OpenXM/src/kan96xx/Doc/slope.sm1 - diff

Return to slope.sm1 CVS log

Up to [local] / OpenXM / src / kan96xx / Doc

version 1.2, 2001/06/18 03:12:21

version 1.4, 2012/06/11 05:23:52

Line 1

% $OpenXM: OpenXM/src/kan96xx/Doc/slope.sm1,v 1.1 2000/11/01 01:57:55 takayama Exp $

% $OpenXM: OpenXM/src/kan96xx/Doc/slope.sm1,v 1.3 2005/10/20 11:22:27 takayama Exp $

(oxasir.sm1.loaded) boundp not {

[(parse) (oxasir.sm1) pushfile] extension

} { } ifelse

Line 6

[(parse) (cohom.sm1) pushfile] extension

} { } ifelse

$slope.sm1, computing the slopes of a D-ideal: June 15, 2000$ message

$ (C) F.Castro-Jimenez, N.Takayama$ message

$ (C) N.Takayama, F.Castro-Jimenez$ message

$Imported commands: slope $ message

/slope.verbose 1 def

/gb.warning 0 def

Line 582 $Imported commands: slope $ message

(When slope.geometric is one, it outputs the geometric slopes.)

(As to the algorithm, see A.Assi, F.J.Castro-Jimenez and J.M.Granger)

( How to calculate the slopes of a D-module, Compositio Math, 104, 1-17, 1996)

(Note that the signs of the slopes are negative, but the absolute values)

(Note that the signs of the slopes s' are negative, but the absolute values -s')

(of the slopes are returned.)

$In other words, when pF+qV is the gap, -s'=q/p is returned.$

$Note that s=1-1/s' is called the slope in recent literatures. Solutions belongs to O(s).$

$The number s satisfies 1<= s.$

$We have r=s-1=-1/s', and kappa=1/r=-s',$

$which is used 1/Gamma(1+m*r) factor and exp(-tau^kappa)$

$in the Borel and Laplace transformations respectively.$

$Example 1: [ [(x^4 Dx + 3)] (x) [0 1] [-1 1]] slope :: $

$ The solution is exp(x^(-3)). $

$Example 2: [ [(x^3 Dx^2 + (x + x^2) Dx + 1)] [(x)] $