[Date Index]
[Thread Index]
[Author Index]
Re: FindRoot and parameters in NIntegrate
*To*: mathgroup at smc.vnet.net
*Subject*: [mg124508] Re: FindRoot and parameters in NIntegrate
*From*: Bill Rowe <readnews at sbcglobal.net>
*Date*: Sun, 22 Jan 2012 07:21:53 -0500 (EST)
*Delivered-to*: l-mathgroup@mail-archive0.wolfram.com
On 1/21/12 at 5:18 AM, g.crlsn at gmail.com (gac) wrote:
>I am using Mathematica 8 to solve for parameters in an integral and
>an auxiliary equation. A prototypical problem is
>FindRoot[{1 + a + b , NIntegrate[a + b x, {x, 0, 1}]}, {{a, 0}, {b,
>0}}]
>Mathematica reports the correct answer for the parameters a and b:
>{a -> 1., b -> -2.}
>But it also reports two error messages:
There are several ways around this. Best would be:
In[13]:= int = Integrate[a + b x, {x, 0, 1}];
FindRoot[{1 + a + b, int}, {{a, 0}, {b, 0}}]
Out[14]= {a -> 1., b -> -2.}
Not only will this eliminate the error messages, but the
integral is only computed once. But, this does require that the
integral be one Mathematica can solve symbolically. If that is
not the case you could do:
In[15]:= f[a_?NumericQ, b_?NumericQ] :=
Integrate[a + b x, {x, 0, 1}];
FindRoot[{1 + a + b, f[a, b]}, {{a, 0}, {b, 0}}]
Out[16]= {a->1.,b->-2.}
The error message results since NIntegrate tries to solve the
integral before FindRoot supplies numeric values for a and b.
This solves the problem since f[a,b] will remain unevaluated
until a and b are given numeric values. But while this works as
intended, it is much less efficient than the first approach
since the integral will be repeatedly solved for each time
FindRoot provides numeric values for a and b. This approach
should only be used in those cases Mathematica cannot provide a
symbolic solution to the integral.
Finally, you could do:
In[17]:= Quiet@
FindRoot[{1 + a + b,
NIntegrate[a + b x, {x, 0, 1}]}, {{a, 0}, {b, 0}}]
Out[17]= {a->1.,b->-2.}
Which simply suppresses the error messages. This is suitable
whenever you know the error messages are unimportant as is the
case here.
Prev by Date:
**Re: log-like symbols?**
Next by Date:
**Re: Derivatives Output as TraditionalForm**
Previous by thread:
**Re: FindRoot and parameters in NIntegrate**
Next by thread:
**Re: FindRoot and parameters in NIntegrate**
| |