Re: Experimental`NumericalFunction::dimsl
- To: mathgroup at smc.vnet.net
- Subject: [mg109388] Re: Experimental`NumericalFunction::dimsl
- From: DrMajorBob <btreat1 at austin.rr.com>
- Date: Sat, 24 Apr 2010 04:03:17 -0400 (EDT)
You are correct that we're rarely willing to wade through a lot of code,
and you're correct that it's not always necessary.
But frequently, it may be a lot of code, but it doesn't matter; I just
evaluate it, I get to the point where your error appears, and then I can
start sleuthing... without having to understand all your code.
And sometimes I _do_ wade through and understand all the code (out of
curiosity), and if so, I usually improve it. Not in your case,
necessarily... but often.
So... use your own judgment, of course.
Bobby
On Fri, 23 Apr 2010 21:57:46 -0500, Joe Hays <hays.joe at gmail.com> wrote:
> Fair enough. I had considered including my code but chose not to so I
> didn't
> dump a blob of code for you all to wade through. I was hoping that
> someone
> had at least seen the mystery error and had a general suggestion.
>
> Here is what I've learned in the past day. The mystery error appears to
> be
> starting point related. In all scenarios that I had experienced the
> error I
> was letting FindMinimum select the starting point based on my
> constraints.
> For whatever reason, the starting point that FindMinimum was selecting
> for
> my problem planted the seed for this unknown error that rears its head
> late
> in the execution.
>
> When I specified my own starting point(s) I stopped experiencing the
> error.
> I have no explanation. My problem is mildly nonconvex but continuous so I
> expected FindMinimum to settle on a local minimum (not to die on an
> error).
> Anyway, as of right now I appear to be up and running.
>
> So, if anyone gets this problem in the future my suggestion is to
> investigate specifying your own starting point(s). If that doesn't help
> then
> I'm not sure what to suggest.
>
> Thanks for the response DrMajorBob. I'd be happy to share my code if
> you're
> curious but as I said I left it out so the group didn't have to wade
> through
> it.
>
> Joe
>
> On Fri, Apr 23, 2010 at 11:40 AM, DrMajorBob <btreat1 at austin.rr.com>
> wrote:
>
>> You gave us NOTHING to work with. How can we guess the problem, without
>> seeing the code?
>>
>> Bobby
>>
>>
>> On Fri, 23 Apr 2010 02:50:07 -0500, Joe Hays <hays.joe at gmail.com> wrote:
>>
>> Hello all,
>>>
>>> I'm stumped with an error code I'm getting from FindMinimum after it
>>> runs
>>> for over 7 minutes. Its,
>>>
>>> <snip>
>>> Experimental`NumericalFunction::dimsl: {-0.8} given in -0.8 should be a
>>> list
>>> of dimensions for a particular argument.
>>>
>>> FindMinimum::dimsl: {-0.8} given in -0.7999999999999636` should be a
>>> list
>>> of
>>> dimensions for a particular argument.
>>> </snip>
>>>
>>> This is all I'm given. There is no documentation on this error and it
>>> seems
>>> odd that I'd get it so long after the function has already been
>>> executing.
>>> So, I'm completely stumped...
>>>
>>> Any thoughts/advice would be greatly appreciated.
>>>
>>> Joe
>>>
>>>
>>>
>>
>> --
>> DrMajorBob at yahoo.com
>>
--
DrMajorBob at yahoo.com