Research On Fluent Calculus And Its Application In Intelligent Agents

The design of intelligent agents is surely one of the most exciting and challenging goals in Artificial Intelligence. Intelligent agents are embedded in an incomplete, dynamic environment, and this requires the agents have higher cognitive capabilities such as reasoning about action, planning, drawing logical conclusions from sensor information acquired, and reacting sensibly to unexpected accidents.As a development of Situation Calculus, Fluent Calculus adds the notion of state which can represent the incomplete environment, and solves the representational and inferential aspect of the classical frame problem. FLUX as a logic programming language for intelligent agents is based on the theory of Fluent Calculus. Fluent Calculus and FLUX provide formalism and realization methods for reasoning about action of the agents for achieving the goal in incomplete environments.The major achievements in this paper are listed here:(1) Systematical research of the formal theory of Fluent Calculus is given in this paper. Fluent Calculus and FLUX can formalize the dynamic environment, and give intelligent agents cognitive ability.(2) The problems such as troubleshooting and planning are discussed in the framework of Fluent Calculus, which are applied in the instance by using FLUX.(3) The incomplete office environment, the preconditions of actions, the effects of actions and knowledge that needed for the office intelligent agent to reason are formalized in this paper by using Fluent Calculus. With the help of idea of constraint program and FLUX, the encoding and realization for the formalism description are completed, and the action sequences are reasoned for the agent to complete the goal, therefore, Fluent Calculus applied in the intelligent agents of the office scenario is realized.(4) The external world the agent inhabits and the inner reasoning are intergrated by using the Java-ECLiPSe Prolog interface, and the simulation of reasoning about action of intelligent agents in office scenario is experimented.The simulation experiment indicates that the agent can automatically reason about action by themselves for any given goal based on Fluent Calculus and FLUX. The research in this paper does some tries on the integrations between the theory of Fluent Calculus and its practical applications beneficially and explores a feasible path.