Shape structural optimization with an interior point nonlinear programmingalgorithm

The aim of this paper is to study the implementation of an efficient and re liable technique for shape optimization of solids, based on general nonline ar programming algorithms. We also study the practical behaviour for this k ind of applications of a quasi-Newton algorithm, based on the Feasible Dire ction Interior Point Method for nonlinear constrained optimization. The opt imal shape of the solid is obtained iteratively. At each iteration, a new s hape is generated by B-spline curves and a new mesh is automatically genera ted. The control point coordinates are given by the design variables. Sever al illustrative two-dimensional examples are solved in a very efficient way . We conclude that the present approach is simple to formulate and to code and that our optimization algorithm is appropriate for this problem.