## Using Steepest Descent to Find Energy-Minimizing Maps Satisfying Nonlinear Constraints

Description:
The method of steepest descent is applied to a nonlinearly constrained optimization problem which arises in the study of liquid crystals. Let Ω denote the region bounded by two coaxial cylinders of height 1 with the outer cylinder having radius 1 and the inner having radius ρ. The problem is to find a mapping, u, from Ω into R^3 which agrees with a given function v on the surfaces of the cylinders and minimizes the energy function over the set of functions in the Sobolev space H^(1,2)(Ω; R^3) having norm 1 almost everywhere. In the variational formulation, the norm 1 condition is emulated by a constraint function B. The direction of descent studied here is given by a projected gradient, called a B-gradient, which involves the projection of a Sobolev gradient onto the tangent space for B. A numerical implementation of the algorithm, the results of which agree with the theoretical results and which is independent of any strong properties of the domain, is described. In chapter 2, the Sobolev space setting and a significant projection in the theory of Sobolev gradients are discussed. The variational formulation is introduced in Chapter 3, where the issues of differentiability and existence of gradients are explored. A theorem relating the B-gradient to the theory of Lagrange multipliers is stated as well. Basic theorems regarding the continuous steepest descent given by the Sobolev and B-gradients are stated in Chapter 4, and conditions for convergence in the application to the liquid crystal problem are given as well. Finally, in Chapter 5, the algorithm is described and numerical results are examined.

Date:
August 1994

Creator:
Garza, Javier, 1965-

Partner:
UNT Libraries