An Optimum Design Procedure Based On Multipoint Approximations And A Global Optimisation Method

1 Abstract In this paper a procedure for global optimisation is presented. The procedure combines on two methods, viz. the Multipoint Approximation based on Response Surface fitting (MARS) Method and a global optimisation method. The MARS method is based on the approximation concepts according to which the original minimization problem is replaced with a succession of simpler ones formulated for approximations of the original objective and constraint functions. To solve the approximation problem with high quality approximations a global optimisation method has to be used. The global optimisation (GO) method used here is based on a one dimensional unconstrained global optimum search algorithm. Constrained optimisation problems are transformed to the unconstrained ones using the index method. The problem of dimension reduction is done by using the mappings based on Peano curves. To estimate effectiveness of the proposed procedure a number of test problems have been considered. The results are presented and discussed.