Programming Goal-driven Web Sites using an Agent Logic Language

In this paper we show how an agent programming language, based on a formal theory of actions, can be employed to implement adaptive web applications, where a personalized dynamical site generation is guided by the user’s needs. For this purpose, we have developed an on-line computer seller in DyLOG, a modal logic programming language which allows one to specify agents acting, interacting, and planning in dynamic environments. Adaptation at the navigation level is realized by dynamically building a presentation plan for solving the problem to assemble a computer, being driven by goals generated by interacting with the user. The planning capabilities of DyLOG are exploited to implement the automated generation of a presentation plan to achieve the goals. The DyLOG agent is the “reasoning” component of a larger system, called WLog, which is described in this paper.