The USA-Advisor: A Case Study in Answer Set Programming

Abstract

Answer set programming [8] is a new declarative programming paradigm suitable for solving a large range of problems related to knowledge representation and search. The paradigm is rooted in recent developments in several areas of artificial intelligence. Answer set programming starts by encoding relevant domain knowledge as a (possibly disjunctive) logic program, Π. The connectives of this program are normally understood in accordance with the answer set (stable model) semantics [4], [5]. The language’s ability to express defaults, i.e. statements of the form “normally, objects of class C have property P”, coupled with its natural treatment of recursion, and other useful features, often leads to a comparatively concise and clear representation of knowledge. Insights on the nature of causality and its relationship with the answer sets of logic programs [6], [7],[10] allows description of the effects of actions which solves the frame, ramification, and qualification problems, which for a long time have caused difficulties in modeling knowledge about dynamic domains.

In the second stage of the programming process, a programming task is reduced to finding the answer sets of a logic program Π ∪ R where R is normally a simple and short program corresponding to this task. The answer sets are found with the help of programming systems [9],[2], [3] implementing various answer set finding algorithms.

During the last few years the answer set programming paradigm seems to have crossed the boundaries of artificial intelligence and has started to attract people in various areas of computer science.

In this talk I will briefly describe the basic idea of the approach and outline its use for the development of the USA-Advisor decision support system for the Space Shuttle. The largest part of this work was done by my former and current students Monica Nogueira, Marcello Balduccini, and Dr. Richard Watson, in close cooperation with Dr. Matt Barry from the USA Advanced Technology Development Group [1]. Our goals in creating the USA-Advisor were two-fold. From a scientific standpoint we wanted to test if the rapidly developing answer set programming methodologies, algorithms, and systems could be successfully applied to the creation of medium size, knowledge intensive applications. From the standpoint of engineering, the goal was to design a system to help flight controllers plan for correct operation of the shuttle in situations where multiple failures have occurred. Even though the engineering part of the project is not yet fully completed it is clear that the approach proved to be successful. In this talk I’ll share some lessons and observations learned from this work.