Discovery Learning in Autonomous Agents Using Genetic Algorithms
Abstract
As the new Distributed Interactive Simulation (DIS) draft standard evolves into a useful document and distributed simulations begin to emerge that implement parts of the standard, there is renewed interest in available methods to effectively control autonomous aircraft agents in such a simulated environment. This investigation examines the use of a genetics-based classifier system for agent control. These are robust learning systems that use the adaptive search mechanisms of genetic algorithms to guide the learning system in forming new concepts (decision rules) about its environment. By allowing the rule base to evolve, it adapts agent behavior to environmental changes. Addressed are the learning needs of autonomous aircraft agents, showing how multiple learning strategies are possible and that the best approach is a coherent combination of these. A design is described for a control system using a distributed filtering architecture and a genetics-based classifier system modified to support a phasing-rule niching system based on phase tags. Finally, a prototype system called the Phased Pilot Learning System (PPLS) is implemented based on this design and tested within a limited simulation environment. Results from empirical tests show that this approach is a viable alternative to other control methods. Discovery learning, Genetic algorithms, Autonomous agents.
Document Details
- Document Type
- Technical Report
- Publication Date
- Dec 01, 1993
- Accession Number
- ADA274083
Entities
People
- Edward O. Gordon
Organizations
- Air Force Institute of Technology