高精度的电子结构计算在小分子领域已经取得了巨大成功，如何将高精度的量子化学方法应用于复杂体系计算是当前的重点和难点课题。本项目致力于进一步发展目前已有的用于复杂体系计算的方法，并开发高效的计算程序，为复杂体系的高精度计算提供有效的工具。本项目拟在前期对分子中的簇（cluster-in-molecule, CIM）局域相关方法发展的基础之上，通过引入远程局域轨道间的相互作用，对能量进行校正，进一步提高CIM局域相关方法的计算精度。局域相关方法在单点能的计算上已经比较成熟，但是相应能量梯度的发展却比较滞后。局域相关方法的梯度发展依赖于传统方法梯度的发展。本项目中，我们将进一步改进标准MP2和RI-MP2方法的梯度计算程序，使其能非常高效地处理中小分子体系。在此基础上，我们拟改进CIM框架下MP2和RI-MP2方法的梯度算法与程序，使CIM局域相关方法能够独立地应用于复杂体系的理论研究。

Highly accurate electronic structure calculations have been applied to small molecules successfully. To extend the calculations to complex systems is one of the attractive topics. This proposal aims at further developing algorithms for calculating complex systems and implementing highly efficient quantum chemistry program. We expect it to be a practical tool for the simulations of complex systems. In the cluster-in-molecule (CIM) local correlation method, we will introduce the correlation energy corrections from distance orbital pairs to further improve the accuracy of the CIM approach. Energy gradient of local correlation methods has not been well developed yet. Energy gradient in the local correlation framework relies on the development of the conventional gradient method. In the project, we will further improve the program of MP2 and RI-MP2 energy gradients to make it treat medium-sized systems efficiently. Based on the development of conventional MP2 and RI-MP2 energy gradients, we will try to improve the energy gradient algorithm and program of the CIM-MP2 and CIM-RI-MP2 methods. With the efficient energy and energy gradient programs, the CIM local correlation method is expected to be applied to complex systems independently.