Integrate gives wrong answer

*To*: mathgroup at smc.vnet.net*Subject*: [mg57052] Integrate gives wrong answer*From*: "Brian Rogers" <brifry at gmail.com>*Date*: Thu, 12 May 2005 22:44:45 -0400 (EDT)*Sender*: owner-wri-mathgroup at wolfram.com

Hi, I have a problem on which Integrate misses badly. The code consits of: fmin[x_, p_]:=[p(1-(x-a+e)/(2e))^(p-1)]/2e which is the pdf of the minimum of p draws from a uniform distribution on [a-e,a+e], and Integrate[Abs[Min[mi,mc]-Min[mj,mc]]*fmin[mc,k]*fmin[mi,ni-k]*fmin[mj,nj-k], {mc,a-e,a+e},{mi,a-e,a+e},{mj,a-e,a+e},Assumptions->{k\el\_Integers, ni\el\_Integers,nj\el\_Integers,1<=k<=Min[ni,nj],0<a-e<a+e<1}] This produces the output: \!\(k\ \((\(a + e\)\/ni - \(2\ e\)\/\(1 + ni\) + \(a + e\)\/\(k - ni - nj\) + \(2\ \ e\)\/\(1 - k + ni + nj\))\)\) which is not symmetric betwenn ni and nj, as it should be. Any ideas? Thanks! Brian Rogers brifry at gmail.com