Uploaded by fcampelofp on Oct 29, 2009
This short video illustrates the artificial evolution of mimetism in a population of lizards. Each (RGB) component of the lizards' color was represented as an 8-bit "gene" (haploid), which resulted in each lizard being represented by its 24-bit "genome". In the video we can see four lizards sampled from the population, and the decrease of the population's average "visibility" against the background. As a measure of this visibility the distance between the RGB vectors of the lizards and the average color of the background was used. The evolutionary process was modeled as a very, very simple genetic algorithm: 1) An initial population of 200 lizards (with random colors) was assumed to exist; 2) This population was subject to mortality: 30% of the population died randomly (e.g., due to disease, starvation, accidents, etc.) and, starting at generation 25, the 20% worst (less-fit) also perished (e.g., introduction of a new predator in the environment) 3) The surviving lizards were allowed to mate (with randomly selected partners), generating a number of offspring equal to the initial size of the population, and then died off; The offspring are generated through genetic recombination of their parents' "genomes", and are subject to mutations (the probability of a mutation happening at a given point was set as 1 in 1000) 4) The cycle then continues from step 2. After a certain time the background pattern changes (simulating a sudden environmental change). It can be seen that, as the selective pressure starts pulling in a different direction, the average color of the population tends to follow after a few generations. This little simulation is intended only to illustrate the ability of very simple evolutionary systems to produce a population with good mimetism characteristics. In the next ones I intend to include characteristics such as: - diploidy and dominance; - gender; - generation gaps (lizards living and reproducing across multiple generations) Suggestions and comments are welcome.
-
1 likes, 0 dislikes
0:46
Evolution of mimetismby fcampelofp144 views
1:19
Genetic Algorithms: Mario level 1-1 Completely Random Jumping.by evolsoulx24,272 views
0:25
Super Mario AI trained by Genetic Algorithm. Playing a random level.by tws2205847,898 views
1:00
ann_452.aviby Alkanen110 views
2:57
Auto adaptable walking robot with genetic algorithmby LUKASFREISTYLE18,164 views
0:33
ZombieSim 2.0 - 500 generationsby neosprockets790 views
1:13
NXT Self-learning SlotCar Autopilotby AlbU2B1,542 views
0:42
Algorithmic Architecture - Pareto Optimalby dmmd123565 views
0:31
Super Mario Brosby Protarget278 views
0:21
Mario AI trained by Genetic Algorithmby tws2205845,015 views
1:55
GA Crawler (learning how to walk)by teknowledgeable4,932 views
1:10
Genetic Algorithm Fishby ravivyatom54,027 views
9:18
Simulated Evolution of a Simple Car Using a Genetic Algorithmby AlexPleaseHelpDotCom3,096 views
0:27
Super Mario AI trained with Genetic Algorithm. Playing on trained levelby tws220584454 views
9:44
Evolving Wind Turbine Bladesby sjh71329,930 views
1:50
Creatures Population Simulatorby hsauro207 views
3:35
6 Degrees of Freedom Flight Simulator Platformby RunRabbitRunRunRun35,237 views
1:39
Genetic Algorithmby KipRicker1,355 views
3:42
Artificial Evolution in Robotic Islandsby mpeniak920 views
2:28
Tetris AI - Genetic Programming Vs Tetris Gameby HeDo88TH14,607 views
- Loading more suggestions...
Link to this comment:
All Comments (0)