Genetic algorithm to solve simple levels in Guarding Nag – Video

Posted: November 7, 2012 at 6:46 pm




Genetic algorithm to solve simple levels in Guarding Nag
This video shows a genetic algorithm used to solve a simple level in Guarding Nag. It displays the evolution of Nag (blue guy) and Stan (orange guy) from being totally helpless to being able to reach the goal. This algorithm uses 20 genomes per generation and shows the best genome of this generation. In the evolution process the four best genomes are taken into the next generation. The other 16 genomes are created by multipoint-crossover of two parents who are selected by roulette-wheel-selection. After the recombination the offspring are mutated. The fitness is based on the distance of Nag to the goal and the distance of Stan to Nag over all update steps; the lower the better. Each genome may live up to 1800 update steps. After these a genome dies to prevent it from living forever.From:MuDGamesViews:0 0ratingsTime:01:37More inGaming

Excerpt from:
Genetic algorithm to solve simple levels in Guarding Nag - Video

Related Posts