Computer-Aided Molecular Design: Fundamentals, Methods, and Applications

Abstract

Computer-aided molecular design (CAMD) includes the identification of molecular structures based on the specification of desired performance indices in a reverse problem formulation. This article provides an overview of fundamental concepts used in CAMD pertaining to molecular representations, property prediction models, and algorithms for generation of structures with respect to performance criteria. This is followed by a classification of molecular design problems considering single- and multiobjective formulations, mixtures and blends, uncertainty, incorporation of process and control design into molecular structure decisions, reactive systems, and quantum mechanical molecular models. The solution of these problems is further discussed through a review of algorithmic approaches that include generate and test methods, deterministic and stochastic optimization algorithms, and property clustering techniques. A review of advanced methodologies is then presented, which combine different algorithms, problem representation approaches, and computational technologies to enable efficient solution of various problem classes. A review of applications in numerous domains is finally provided, including solvents, adsorbents, ionic liquids, materials for CO2 capture, heat exchange fluids, polymers, bio-based fuels, formulated products, and other chemicals.