Abstract:

We propose a new collaborative optimization (CO) method with global fast optimization to solve the problems of too many times of iterations and local optimal solutions of CO. A new slack factor is introduced into system optimization, and optimal design points can be fast converged to extreme points by the improved dynamic slack factor. Static slack factor enables optimal design points to jump out of local extreme points, guaranteeing that the results of the system objective function are global optimal solutions. The objective function of subsystem is divided into two parts: consistent objective function and subsystem optimal objective function, which are added up with different weights as the subsystem objective function. Thus both the consistence and the independence of subsystems are taken into account. The improved CO (ICO) is validated via the examples of reducer. Simulation results show that on the premise of ensuring a smaller constrained maximum value, the ICO can quickly get the global optimal solution and has good robustness.

Key words: collaborative optimization, slack factor, consistency, stability