On dual-complementary variational principles in mathematical physics (Q2264984)

A general theory of complementary-dual variational principles is developed for a wide class of linear boundary value problems in mathematical physics. In particular, the class of problems considered is described by equations of the form \(T^*ETu+ f= 0\), where \(T\) is a linear operator from a Hilbert space \(\mathcal U\) into \(\mathcal G\), \((T^*\) is the adjoint of \(T\), \(T^*\colon \mathcal V\subset \mathcal G\to \mathcal F\), \(E\) is a symmetric operator \(E\colon \mathcal V\to \mathcal V\), and \(f\) is a given function in \(\mathcal F\). It is shown that this equation can be decomposed into three canonical equations, and that there exist a dual pair of canonical equations (with each pair of primal set) that give rise to equations of the type \(S^*E^{-1}S\chi + \eta = 0\). A number of variational principles are constructed for each set of canonical equations. These variational principles can be used to construct mixed models for approximating two or more variables independently. Examples of the theory are given from particle dynamics, viscoelasticity, electrostatics, magnetostatics, fluid mechanics, heat conduction, resistive net-works, transportation networks, etc.