OPTIMAL COMPUTER-AIDED MOLECULAR DESIGN - A POLYMER DESIGN CASE-STUDY

This paper addresses the problem of optimally designing molecular prod ucts. A systematic analysis framework is presented for transforming a class of optimal computer-aided molecular design problems with nonline ar structure-property functionalities into equivalent mixed-integer li near (MILP) problems. While, in general, it is not possible to solve t he original problem formulation for the best molecular design with mat hematical certainty, the equivalent (MILP) reformulation can be solved efficiently with existing solvers and identify not only the best, but also the second, third, etc., best molecular designs. Two alternative design objectives are considered: (i) minimization of the scaled devi ation of design properties from some target values, property matching, and (ii) minimization/maximization of a single property subject to lo wer and upper bounds on the rest of the properties, property optimizat ion. The framework is applied to the design of polymers where thermoph ysical and mechanical properties are estimated using group contributio n methods; Three case studies, including comparisons with existing, me thods, illustrate the computational efficiency and feasibility of the proposed methodology.