團隊負責人舒適教授
團隊所研發的JXPAMG解法器對比實驗結果
團隊科研成果獲獎
團隊人才培養獲獎
偏微分方程大規模離散系統的高效并行求解是科學工程計算領域的研究熱點和瓶頸問題。湘潭大學“高效并行算法與解法器”研究團隊在舒適教授的帶領下，長期圍繞代數多重網格（AMG）法和非重疊區域分解法這兩類國際上流行的求解算法開展研究，在國內率先系統研究了基于部分幾何與分析信息的AMG法，是AMG解法器的主要研發團隊之一。團隊成員主要有：舒適教授、馮春生副教授、岳孝強副教授、王俊仙副教授、喻海元教授、袁海專副教授、張建平副教授、楊玉月副教授、朱瑋博士、周潔博士等。
團隊面向我國國防建設和社會發展的重大需求，承擔了國家科技重大專項課題（軍口）、國家863計劃和973計劃子課題、國家自然科學基金和部省級重點項目等10余項，多個課題被評為優秀。在國際計算數學權威刊物SIAM系列期刊、Math. Comput.、Numer. Math.和中國科學等國內外學術刊物發表論文150余篇，其中被SCI收錄80余篇；獲教育部自然科學一等獎和湖南省自然科學一等獎各一項。與北京應用物理與計算數學研究所共同研制的并行AMG解法器已成為高性能并行計算軟件平臺JASMIN的默認線性解法器，綜合性能優于國際上流行的同類解法器，大幅提高了一些國防應用模型問題的計算效率。團隊培養博士和碩士研究生50余名，多人獲全國優秀博士學位論文、數值分析Fox獎、湖南省優秀博士和碩士論文、湘潭大學研究生校長特等獎等。
Research Team for "Efficient Parallel Algorithms and Solvers" of Xiangtan University
Efficient parallel computation of large-scale discrete systems of partial differential equations is a hot topic and bottleneck problem in the field of scientific engineering computing. Our group(main members: Shi Shu, Chunsheng Feng, Xiaoqiang Yue, Junxian Wang, Haiyuan Yu, Haizhuan Yuan, Jianping Zhang, Yuyue Yang, Wei Zhu, Jie Zhou) have conducted long-term research around the Algebraic Multigrid (AMG) method and the non-overlapping domain decomposition method which are two classes of popular algorithms. We took the lead in studying systematically the AMG method based on part geometry and analysis information, and is one of the most influential teams to study the AMG method.
The group has long carried out research on the major needs of national defense construction and social development, and has undertaken more than 10 projects including National Major Scientific and Technological Special Project (military), 863 and 973 sub-projects, National Natural Science Foundation and Province (or Ministry) key projects. Many projects were rated as excellent. The group has published more than 150 papers in the journals such as SIAM series, Math. Comput., Numer. Math., Science China and so on, over 80 papers of which were collected by SCI, and was awarded the first prize in Natural Science of the Ministry of education and the first prize of Natural Science in Hunan province. The parallel AMG solver which was developed by cooperating with the Institute of Applied Physics and Computational Mathematics in Beijing has become the default linear solver of high performance parallel computing software platform JASMIN, and its comprehensive performance is better than that of the similar international popular solvers. The solver greatly improves the computational efficiency of some defense application model problems. More than 50 PhDs and masters have been trained, several students among them were awarded the National Excellent Doctoral Dissertations, the Leslie Fox prize for numerical analysis, excellent doctorate and master's thesis in Hunan, and the special prize of the graduate principal of the Xiangtan University.
