遗传程序设计算法在水溶于有机溶剂过程研究应用

Dissolving Water in the Organic Solvent under Genetic Programming Algorithm

对若干物性体系进行了有机溶剂分子结构与水溶解度之间的定量研究,选取了39个有机溶剂分子,分为两组,一组30个分子作为训练集,另外9个分子作为测试集组.采用28个分子结构描述符来定量描述分子的结构,最终通过遗传程序设计筛选,得到一个4参数的非线性方程,关联和预测相关系数的平方均在0.95以上,研究取得了较满意的计算结果.

The relationship of molecular structure and its property was studied, and the changes rule of the properties with molecular structure were summed up. At the same time, the quantitative association model of the structure and properties was constructed. The discussion above not only provides quantitative forecast for the properties of the materials that have not been studied, but also saves cost and time. Moreover, it provides a necessary foundation data for the actual chemical and chemical process. Thus the quantitative research was done between the organic solvent molecule and the solubility of water. 39 organic solvent molecules were selected as study objects, and were divided into two groups. 30 elements as a group for training set, and another nine elements as a group for testing set. 28 molecular structure descriptors were used to describe quantitatively the molecular structure. At last, a nonlinear equation of four parameters was obtained through genetic screening. The square of the association and the correlation coefficient are above 0.95, and the results are satisfactory.