09/01/1994 09:00 AM
Computer Science
This paper is a blueprint for the development of a fully domain-independent single agent and multi-agent heuristic search system. The paper gives a graph-theoretic representation of search problems based on conceptual graphs, and outlines two different learning systems. One, an ``informed learner\" makes use of the the graph- theoretic definition of a search problem or game in playing and adapting to a game in the given environment. The other a ``blind learner\" is not given access to the rules of a domain, but must discover and then exploit the underlying mathematical structure of a given domain. These learning agents are based on generalizing the understanding obtained with the Morph chess system to all games involving the interactions of abstract mathematical relations. An effective monitor for such domains is presented, along with implementations of blind and informed learners. Performance results are very encouraging.
