Design Theory and Methods using CAD/CAE

3.7. Non-Gradient Approach

Unlike the gradient-based approach, the non-gradient approach uses only the function values in the search process, without the need for gradient information. The algorithms developed in this approach are very general and can be applied to all kinds of problems—discrete, continuous, or nondifferentiable functions. In addition, the methods determine global optimal solutions as opposed to the local optimal determined by a gradient-based approach. Although the non-gradient approach does not require the use of gradients of objective or constraint functions, the solution techniques require a large amount of function evaluations. For large-scale problems that require significant computing time for function evaluations, the ...