A fuzzy utility-based agent for the design problem

Abstract

This article conceptualizes, formalizes, and tests a utility-based agent with multiple goals capable of designing sequences of actions, in an environment where perceptions regarding the present and desired situations of the agent, its actions, and information that it maintains internally with respect to the effects of these actions are ambiguous and vague concepts. Fuzzy sets, the compositional rule of inference, several theoretical results involving fuzzy relational equations solutions, a measurement for fuzzy sets similarity, problem-solving and decision making theories and others results from heuristic search were used in the outline of the agent and the formalization of its actions selection mechanism.