An Agent Formal Model for Autonomous Decision-Making

Abstract

With the development of artificial intelligence, agent is widely applied to the intelligent system. As the intelligent decision at a higher sense, autonomous decision-making directly affects the intelligent level of intelligent system. In order to improve autonomous decision-making ability of intelligent system, the paper aims to use formal method to research on the process of autonomous decision-making. Combining the feature of autonomous decision-making and agent formal method, the paper puts forward an agent model for autonomous decision-making, namely-ADAM(Autonomous Decision-making Agent Model). The model mainly includes five parts: data pretreatment, decision-making response, decision-making execution, decision-making evaluation and learning feedback. Then the paper gives the formal description of ADAM, which realizes formal analysis for the process of autonomous decision-making. Finally, the paper points out the realized critical point of the model and preliminary demonstrates validity of the model.