Presentation on theme: "Coevolution of Human-Competitive Robocode Tanks Using Genetic Programming with Exogenous Fitness Jason Owens & Ron Bowers."— Presentation transcript:
Coevolution of Human-Competitive Robocode Tanks Using Genetic Programming with Exogenous Fitness Jason Owens & Ron Bowers
Why? Possible relevance to our day jobs.
Difficulties with Coevolution
Previous Work with Robocode Eisenstein  Used a GA to evolve a subsumption architecture. Was successful in developing bots that could fight a specific adversary given a specific starting condition Attempted to use coevolution but after several generations I found the populations rife with catatonics Hong and Cho  Used a GA that consisted of 6 chromosomes, representing the behavior in the main loop and in 5 of the event handlers. Each chromosome consisted of six genes, corresponding to actions such as move or shoot. Each action could be one of 2 or more hand-coded implementations. Were successful in consistently defeating 3 of the standard bots.
Previous Work With Robocode Shichel, Ziserman, and Sipper  Used Koza-style GP Limited investigation to Haiku Bot (4 lines of code) Evolved bots were entered into a Haiku Bot tournament where they placed third out of 27.
Hypothesis We hypothesize that by using genetic programming and coevolution with an exogenous fitness function we can evolve Robocode agents that can compete successfully against human-coded bots. But...
Initial Results We did not entirely succeed. We did not produce any competitive agents in time to report in the paper. We have continued our efforts!
Algorithm Configuration Strongly-typed tree-based GP Linear-rank selection using stochastic universal sampling 80% crossover, 20% mutation (whoa!) Elitism (one individual) Initial tree depth of 5
Fitness Functions simple (shoot and dodge) movement, enemy sensing, wall avoidance re-proportioned movement value final, emphasis on damage with firing/scanning efficiencies