Heya. Figured this would interest some in this forum (especially those interested in genetic algs/ANNs area). My dissertation project is on Automatic Game Content Generation. Specifically, an algorithm that automatically evolves unique particle system weapons based on those players preferred in the past. Link to game download and pics/info: gar.eecs.ucf.edu It is a multi-player space shooter with unique evolving weapons. The main goal is to show that, unlike games where there are maybe 8-10 developer designed weapons, it is feasible to automatically evolve unique weapons (or other content) in a game based on player preferences. The weapons evolve in the following manner: - CPPNs (a type of neural network) represents and controls particle system weapons - each weapon has a unique CPPN - each particle fired by the weapon, is controlled by the weapon CPPN - when players destroy enemies, new weapons are produced - the CPPNs for the new weapons are evolved from those weapons that players favored in the past Anyway, let me know what you think. Especially if there are any problems with the installer or game itself. So far it seems that the algorithm really works. A lot of cool weapons have been produced. BTW, playing on the GAR Official Server 1 will help me collect dissertation data. =) Thanks for reading. Any feedback would be greatly appreciated. Also, any thoughts on the future of evolving game content (via neuroevolution or any other method)?