The general CAMD problem is first decomposed into two parts molecular design and mixture design.If we are interested in pure component solvent design then only the first part is followed while if we are interested in mixture design both parts are fol

The general CAMD problem is first decomposed into two parts molecular design and mixture design.If we are interested in pure component solvent design then only the first part is followed while if we are interested in mixture design both parts are fol
The general CAMD problem is first decomposed into two parts molecular design and mixture design.If we are interested in pure component solvent design then only the first part is followed while if we are interested in mixture design both parts are followed (see Fig.1).In the first part (pure component design) the problem is decomposed into four sub-problems.The first sub-problem considers the structural constraints that result in generation of feasible molecular structures.The second sub-problem considers the pure component properties,and the feasible molecular structures from the first sub-problem are solved for the pure component properties.Those molecules,which satisfy the pure component property constraints,are then passed into the third sub-problem,which considers the mixture properties.Those satisfying the mixture property constraints are passed on to the final sub-problem where the process model constraints are considered along with the objective function and the optimal solvent is identified by either solving a smaller MINLP problem or a series of NLP problems or just ranking the solvents according to the performance index,depending on the number of feasible solutions that are passed into the final sub-problem.

The general CAMD problem is first decomposed into two parts molecular design and mixture design.If we are interested in pure component solvent design then only the first part is followed while if we are interested in mixture design both parts are fol
一般CAMD问题首先被分解为两部分的分子设计和混合设计.如果我们在纯组分溶剂设计有兴趣的则只有第一部分是跟随而如果我们在混合料设计有兴趣的两个部分之后（见图.1）.在第一部分（纯组件设计）的问题分为四个子问题分解.第一子问题认为,在可行的分子结构产生的结果结构性制约.第二次问题认为纯组件属性,并从第一个分问题的可行分子结构的纯组件属性解决.这些分子,满足纯组件的属性限制,然后传递到第三次的问题,认为混合属性.那些满足混合属性限制传递到最后的子问题,即该过程模型的限制,我们考虑了目标函数和最优溶剂沿要么是解决一个问题或一个较小的MINLP NLP的问题只是确定排名或系列根据溶剂的性能指标,具体取决于正在进入最后的子问题,通过可行的解决办法.