Computing the Sign of a Dot Product Sum

A real number usually cannot be exactly represented by a floating-point number in a computer. Namely, a floating-point number frequently stands for any real number in a specific interval. In this paper, we present a method for computing the sign of a dot product sum. Each initial datum that is a floating-point number is considered as an interval. With interval analysis and floating-point summation methods, an explicit formula for calculating the minimal interval of a dot product sum is presented. Error analysis and some examples are provided as well.

[1]  Luc Jaulin,et al.  Applied Interval Analysis , 2001, Springer London.

[2]  Glauco Masotti.  Floating-point numbers with error estimates , 1993, Comput. Aided Des..

[3]  Ansi Ieee,et al.  IEEE Standard for Binary Floating Point Arithmetic , 1985 .

[4]  Nicholas J. Higham,et al.  The Accuracy of Floating Point Summation , 1993, SIAM J. Sci. Comput..

[5]  I. J. Anderson,et al.  A Distillation Algorithm for Floating-Point Summation , 1999, SIAM J. Sci. Comput..

[6]  Nicholas J. Higham,et al.  INVERSE PROBLEMS NEWSLETTER , 1991 .

[7]  Jon G. Rokne,et al.  Exact computation of the sign of a finite sum , 1999, Appl. Math. Comput..

[8]  Douglas M. Priest,et al.  Algorithms for arbitrary precision floating point arithmetic , 1991, [1991] Proceedings 10th IEEE Symposium on Computer Arithmetic.

[9]  William Kahan,et al.  Pracniques: further remarks on reducing truncation errors , 1965, CACM.

[10]  Marina L. Gavrilova,et al.  Reliable line segment intersection testing , 2000, Comput. Aided Des..