Computer Algebra of Polynomials and Rational Functions
Computer Algebra of Polynomials and Rational Functions
复制标题
多项式和有理函数的计算机代数
DOI:
10.1080/00029890.1973.11993365
复制
发表时间:
1973
影响因子:
0.5
通讯作者:
G. Collins
中科院分区:
文献类型:
--
作者:
G. Collins
1. Introduction. Computer programs are now available for performing many important algebraic processes of interest and utility to pure and applied mathematicians. Also, many interesting mathematical problems arise in the development and analysis of algorithms for use in such programs. As a mathematician-turned-computerscientist who has contributed to realizing the current state of computer algebra capabilities and who is intrigued with the mathematical problems whose solutions will contribute to further progress, I hope in this brief survey article to impart some of my knowledge a bout this subject and to transmit some of my enthusiasm for it to other mathematicians.1 he kind of computer algebra I will discuss is concerned mainly with polynomials and rational functions, for the following reasons. First, the allotted space does not permit me to be more ambitious. Second, this is the only part of the subject in which I can really be authoritative. Third, this is the part of the subject about which the most is known and on which other parts most depend. I shall be concerned with polynomials and rational functions in several variables, primarily with rational integer coefficients, but this will also lead to consideration of rational number coefficients and finite field coefficients; algebraic number coefficients will receive some mention as an advanced topic of current research. Operations on polynomials and rational functions which will be considered include the" arithmetic" operations of addition, subtraction, multiplication and division. I shall show in Section 3, perhaps to the reader's surprise, that even polynomial arithmetic is non-trivial and interesting when the objective is to design and analyze optimal algorithms.Algorithms for the arithmetic operations on rational functions require an efficient algorithm for polynomial gcd (greatest common divisor) calculation. Research over the last seven years has revealed that there are numerous versions of the" Euclidean algorithm" for polynomials which differ dramatically in their efficiency. Also, within the last four years," modular" polynomial gcd algorithms have been devised, which depend on use of the Chinese remainder theorem, and which are orders of magnitudes faster than any of the non-modular Euclidean algorithms.