Integration problem / internal error ?
- To: mathgroup at smc.vnet.net
- Subject: [mg31328] Integration problem / internal error ?
- From: Madhusudan Singh <ar at sc.hc.de>
- Date: Sun, 28 Oct 2001 03:07:08 -0500 (EST)
- Organization: University of Michigan Electrical Engineering and Computer Science
- Sender: owner-wri-mathgroup at wolfram.com
The problem is best stated as the following : I have a monotonically decreasing function f[x+y] that is used in the evaluation of a triple integral. The values of the function at the endpoints are estimated to be 10^(-31) and 0.5 (taking max(x+y) and min(x+y) over the integration interval). Obviously the invocations of the function during the Monte Carlo integration should lead to values intermediate between these. However, an attempt to integrate leads to errors like "10^(-680) should be a machine sized real number". When f[z] is replaced with 1, or Chop[f[z],10^(-50)], the errors go away. I am puzzled. How can this possibly happen ? Is it an internal Mathematica error ?