on top of conventional components. This new feature is particularly suited for long-term planning and evaluation of DMU performance. Read more on this in DEA Solver Newsletter No. 8. Earlier releases of version 7 repaired problems with the zero measure in Network DEA and uniqueness checking in Dynamic DEA.
Version 6.0 introduced Network DEA for comparing interconnected DMUs where outputs of some DMUs become inputs to other DMUs. Network DEA is particularly useful for evaluating internal divisions and departments of large organizations. For more information on this feature, click DEA Solver Newsletter No. 7.
The last release 6.0k of Version 6 repaired a problem recently found in Bilateral SBM-C(V) models. For Vista users, use the release 6.0j or 6.0k. Note that if you use 6.0f or an earlier one, you need to create a new directory C:\SAITECH\DEA-Solver\sec before upgrading to 6.0j or 6.0k.
For the previous newsletters, click DEA Solver Newsletters.
近年DEA (資料包絡分析) 應用繁多，用於評估各種總類的效益，在不同的活動、不同的國家、不同的情境下。
Version 9.0g with Dynamic SBM with Network Structure (DNSBM) model is available for downloading. Also if you are a licensed user, you can get the revised documentation of 9.0 from the link above right. Note that the release 9.0g repairs a minor problem in Malmquist model.