个人信息Personal Information
教授
博士生导师
硕士生导师
性别:男
毕业院校:吉林大学
学位:博士
所在单位:数学科学学院
学科:计算数学. 金融数学与保险精算
电子邮箱:yubo@dlut.edu.cn
A SYMMETRIC HOMOTOPY AND HYBRID POLYNOMIAL SYSTEM SOLVING METHOD FOR MIXED TRIGONOMETRIC POLYNOMIAL SYSTEMS
点击次数:
论文类型:期刊论文
发表时间:2014-07-01
发表刊物:MATHEMATICS OF COMPUTATION
收录刊物:SCIE
卷号:83
期号:288
页面范围:1847-1868
ISSN号:0025-5718
关键字:Mixed trigonometric polynomial system; polynomial system; symmetry; homotopy method; hybrid algorithm; symbolic-numeric computation
摘要:A mixed trigonometric polynomial system, which rather frequently occurs in applications, is a polynomial system where every monomial is a mixture of some variables and sine and cosine functions applied to the other variables. Polynomial systems transformed from the mixed trigonometric polynomial systems have a special structure. Based on this structure, a hybrid polynomial system solving method, which is more efficient than random product homotopy method and polyhedral homotopy method in solving this class of systems, has been presented. Furthermore, the transformed polynomial system has an inherent partially symmetric structure, which cannot be adequately exploited to reduce the computation by the existing methods for solving polynomial systems. In this paper, a symmetric homotopy is constructed and, combining homotopy methods, decomposition, and elimination techniques, an efficient symbolic-numerical method for solving this class of polynomial systems is presented. Preservation of the symmetric structure assures us that only part of the homotopy paths have to be traced, and more important, the computation work can be reduced due to the existence of the inconsistent subsystems, which need not to be solved at all. Exploiting the new hybrid method, some problems from the literature and a challenging practical problem, which cannot be solved by the existing methods, are resolved. Numerical results show that our method has an advantage over the polyhedral homotopy method, hybrid method and regeneration method, which are considered as the state-of-art numerical methods for solving highly deficient polynomial systems of high dimension.