Optimization of Continuous Heterogeneous Models

A heterogeneous model consists of a solid model and a number of spatially distributed material attributes. Much progress has been made in developing methods for construction, design, and editing of such models. We consider the problem of optimization of a heterogeneous model, and show that its representation by a continuous function defined over a constructively represented domain naturally leads to simple and effective optimization procedures. Using minimum compliance optimization problem as an example, we show that the design sensitivities are directly obtainable in terms of material and geometric parameters, which can be used in any standard gradient-based optimization procedures. The proposed approach allows both local control of the material properties and global control of geometric variations, and can be used with many existing techniques for material modeling. Numerical experiments are given to demonstrate these representational advantages.