(61g) Adjustable Robust Optimization for the Synthesis of Continuous Rufinamide Manufacturing Process Under Uncertainty

Rufinamide is a new class of antiepileptic drugs working by regulating sodium ion channels in the people’s brain voltage gate. It can assist to treat a kind of typical epilepsy, Lennox-Gastaut Syndrome. Rufinamide was first successfully developed by Eisai's North American subsidiary and approved by the U.S. Food and Drug Administration in 2008. The drug Banzel containing rufinamide as the API, is currently priced at ~ US $3.75 / mg tablet and ~ US $3.90 /mL liquid suspension according to its seller.

The traditional batch synthesis process of rufinamide has some disadvantages such as low yield, complicated procedure, high labor cost and poor environmental benefit. Therefore, the continuous manufacturing methods of rufinamide have attracted extensive attention. In recent years, the continuous routes for synthesizing rufinamide have been improved and enriched continuously. However, each route at different synthesis stages has its own advantages and disadvantages. It is of great significance to identify the optimal rufinamide synthesis process from the viewpoint of process systems engineering via the superstructure optimization.

In this work, we developed a superstructure contains more than 500 possible synthetic routes for the continuous rufinamide manufacturing. The deterministic mixed-integer nonlinear programming (MINLP) was correspondingly formulated to identify the best synthetic route. Furthermore, we formulated a two-stage adjustable robust optimization model to hedge against potential uncertain factors such as yields of key conversion step as well as the energy consumption coefficient. The tailored column-and-constraints algorithm was proposed to solve the above two-stage adjustable robust optimization model. The price of robustness of the adjustable decisions on optimal synthesis of the continuous manufacturing process was finally examined.