Method for optimizing engineering parameters by using Bayesian optimization algorithm

Abstract

The invention discloses a method for optimizing engineering parameters by using a Bayesian optimization algorithm. The method is used for optimizing performance parameters having a function relationship with the engineering parameters under an engineering constraint condition. According to the method, a new acquisition function is defined, an adaptive target value ytarget (k) is introduced into the acquisition function to replace an optimal solution item ymin of a traditional expectation improvement criterion in the prior art, and an adaptive weight coefficient item wk is introduced, so that the traditional expectation improvement criterion is evolved into an adaptive jump weighted expectation improvement criterion. According to the method, the situations of excessive development and excessive exploration can be reduced to a certain extent, the optimization efficiency is improved, the waste of computing resources is reduced, and a better optimization effect is obtained.

Description

technical field [0001] The present application relates to the field of engineering design, in particular to a method for optimizing engineering parameters using a Bayesian optimization algorithm, especially a method for optimizing shape parameters of a turbine disk. Background technique [0002] In the field of engineering design, it is often faced with the situation that under the condition of engineering constraints, it is necessary to design engineering parameters to make the performance close to the limit. At this time, if there is a known functional relationship between performance and engineering parameters, and the function is a complex function or the extremum needs to be expensive, the Bayesian optimization algorithm can be used to determine the engineering parameters to solve the problem. For example, in the design process of a turbine disk, the maximum allowable equivalent stress is the engineering constraint, and the engineering designer needs to design the geome...

AI Technical Summary

Problems solved by technology

When the traditional expectation improvement criterion selects new evaluation points, there will be "over-exploitation" or "over-exploration", which will not only cause Bayesian optimization to easily fall into local optimum (that is, converge to a local optimal solution) , and it will also waste computing resources and reduce optimization efficiency

Therefore, in view of the inadequacy of the expectation improvement criterion that cannot be adjusted according to actual needs for local development or global exploration, it is urgent to invent a strategy that can automatically balance local development and global exploration on the basis of the traditional expectation improvement criterion. Improving the optimization performance of Bayesian optimization methods

Images