       FindRoot failing in v8 on a system easily solvable in v7

• To: mathgroup at smc.vnet.net
• Subject: [mg114743] FindRoot failing in v8 on a system easily solvable in v7
• From: vasil michev <michev at gmail.com>
• Date: Wed, 15 Dec 2010 07:50:58 -0500 (EST)

Consider the following input

eq1:=Block[{u1,u2,u3,u4},u1:=uX-2*NIntegrate[Cos[x]/(Exp[1/1.529266/
T*(uX*0.004559*(1-Cos[x])+uY*0.004559*(1-Cos[y])+uZ*0.012095*(1-
Cos[z]))]-1),{x,-\[Pi],\[Pi]},{y,-\[Pi],\[Pi]},{z,-\[Pi],\
[Pi]},PrecisionGoal->4, MaxPoints->15000000,Method ->
u2:=uY-2*NIntegrate[Cos[y]/(Exp[1/1.529266/T*(uX*0.004559*(1-Cos[x])
+uY*0.004559*(1-Cos[y])+uZ*0.012095*(1-Cos[z]))]-1),{x,-\[Pi],\[Pi]},
{y,-\[Pi],\[Pi]},{z,-\[Pi],\[Pi]},PrecisionGoal->4, MaxPoints-
u3:=uZ-2*NIntegrate[Cos[z]/(Exp[1/1.529266/T*(uX*0.004559*(1-Cos[x])
+uY*0.004559*(1-Cos[y])+uZ*0.012095*(1-Cos[z]))]-1),{x,-\[Pi],\[Pi]},
{y,-\[Pi],\[Pi]},{z,-\[Pi],\[Pi]},PrecisionGoal->4, MaxPoints-
u4:=1.529266-NIntegrate[1/(Exp[1/1.529266/T*(uX*0.004559*(1-Cos[x])
+uY*0.004559*(1-Cos[y])+uZ*0.012095*(1-Cos[z]))]-1),{x,-\[Pi],\[Pi]},
{y,-\[Pi],\[Pi]},{z,-\[Pi],\[Pi]},PrecisionGoal->4, MaxPoints-
FindRoot[{u1,u2,u3,u4},{{uX,2,2.2},{uY,.5,.8},{uZ,1.5,2.5},{T,
0.01,0.1}},AccuracyGoal->5,PrecisionGoal->5]//Quiet]

In v7 I get

{uX->1.08147,uY->1.08147,uZ->2.08043,T->0.023582}

while in v8 I get NIntegrate::inumr errors. Help browser shows no
changes to FindRoot in v8 and I cant figure out whats failing. A quick
test shows that the equations dont have any undefined parameters

In:= Block[{uX=1,uY=1,uZ=1,T=1},u1]
Out= -69.385

so it must be FindRoot thats causing the problem. Any ideas?

• Prev by Date: Maximum optimization problem
• Next by Date: Re: How to create a notebook outside of Mathematica?
• Previous by thread: Maximum optimization problem
• Next by thread: Re: FindRoot failing in v8 on a system easily solvable in v7