Abstract: The compound method is a direct way for solving constrained nonlinear planning problems. It is simple in calculation, wide in application, and common in engineering technology. But, a weakness of this method is the reducing dimen sion phenomenon when you choose the initiating compound inappropriately, which directly offects the convergence of the optimum solution. The reducing dimension control theorem in the space of n dimensions is given, by wh ich it is possible to avoid the reducing dimension phenomenon during constructing a suitable initiating compound.

Key words: nonlinear planning, compound method, reducing dimension control