Stochastic, iterative search methods such as Evolutionary Algorithms (EAs) require evaluation of the candidates which may be prohibitively expensive in many real world optimization problems. Use of meta-models or surrogates is being experimented to reduce the number of such evaluations. In this paper we investigated two such methods. The first method (DAFHEA) partially replaces expensive function evaluation by its approximate model. The approximation is realized with a support vector machine (SVM) regression model. The second one uses surrogate ranking with preference learning or ordinal regression. The fitness of the candidates is estimated by modeling their rank. Some of the classical numerical optimization functions have been used to test the techniques. The comparative benefits and shortcomings of both techniques have been identified.
|Number of pages||6|
|Journal||Australian Journal of Intelligent Information Processing Systems|
|Publication status||Published - 2010|