A ava ased Implementation of nowledge and ecision Making in Environments with u y ariables

nowledge of intelligent agents can be described at three different levels, the knowledge level or the epistemological level, the logical level, and the implementation level. The implementation level is the most important level for an AI agent to act efficiently. This paper describes a ava-based implementation of an agent´s knowledge in an environment characterized by its fuzzy variables and a ava-based implementation of decision trees to facilitate decision making in such environments. An agent´s knowledge in this paper is defined in the form of fuzzy sets, linguistic variables and fuzzy if-then rules. Fuzzy sets and linguistic variables are constructed as objects of ava classes. Decision trees are constructed from fuzzy if-then rules. Each tree can yield one decision or more. The decision is then chosen based on a utility function that uses a min-max approach.