An Action Planning Module Based on Vague Knowledge Extracted from Past Experiences

摘要

In this paper, an agent architecture is presented and a formal model for a preference choice module and planning module based on vague knowledge is described. An agent situated in an environment gathers information about objects placed around it and stores this information inside it's own database. According to it's preferences, an agent plans it's actions in order to change surrounding environment into it's preferences. Because of a partial lack of knowledge about an environment, an agent chooses a plan with a highest success ratio due to it's previous observations.