Re: [Repast-developer] [Repast-interest] ValueLayerDiffuser in Repast HPC?

[Murphy, J. T. <jtm...@an...>]

Nathan,

Thanks very much for all of these- this will be extremely helpful.

Best,
John 
-- 
John T. Murphy
Anthropologist/Social Science Modeling Specialist
Computational Engineer
Global Security Sciences Division
Systems Science Center, Social and Network Analysis Section
Argonne National Laboratory
jtm...@an...




On 4/1/15 11:02 AM, "Nathan Liles" <nm...@gm...> wrote:

>Hi John,
>
>Attached is version 1 of the ValueLayerDiffuser class. It appears to be
>working in our model, but I haven't gotten around to writing tests for
>it. So, caveat emptor.
>
>While I was working on the class, I found a couple of issues:
>
>1. In RepastErrors.h, Repast_Error_47 has a typo in the REASON stanza
>that causes creation of DiscreteValueLayer objects to fail due to its
>reliance on matrix.h (which is responsible for throwing
>Repast_Error_47). The fix is to wrap matrixDimensions in a VAL().
>
>2. The DiscreteValueLayer constructor is missing an initializer for the
>WrapAroundBorders object
>
>3. Also in the DiscreteValueLayer constructor, the matrix objects it
>creates expect the dimensions to be Point<int>, but GridDimensions only
>accepts Point<double>. I modified it to convert the coordinates
>accordingly.
>
>4. ValueLayer::shape() returns Point<int>. But, once again,
>GridDimensions can only be created using Point<double>. So calling
>shape() will throw an error. There technically is a way around this by
>calling ValueLayer::dimensions()::extents(). This doesn't affect the
>ValueLayerDiffuser, but I thought I'd fix it while I was in there.
>
>I couldn't find a bug tracker anywhere online. If there is one, I'd be
>happy to add these to it. I also attached patches for RepastErrors.h and
>ValueLayer.h. They are necessary to get ValueLayerDiffuser up and running.
>
>Let me know if there's anything I missed,
>-Nathan
>
>
>Murphy, John T. wrote:
>> Hi Nathan,
>>
>> There is not a diffuser class in Repast HPC, though we have had several
>> requests for something like it.
>>
>> If you generate some code for this, please share it; it may be something
>> we could incorporate into the Repast HPC tools.
>>
>> Best,
>> John
>>