一种带有隐藏约束昂贵黑箱优化问题的改进响应面方法

An Improved Response Surface Method for Expensive Black-Box Optimization with Hidden Constraints

【目的】提出一种求解带有隐藏约束的昂贵黑箱优化问题的新响应面方法。【方法】对SHEBO算法进行了改进,取消了MADS强化搜索这一步骤,节约了昂贵黑箱目标函数的估值次数,并改善了响应面模型的更新策略,从而提高了算法效率。【结果】得到了新的带有隐藏约束昂贵黑箱优化问题的响应面方法。【结论】通过50个标准的测试问题对新算法的数值表现进行了评估,结果表明新算法优于原有的SHEBO方法。

[Purposes]To propose a new algorithm for solving expensive black-box optimization problems with hidden constraints.[Methods]Based on the framework of the SHEBO algorithm,the MADS intensified search step is dropped,which saves the number of expensive black-box objective function evaluations,and the update strategy of the response surface model is modified,to improve the efficiency of the algorithm.[Findings]A new response surface method for expensive black-box optimization problems with hidden constraints is obtained.[Conclusions]The numerical performance of the new algorithm is evaluated on 50 standard test problems,and the results show that the proposed algorithm is better than the original SHEBO algorithm.