摘要
差分进化算法(Differential Evolution Algorithm)有较强的全局收敛能力和稳健性,应用该算法可解决工程学、计算机科学等领域的一些复杂优化问题。本文介绍一个改进DE算法,该算法适用于混料模型的近似最优设计问题,对于具有附加约束试验域的混料问题也可高效求解。最后,本文给出应用改进DE算法求解混料模型D-最优试验设计的例子。
DE(Differential Evolution)algorithm holds strong global convergence capability and robustness.which make it popular in solving complex optimization problem of many fields,such as engineering.computer science,etc.This paper shows a modified DE algorithm,which can be used for finding the optimal experimental design for mixture models on mixture region with or without additional constraints.At last,examples of solving D-optimal design for mixture model will be given to show the effectiveness of the modified DE algorithm.
作者
张新风
朱志彬
李光辉
张崇岐
ZHANG Xin-feng;ZHU Zhi-bin;LI Guang-hui;ZHANG Chong-qi(School of Economics and Statistics,Guangzhou University,Guangzhou 510006,China;Undergraduate School of Medical Business,Guangdong Pharmacentical University,Guangzhou 510006,China;School of Science,Kaili University,Kaili 556011,China)
出处
《数理统计与管理》
CSSCI
北大核心
2021年第1期26-35,共10页
Journal of Applied Statistics and Management
基金
国家自然科学基金(12071096.11671104,11926305.11901260)。
