Ordering in Solving Systems of Equations 1).pdfVIP

MM Research Preprints,33–41 No. 19, Dec. 2000. Beijing 33 Ordering in Solving Systems of Equations 1) Kun Jiang, Xiao-Shan Gao Institute of System Sciences, AMSS, Academia Sinica Beijing 100080, P.R. China E-mail: (kjiang, xgao)@mmrc.iss.ac.cn Abstract. To use a “good” variable order is one of the effective ways to prevent the occurrence of large polynomials in an elimination method. In this paper, we present 3 an algorithm of O(n ) to find such a “good” order using algorithms from graph theory. Based on this algorithm, we present a new version of Wu-Ritt’s zero decomposition algorithm, which allows different orders for different triangular sets in the decomposition formula. 1. Introduction It is well known that Wu’s method is very fast in proving geometry theorems[?]. Many difficult theorems whose traditional proofs need an enormous a mount of human intelligence, such as Feuerbach’s theorem, Morley’s trisector theorem, etc., can be proved by computer programs based on Wu’s method within seconds[?]. This method first introduces a coordinate system so that the geometric entities and relations are turned into a system of equations, and then decides in an algorithmic man- ner whether the conclusion is valid on the solutions under some non-degenerate conditions. Therefore, Wu’s method is essentially based on an algebraic elimination theory. One of the reason behind the success of Wu’s method is that we may choose a nice order for the points in a geometry theorem. This order will lead to a variable order which will make the elimination procedure easy.