Upgrading inefficient decision making units (with negative data) towards common weights (using DEA)

The main purpose of this paper is to upgrade and improve inefficient units by common weights obtained from all units studied. In fact, we consider the common weight vector as the direction in which inefficient units rise. The methodology of this research is to consider the semiessential radial model and we want to use the duality of this model to find the common weights of inputs and outputs, some of which are negative. For this purpose, we present a multi-objective problem of generating common weights and use ideal programming to solve it, which leads to the production of a nonlinear problem, which for this particular problem, by a linearization method, is called We turn a linear programming problem. Since the necessary and sufficient condition for the boundary of the semi-essential radial model in the nature of input (output) is that there is an input (output) with at least one positive value, so we observe this condition here. Finally, we will explain our method with an example and the remarkable thing about the promotion method in the present study is that negative data is promoted and improved as negative data.