In this paper, a new method for the identification of self-optimizing control structure designs (CSDs) based on generalized singular value decomposition (GSVD) is proposed. The method is primarily dedicated to find optimal CSDs where all controlled variables (CVs) are represented by a common set of linear combinations of process variables (PVs). It is shown that the implementation of the GSVD into iterative solution approaches is beneficial in order to find CSDs where an individual PV subset is mapped to each CV. The developments will be tested on a simple process.