[Paper Review] Polygraphs for termination of left-linear term rewriting systems

Yves Guiraud|ArXiv.org|Feb 7, 2007

Logic, programming, and type systems11 references3 citations

TL;DR

This paper introduces a novel methodology for proving termination of left-linear term rewriting systems (TRSs) using Albert Burroni’s polygraphs—higher-dimensional rewriting systems that model algebraic circuits. By translating a TRS into a minimal polygraphic representation and applying polygraphic interpretations, the method achieves finer implicit complexity bounds than traditional polynomial interpretations, particularly for functional programs, and can handle systems that resist simplification orders.

ABSTRACT

We present a methodology for proving termination of left-linear term rewriting systems (TRSs) by using Albert Burroni's polygraphs, a kind of rewriting systems on algebraic circuits. We translate the considered TRS into a polygraph of minimal size whose termination is proven with a polygraphic interpretation, then we get back the property on the TRS. We recall Yves Lafont's general translation of TRSs into polygraphs and known links between their termination properties. We give several conditions on the original TRS, including being a first-order functional program, that ensure that we can reduce the size of the polygraphic translation. We also prove sufficient conditions on the polygraphic interpretations of a minimal translation to imply termination of the original TRS. Examples are given to compare this method with usual polynomial interpretations.

Motivation & Objective

• Address the limitations of traditional polynomial interpretations in term rewriting, which conflate computation size and result size, leading to overly conservative bounds.
• Overcome the challenge of proving termination for TRSs that lack simplification orders, such as division computation on natural numbers.
• Develop a polygraphic framework that differentiates between constructors and functions, enabling more precise complexity analysis.
• Provide a systematic method to reduce the size of polygraphic translations while preserving termination guarantees for left-linear TRSs.
• Establish theoretical links between polygraphic termination and TRS termination, enabling implicit complexity bounds for functional programs.

Proposed method

• Translate a left-linear TRS into a minimal polygraph using a refined version of Yves Lafont’s general translation, minimizing the number of rewriting rules.
• Apply polygraphic interpretations—assignments of polynomials to 1-cells (gates) with monotonic heat functions and multiset orders on current flows.
• Use structure 3-cells (polygraphic cells) to model duplication, erasure, and permutation operations explicitly, ensuring correctness in the translation.
• Leverage the multiset order on heat production to compare interpretations of terms: if the source term produces strictly more heat than the target, it is considered greater.
• Prove termination by showing that every rewriting step strictly decreases the interpretation under a multiset order, relying on monotonicity of the heat function and current propagation rules.
• Ensure that the polygraphic interpretation respects the structure of the original TRS by encoding rule dependencies and recursive patterns via 2-paths and 3-cells.

Experimental results

Research questions

• RQ1 Can polygraphic interpretations provide tighter implicit complexity bounds than standard polynomial interpretations for functional programs?
• RQ2 How can the size of the polygraphic translation be minimized while preserving termination guarantees for left-linear TRSs?
• RQ3 Can polygraphic interpretations prove termination for TRSs that do not admit simplification orders, such as the division function?
• RQ4 What conditions on the polygraphic interpretation ensure that termination of the polygraph implies termination of the original TRS?
• RQ5 How can the mathematical structure of polygraphic interpretations be further developed to support conditional rules and advanced program constructs?

Key findings

• For the double function, the polygraphic interpretation yields a complexity bound of $X$, significantly better than the $3X$ bound from standard polynomial interpretations.
• The method successfully proves termination of the division-on-naturals TRS, which resists simplification orders due to embedding in the right-hand side of a rule.
• Polygraphic interpretations can differentiate between function symbols and constructors, enabling more precise complexity analysis, similar to dependency pairs.
• The minimal polygraphic translation ensures that only essential structural rules (for duplication, erasure, permutation) are included, reducing the size of the generated system.
• The multiset order on heat production, combined with monotonicity of the heat function, ensures that every rewriting step strictly decreases the interpretation, thus proving termination.
• The method generalizes beyond standard polynomial interpretations and can be extended to handle conditional rules and if-then-else constructs in functional programs.