Jump to content

  • Log In with Google      Sign In   
  • Create Account

Tasaq

Member Since 05 Nov 2011
Offline Last Active May 24 2016 02:08 PM

Posts I've Made

In Topic: Objects distribution within concave polygon

26 August 2015 - 07:41 AM

Finally I found something satisfying, Particle Swarm Optimisation was good for my problem. I used the same fitness function as I used for GA but PSO was much faster and gave a little bit better solutions.


In Topic: Objects distribution within concave polygon

19 August 2015 - 03:13 AM

If you use an optimization method like simulated annealing or genetic algorithms, you can divide the problem in two parts:
1) Generate candidate distributions (usually by modifying an existing distribution a bit)
2) Evaluate their utility (how happy you are with the result)

If you are not happy with what the results look like, try to define what you don't like about them very precisely, then add that as an extra term in the utility function.

The candidate generation can be used to speed up the process dramatically. For instance, if having objects on corners or on walls is rewarded by the utility function, you should try to place objects there with high probability.

 

This sounds like a nice idea. So do you think something like combining GA with gradien descent would be a good way to do it?

I also found something like this:

https://www.youtube.com/watch?v=HQ41szO1AMc

 

The first part of the video (0:00 - 0:25) shows something I need. The points are distributed in optimum way and can be easily aligned afterwards (even manually). I have no idea what was used here, it looks like some kind of coverage optimization using independent inteligent agents?


In Topic: Terrain LOD tessellation cracks

02 June 2014 - 02:03 PM

Are you using hardware tessellation? If so, can you share the code?


In Topic: Are you a gamer?

04 April 2014 - 01:50 PM

I play some games. Or should I say, re-play, since I prefer games from late 90's smile.png From modern games I play Smite and some games I find cool (I am very demanding gamer). But I must say I mostly spent my time coding than playing. Sometimes when friends show me a gameplay of new game I am like "WOW, they got nice shading". Same thing happens when I play, I am like "wow, what a nice material on that trash can..." or "this shadow artifacts are strange, wonder what's causing them" and at the same time I die or crash or whatever biggrin.png

 

Oh, and I once heard something like that: "I am not playing games, I am checking what kind of technology competition has" :)


In Topic: Real-time Ray Tracer in GLSL

02 April 2014 - 02:18 PM

I think you did it in a similar way as I did mine.

 

I don't see part for triangle intersection? Does that mean the walls in your video are just enormous speheres? smile.png

 

More reasonable approach for ray tracing in real-time is to use either OpenCL or CUDA, they are much more flexible than fragment shaders. As you experienced you got 10 fps for scene with 10(?) spheres and 2 lights. I don't want to discourage you in any way, but imagine you want to render scene with, for instance, 5000 triangles smile.png For that you are bound to use acceleration structures, and they might be tough to implement with GPGPU.

 

If you want to follow the path of writing real-time ray tracers, you can try implementing geometry shader ray tracing:

http://ar.in.tum.de/pub/ralovich2010rsigs/ralovich2010rsigs.pdf

 

Note that this technique is slower than fragment shader method (ready article's results section for details).


PARTNERS