Generalized LR parsing and the shuffle operator

TL;DR

This paper extends Tomita's Generalized LR parsing algorithm to handle context-free grammars with a shuffle operator, enabling more complex language recognition, motivated by AI plan recognition applications.

Contribution

It introduces modifications to the handle-finding automaton and parser table construction for grammars with shuffle, advancing parsing techniques for enriched languages.

Findings

Successfully adapted the algorithm for shuffle operators

Ensured correctness of the extended parser

Discussed potential future improvements

Abstract

We adapt Tomita's Generalized LR algorithm to languages generated by context-free grammars enriched with a shuffle operator. The change involves extensions to the underlying handle-finding finite automaton, construction of parser tables, and the necessary optimizations in constructing a deterministic parser. Our system is motivated by an application from artificial intelligence plan recognition. We argue for the correctness of the system, and discuss future extensions of this work.