Optimization of Formulations Using Robotic Experiments Driven by Machine Learning DoE
Optimization of Formulations Using Robotic Experiments Driven by Machine Learning DoE
复制标题
DOI：
10.1016/j.xcrp.2020.100295
复制
发表时间：
2021-01-20
影响因子：
8.9
通讯作者：
Lapkin, Alexei A.
中科院分区：
文献类型：
--
作者：
Cao, Liwei;Russo, Danilo;Lapkin, Alexei A.
配方产品是成分的复杂混合物，由于缺乏所需特性的一般可预测物理模型，其上市时间可能难以加快。在这里，我们报告了机器学习分类算法与汤普森采样高效多目标优化（TSEMO）算法的耦合，用于同时优化连续和离散输出。该方法成功地应用于没有物理模型可用的商业利益的配方液体产品的设计。实验以半自动的方式进行，使用由机器学习算法触发的机器人平台。该程序允许在15个工作日内找到9个符合客户定义标准的合适配方，在配方的目标性能方面优于人类直觉。
Formulated products are complex mixtures of ingredients whose time to market can be difficult to speed due to the lack of general predictable physical models for the desired properties. Here, we report the coupling of a machine learning classification algorithm with the Thompson sampling efficient multiobjective optimization (TSEMO) algorithm for the simultaneous optimization of continuous and discrete outputs. The methodology is successfully applied to the design of a formulated liquid product of commercial interest for which no physical models are available. Experiments are carried out in a semiautomated fashion using robotic platforms triggered by the machine learning algorithms. The procedure allows one to find nine suitable recipes meeting the customer-defined criteria within 15 working days, outperforming human intuition in the target performance of the formulations.