### Development and Application of Logical Actors Mathematical Apparatus for Logic Programming of Web Agents [chapter]

Alexei A. Morozov
2003 Lecture Notes in Computer Science
One of the most interesting and promising approaches to programming Internet agents is logic programming of agents. This approach has good prospects, because the ideology and principles of logic programming are very convenient for searching, recognition, and analysis of unstructured, poorly structured, and hypertext information. Many ideas and methods of logic programming of Internet agents based on various modifications of Prolog and non-classical logic (linear, modal, etc.) were developed
more » ... ng the recent decade. Nevertheless, there has been no mathematical apparatus providing sound and complete operation of logic programs in the dynamic Internet environment (i.e., under conditions of permanent update and revision of information). To solve this problem, we have created a mathematical apparatus based on the principle of repeated proving of sub-goals (so-called logical actors). Our mathematical apparatus for logic programming of Internet agents includes: 1. A model of intelligent agents that operate in a dynamical environment; 2. A classical declarative (model-theoretic) semantics of agents; 3. Control strategies for executing logic programs (Internet agents) that are sound and (under some conditions) complete with respect to the modeltheoretic semantics of these agents. Within the framework of our model of intelligent agents, an Internet agent (a group of Internet agents) is a logic program controlled by a special strategy. The control strategy is a modification of standard control strategy of Prolog, enhanced by so-called repeated proving of sub-goals. The idea of repeated proving consists in dividing the program into separate sub-goals (called logical actors) [2, 3] that have the following properties: 1. Common variables are the single channel of data exchange between the actors. 2. Proving of separate actors can be fulfilled independently in arbitrary order. 3. One can defeat the results of proving of any actor while keeping all other sub-goals of the program.