The mixed finite element method has advantages over classical formulations of Lagrange and Castigliano, and ensures high precision and smoothness approximate solutions for strains and stresses.
In practice, using the mixed finite element method is limited by high size of linear equation systems, which leads to significant increasing time costs.
Using the orthogonal finite functions for approximating of unknowns leads to generation of sparse linear systems with a special structure.
That system can be simplified during solution.
This article proposes the algorithm of modificated Gaussian elimination for preliminary transformation an extended matrix of linear system.
This algorithm underlies software complex, which includes LISTSOLVER, the author's solver.
Confirmed the efficiency of created solver in compare with common-type solver.
Keywords: mixed finite element method, sparse linear systems, orthogonal finite functions, Gaussian elimination