Particle Swarm Optimization Mario AI (Second Draft)

Second video taken from the particle swarm optimization Mario agent. This model demonstrates the optimization method searching a high dimensional space. The overview of how it works is for each 3 dimensions that is one single action to take. The dimensionality is a multiple of 3 which means a single particle is a candidate solution to a series of actions. Each successive one assumes the previous action was taken.

This project is the main focus of my undergraduate computer science thesis. Hope you enjoy!

More info: www.adamstirtan.com

Category:

Science & Technology

Tags:

• pso
• mario
• particle swarm
• particle swarm optimization
• ai
• artificial intelligence
• a.i.

[adamstirtan]

I'm going to have to disagree there. PSO is primarily focused on learning through adaptation to the environment. In reference to it being by "random" chance you instead meant stochastic, in which you would be partially right.

[cakefarts]

this shit doesnt actually learn how to play though, fuckers run through until one of them gets it right by random chance

[Patate7Frite]

cool

[gransoporo]

what is your fitness here!? I can't understand

[andtenorio]

man, this is pretty cool!

[mehwoot]

well... it's getting better