Evaluating performance of cross-docks: A data envelopment analysis approach

This research develops an iterative two-stage approach for evaluating the performance of 3rdparty cross-docking services. it is an extension of ras two-stage model and addresses the challenge in data envelopment analysis (dea) where the number of inputs/outputs (both discretionary and nondiscretionary) is large relative to the number of decision making units (dmus). in this approach, each iteration uses dea in first-stage to rank cross-docks using a small subset of discretionary variables and regression in second-stage to adjust rankings based on non-discretionary variables. the key to selecting the variable subsets is using orthogonal arrays so that all discretionary variables and their combinations appear in the same number of iterations, thus ensuring equal consideration in the final analysis. this approach results in a set of adjusted dmu rankings (one for each iteration) that can be statistically analyzed to detect significant differences in dmu performance and to identify the discretionary inputs that have the greatest impact on technical efficiency.

Evaluating performance of cross-docks: A data envelopment analysis approach Academic Article