Mathematical Sciences: Numerical Methods for Molecular Dynamics

首席研究员研究和开发适用于并行和矢量计算机的计算方法来解决分子动力学问题。他强调时间步进技术来计算多体相互作用中的运动，研究辛方法的重要性，并研究采取更大时间步长的策略。为了验证这些想法，首席研究员在一个用于建模大分子的软件包中实现了这些想法。多体问题——一组不同的物体如何在随着物体运动而变化的力的影响下运动——是经典而困难的。例如，太阳系中行星的运动可以被认为是一个十体问题，其中主要的力量是重力。分子的运动尤其困难，因为这类问题涉及如此多的物体，而力的性质又如此复杂。这些问题出现在材料科学、结构生物学和药物设计领域。这项工作旨在开发软件和算法，以便在计算机上快速解决这些问题。

The principal investigator studies and develops computational methods suitable for parallel and vector computers to solve molecular dynamics problems. He emphasizes time-stepping techniques to compute motion in many-body interactions, studies the importance of symplectic methods, and investigates strategies for taking larger time steps. To test the ideas, the principal investigator implements them in a software package for modeling macromolecules. The many-body problem-how a set of distinct objects move under the influence of forces that change as the objects move-is classical and hard. For instance, the motion of planets in the solar system can be thought of as a 10 body problem where the primary force is gravity. The motion of molecules is particularly difficult because such problems involve so many objects and the nature of the forces is so complicated. The problems arise in areas of material science, structural biology, and design of drugs. This work aims to develop software and algorithms to solve such problems rapidly on computers.