step-sizes. It starts out with large jumps, then as it cools the jumps get smaller and smaller as does the likelihood of a retrograde jump being accepted as a valid new state.
Most fitness landscapes have more than one dimension and therefore a 'jump' could include any number between 1 and N, where N is the dimensionality of the landscape. The Drift Rate setting —which may well be poorly named— controls the odds that a jump includes an additional dimension. All jumps must be at least one-dimensional, but 25 percent of them (on average) will include another dimension. 25% of those will include a third dimension and 25 percent of those a fourth and so on and so forth until the dimensionality of the landscape has been reached. Here's a list for 1000 jumps:
Drift Rate: 25%
1D jumps: 750
2D jumps: 187
3D jumps: 47
4D jumps: 12
5D jumps: 3
6D jumps: 1
A good question to ask would be; "Why would you want a jump to include more than one dimension?" and the answer is that the more genes are related, the higher the changes that a multi-dimensional jump will yield an improvement. It's not difficult to imagine that you cannot improve your current state by only modifying a single gene. Sometimes you need to change two in unison in order to reach a better solution. If your genes are highly related (which is bad practice to begin with) then you may need to adjust the Drift Rate to a higher value.
--
David Rutten
david@mcneel.com
Poprad, Slovakia…
Added by David Rutten at 11:09am on April 17, 2012
p://gamelab.mit.edu/games/a-slower-speed-of-light/
I suppose one could replace the current vector library so that whenever it calculated distances it took the metric tensor into account...
I don't think this is something I'll have time for any time soon though!
That said I do think there are applications for some specific 3-manifolds and mappings between them. I recently added to Kangaroo a way to simulate periodic boundary conditions so you can work in a 3-torus (allowing things like relaxing triply periodic minimal surfaces).
The 4d rotation script I posted can also be useful in conjunction with all the circle related geometry optimizations, such as conical meshes, circle packings, etc., because such transformations preserve circles and the angles between them.
…
anually update the definition for the changes to have an effect. Also, it's not very user friendly to change data using commands (you can add them to the toolbar).
I recently did a definition that required custom attribute data per object. Since objects could vary in number it was easier to have the data attached to the object rather than in a spreadsheet. I only needed 3 or 4 values per object, so i just added them to the name of the object in the form of "a;3;500". I have the object properties window open all the time so this way it's easy to quickly change the values.
It would be great if rhino's GUI allowed to add and change custom attribute data easily. Cinema 4D does this very well. I think David was working on a plugin that did this but i doubt it's still in development.…
Added by Vicente Soler at 4:10pm on October 12, 2009
component will draw a 'ghost' copy of the same path, but with one end at Z=0 and the other end at Z=1 (or 1000, depending on units / scale) so that each control point has a Z-value between those two extremes. These values should be adjustable, so that when I export the path, I take the XYZ coordinates of the control points of the original path, and interpret the Z-values of the 'ghost' path as time.
To do this, I need to be able to dynamically adjust the ghost path and be able to edit the control points of the ghost path (however constrained to movement in the Z-axis) so that the user can tweak the time-values.
Is this possible through C#? One way I thought of was to do it by passing the control point index values to a graph component to create the time values separately and then merge them when exporting, however I can't seem to get my head around the graph components in GH...
Any tips / help / suggestions would be much appreciated!
Tom…
he slopes.
Reference : A Cellular Texture Basis Function
Steven Worley
http://graphics.ucsd.edu/courses/cse168_s06/ucsd/cellular_noise.pdf
Cinema 4D textureshttp://cdn.tutsplus.com/cg/uploads/legacy/0000_Freebies/007_C4D_NoisePDF/C4D%20Noise%20Texture%20Reference%20v1.pdf
Laurent Delrieu August 2016l.delrieu@neuf.fr…