Generalized symmetric weight assignment technique: Incorporating managerial preferences in data envelopment analysis using a penalty function

In this paper, we generalize the Symmetric Weight Assignment Technique to incorporate all managerial preferences in Data Envelopment Analysis (DEA). This is a method that promotes managerial preferences, while not changing the feasibility region of the nominal DEA model, unlike standard techniques such as cone ratio and assurance region that potentially yield infeasible problems. We discuss how this generalization is motivated by a real word problem where the decision maker's preference is not precisely known. We also discuss how we are using our generalization in solving a workforce allocation problem for the United States Navy.

► We show how to promote managerial preferences in DEA without changing the feasibility region.
► We discuss the differences of our method and other DEA weight restriction methods.
► We extend SWAT to incorporate arbitrary managerial preferences.
► We discuss how we are using the proposed method to solve a real world problem.