Output details
12 - Aeronautical, Mechanical, Chemical and Manufacturing Engineering
University College London
Approximate multi-parametric programming based B&B algorithm for MINLPs
Model-based optimal synthesis and design of chemical processes and products results in computationally complex combinatorial optimisation problems involving integer variables (representing the existence or not of process units, molecular group etc.) and continuous variables (representing temperature, viscosity etc.). An algorithm which performed better than the state-of-the-art techniques is presented. The novelty is in mathematically reformulating the integer variables as parameters and then obtaining only an approximate parametric solution of the resulting multi-parametric nonlinear program, to efficiently guide the search of the optimal solution. This work was presented at an international conference and then invited as a full length journal paper.