140 likes | 248 Views
Evolving the goal priorities of autonomous agents. Adam Campbell Advisor: Dr. Annie S. Wu School of Electrical Engineering and Computer Science. Outline. Motivation Genetic algorithms Agent control system Experiments and results Conclusion and future work. Motivation.
E N D
Evolving the goal priorities of autonomous agents Adam Campbell Advisor: Dr. Annie S. Wu School of Electrical Engineering and Computer Science
Outline • Motivation • Genetic algorithms • Agent control system • Experiments and results • Conclusion and future work
Motivation • Imagine an insect with two goals • Get food • Avoid predator • When should the get food goal priority be higher than the avoid predator goal? • Action selection How should agents choose between the actions of conflicting goals? • Two general methods • Take one action • Combine actions • Use a genetic algorithm to learn goal weights
Genetic algorithm • Survival of fittest amongst problem solutions • General algorithm… 1) Initialize random population 2) Evaluate population 3) Select individuals 4) Recombine/mutate selected individuals 5) If stopping condition not satisfied 6) GOTO 2
Fitness 4 3 3 4 5 2 Genetic algorithm example • Problem: find all black squares Random population
Crossover & Mutation Fitness 5 4 2 6 4 5 GA example continued Selected population
How is the GA used? • Immediate goal functions • Produce a vector indicating where the agent should move in order to best satisfy the goal • Each immediate goal has a weight associated to it • Five immediate goal functions • Avoid agent • Avoid obstacle • Momentum • Go to area of interest (AOI) • Follow obstacle
Additional parameters 0.00 0.01 0.04 • Randomness • Comfort • Allows obstacle followingto occur
Average fitness • Agents must survive and see as many AOIs as possible • Not much difference in fitness between two scenarios
Summary and conclusion • Discussed action selection problem in artificial intelligence and showed an evolutionary approach to solving • Tested approach on simple problem scenarios • Performed well on both scenarios • New behaviors (goals) can easily be added to the system • The parameters evolved are specific to the environment they were learned in
Future work • Social interactions between agents • Allow communication of data between agents • New immediate goal functions needed • Allow agents to have more than one set of goal weights • Depending on the agent’s state (hungry, low on fuel, in danger, etc.) use a different set of goal weights • Other ways to combine vectors from immediate goal functions • Non-linear combination of vectors • Genetic programming • Currently being worked on at George Mason University • Better test scenarios • Evolve parameters that generalize well to unseen environments