To provide an optimizing method which represents the performance of a system in a search space.

The system has design parameters and restriction conditions of a target function modeled by this system. The restriction conditions and target function are grouped according to plural unique hierarchical levels. The group including the target function has the highest level. According to the grouping of the restriction conditions and target function, the design parameters are evaluated in the increasing order of the levels. As far as a relative level is not the highest level, the performance is immediately allocated to the design parameters when the evaluation of a design parameter becomes negative, the allocated performance is evaluated by the target function, and the search space is divided into sections according to the level.