A hybrid algorithm for the solution of a single commodity spatial equilibrium model

L. Portugal and J. Judice

In this paper we propose a hybrid algorithm for the solution of a large-scale single commodity spatial equilibrium model. This model can be stated as a Linear Complementarity Problem (LCP) with a singular Symmetric Positive Semi-Definite (SPSD) matrix whose structure is closely related with the network structure of the model. The hybrid scheme is a combination of a Predictor-Corrector (PC) and a Parametric Principal Pivoting (PPP) algorithms and its implementation takes full advantage of the structure of the matrix of the LCP. We report computational experience on the solution of large-scale spatial equilibrium problems with up to 1000 regions that shows the great efficiency of the approach discussed in this paper.