Kaedah hibrid arah carian bagi masalah pengoptimuman tak berkekangan

Abstract

The aim of this research is to focus on determining the minimum values of function for unconstrained optimization. Until now, there are various of methods to solve that kinds of problems but just some of them will be discussed. In this study we have tried to built an algorithm which is using hybrid search direction of quasi-Newton and steepest descent. By using this hybrid search direction, we are trying to test on several standard problems by using BFGS, Broyden, PSB and SRI methods. We called that new method which used hybrid search direction as hybrid BFGS, hybrid Broyden, hybrid PSB and hybrid SRI methods. In determination of the efficiency of this algorithm, we compared numerical results for all those methods.