Melnik, R.V.N., Uhlherr, A., Hodgkin, J. and de Hoog, F.
Polymeric materials and polymer-based composites became an intrinsic part of many new industrial materials technologies where structural engineering and material engineering research go hand in hand. This research relies heavily on the fact that the material from which the end-use structure is made can be "designed" in principle and adapted to suit specific engineering purposes. A core component in this research, both from the technological and theoretical point of view, is to be able to extract the information about the macroscopic (thermomechanical, structural) properties of polymers from their microscopic (atomistic) constitution.
Since microscopic, atomistically detailed models of materials provide a key to a better understanding of their behaviour at the macroscopic level, the development of such models and algorithms for their implementation become an important activity in industrial materials research, in particular in polymer and composite sciences. The computer-aided molecular design of structures using effective algorithms leads to the situation where compounds and composites worthy of actual synthesis can be effectively identified.
In solving molecular conformation problems, stochastic methods based on Molecular Dynamics (MD) and Monte Carlo (MC) simulations have become quite popular among the researchers. However, the computer time required for full conformational search by the MD and MC algorithms and many other stochastic search methods (genetic algorithms, e.g.) is often hardly feasible for real polymeric systemsand composites. The situation is not better for the majority of deterministic methods that, due to their exhaustive search of the conformational space, are becoming less feasible computationally when designing complex new material structures. Despite recent progress, all these methods are still computationally very demanding and, most importantly, the success of their applications to realistic polymer systems relies heavily on the choice of initial conformation. Therefore, a real challengefor these types of algorithms is to build plausible initial configurations which can then be relaxed effectively. Although for complex polymeric systems and composites this has often proved to be a difficult task, it has to be tackled in order to provide the real opportunity to design, characterize and optimize these materials before undertaking expensive experimental work. In this contribution we argue that moving towards atomistic simulations of complex polymer systems and composites, another group of stochastic search methods, known under its generic name as distance geometry (DG), will play an increasingly important role. Formulating the general distance geometry problem in atomistic simulation of materials as an optimization problem, we apply two distance geometry algorithms to its solution and analyze their performance for bulk decane molecules. The comparative analysis of the described algorithms in applications to cross-linked polymeric structures is a natural extension of the presented work.
Key words: atomistic simulations; complex systems; distance geometry algorithms; Molecular Dynamics and Monte Carlo; conformational search; stochastic search methods; initial conformations' polymeric materials; polymer-based composites; mathematical modeling; macroscopic properties from microscopic constitution; computer-aided molecular design; optimization.