【Abstract】 In iterative computation s for optimization algorithms, estimation of initia1 sets depends mainly on experimental guiding principles or making a guess at argument sets. In this paper a method for reductions of initial sets that is used to fix initial sets is presented in order to transform g1obal optimization into local optimization in the sense of functional va1ues, that is, an initia1 set is chosen according to functional values. It is possible to reduce blindness for fixing initial sets and improve computative reliability. We state simply the basic principle of the method in this paper. This paper is divided into four passages,the princip1e of reductions is presented in passage 2. In passage 3 we give major steps of the algorithm for reductions of initia1 sets.更多还原