Application of multiobjective optimization in the deisgn of simulated moving bed systems for chiral drug separation

手性药物的销售接近全球所有药物销售的三分之一。用于手性药物的普通化学生产方法产生包含具有相同化学组成但在空间上不同结构取向的对映体的外消旋混合物。手性药物的一种异构体对人体具有治疗作用，而其对映体对人体有害，这是很常见的。当一种异构体是“好”的而另一种是“坏”的时，分离两种对映异构体以提高其安全性和耐受性有明显的益处。模拟移动床（SMB）系统是一种可行的技术，用于获得纯的单一对映体。近年来，由于手性固定相和非线性色谱理论的发展，以及严格的药品管理政策，制药行业对使用SMB技术进行对映体分离的兴趣越来越大。通过对SMB进行优化设计，可以提高生产率，满足严格的产品质量要求，并大幅节省成本。在本计画中，我们拟详细研究手性药物的分离与纯化，特别是含有两种以上对映体的外消旋混合物。我们的主要重点是新的设计和开发，以及多目标优化和先进的实时控制应用到SMB系统。将开发详细的模拟模型，随后将通过实验加以验证。此后，将使用多个目标和约束来计算最佳操作条件。一个适应的国家的最先进的优化技术，遗传算法，将使用。最后，将通过实验验证最佳操作性能。这些研究将有助于优化多个目标，同时满足工业中遇到的众多现实限制。

The sale of chiral drugs is close to one third of all drug sales worldwide. Ordinary chemical production methods for chiral drugs produce a racemic mixture containing enantiomers having identical chemical compositions but different structrtal orientation in space. It is quite common to find that one isomeric form of a chiral drug has a therapeutic effect on the human body while its enantiomer is harmful. When one isomer is 'good' and the other 'bad', there is obvious benefit from separating the two enantiomers to enhance its safety and tolerability. The Simulated Moving Bed (SMB) system is a viable technology for obtaining pure single enantiomers. Lately, there has been increased interest in the pharmaceutical industry for using SMB technology for enantio-separations due to recent developments in chiral stationary phases and nonlinear chromatographic theory, as well as stringent drug administration policy. Enhanced productivity satisfying stringent product qualities and significant cost savings are possible if optimal design of the SMB is made. In this project, we propose to study in detail the separation and purification of chiral drugs, particularly, for racemate mixtures containing more than two enantiomers. Our main emphasis is in new design and development, and application of multi-objective optimization and advanced real-time control to SMB systems. Detailed simulation models would be developed, which will subsequently be verified experimentally. Thereafter, optimal operating conditions will be computed using multiple objectives and constraints. An adaptation of the state-of-the-art optimization technique, genetic algorithm, will be used. Finally, optimal operating performance would be validated experimentally. These studies would help optimize several objectives while simultaneously satisfying numerous real-life constraints encountered in industry.