We consider a new class of Data Envelopment Analysis (DEA) modeling, which we call ‘sequential DEA’. This new approach is a relatively simple generalization of the standard and popular in practice DEA. It allows for analyzing efficiency of the decision making units that consist of a sequence of sub-DMUs (e.g., branches of banks, hospital holding company running a number of hospitals at different locations, hotel chains, etc.). The approach is embedded in the Hilbert sequence space (\(\ell ^{2}\)) and therefore it allows for potentially different numbers of the sub-DMUs as well as different numbers of inputs and outputs used by different decision making units. We hope this approach will open up a new stream of literature in the sense that many existing variations from the already rich literature on DEA can be adapted to this approach.

我们考虑一类新的数据包络分析（DEA）建模，我们将其称为“顺序DEA”。这种新方法是标准的相对简单的概括，并且在实践中非常流行。它允许分析由一系列子DMU（例如，银行的分行，在不同位置经营多家医院的医院控股公司，酒店连锁店等）组成的决策单位的效率。该方法嵌入在希尔伯特序列空间（\（\ ell ^ {2} \）中），因此它允许潜在数量不同的子DMU以及不同决策单元使用的不同数量的输入和输出。我们希望这种方法将开辟新的文献流，因为从已经丰富的DEA文献中可以找到许多现有的变体，可以适应这种方法。