Table of Contents
Fetching ...

IMELL Cut Elimination with Linear Overhead

Beniamino Accattoli, Claudio Sacerdoti Coen

TL;DR

This work addresses the time-cost modeling of cut-elimination in IMELL by building on the Exponential Substitution Calculus (ESC) and its good strategy. The authors design SESAME, an environment-based abstract machine that implements the good strategy with linear overhead, and provide an OCaml implementation to validate the theoretical bounds. They establish key invariants (sub-term property, diamond property, contextual decoding) and prove that SESAME, together with a final garbage-collection pass, yields cut-free terms in a bi-linear overhead relative to the initial term and reduction length. The approach advances the explicit-management of duplication and garbage collection in a strong-evaluation setting for linear logic, with potential implications for implicit computational complexity within IMELL-based formalisms.

Abstract

Recently, Accattoli introduced the Exponential Substitution Calculus (ESC) given by untyped proof terms for Intuitionistic Multiplicative Exponential Linear Logic (IMELL), endowed with rewriting rules at-a-distance for cut elimination. He also introduced a new cut elimination strategy, dubbed the good strategy, and showed that its number of steps is a time cost model with polynomial overhead for the ESC/IMELL, and the first such one. Here, we refine Accattoli's result by introducing an abstract machine for ESC and proving that it implements the good strategy and computes cut-free terms/proofs within a linear overhead.

IMELL Cut Elimination with Linear Overhead

TL;DR

This work addresses the time-cost modeling of cut-elimination in IMELL by building on the Exponential Substitution Calculus (ESC) and its good strategy. The authors design SESAME, an environment-based abstract machine that implements the good strategy with linear overhead, and provide an OCaml implementation to validate the theoretical bounds. They establish key invariants (sub-term property, diamond property, contextual decoding) and prove that SESAME, together with a final garbage-collection pass, yields cut-free terms in a bi-linear overhead relative to the initial term and reduction length. The approach advances the explicit-management of duplication and garbage collection in a strong-evaluation setting for linear logic, with potential implications for implicit computational complexity within IMELL-based formalisms.

Abstract

Recently, Accattoli introduced the Exponential Substitution Calculus (ESC) given by untyped proof terms for Intuitionistic Multiplicative Exponential Linear Logic (IMELL), endowed with rewriting rules at-a-distance for cut elimination. He also introduced a new cut elimination strategy, dubbed the good strategy, and showed that its number of steps is a time cost model with polynomial overhead for the ESC/IMELL, and the first such one. Here, we refine Accattoli's result by introducing an abstract machine for ESC and proving that it implements the good strategy and computes cut-free terms/proofs within a linear overhead.
Paper Structure (64 sections, 22 theorems, 5 figures)

This paper contains 64 sections, 22 theorems, 5 figures.

Key Result

lemma 1

Let $t$ be a proper term.

Figures (5)

  • Figure 1: The Exponential Substitution Calculus (ESC).
  • Figure 2: IMELL has a type system for ESC.
  • Figure 3: Definitions for the good strategy: dominating free variables and good contexts.
  • Figure 4: The Basic Abstract Machine (BAM).
  • Figure 5: The Strong Exponential Substitution Abstract Machine without Erasure (SESAME).

Theorems & Definitions (51)

  • definition 1: Positions
  • definition 2: Clashing, smooth, clash-free terms
  • lemma 1: DBLP:journals/lmcs/Accattoli23
  • proposition 1: Postponement of garbage collection, DBLP:journals/lmcs/Accattoli23
  • definition 3: Basic evaluation, answer
  • lemma 2
  • definition 4: Out cuts, out variables
  • definition 5: Cut-freeness up to garbage
  • definition 6: Good steps, good strategy
  • theorem 1: Properties of the good strategy, DBLP:journals/lmcs/Accattoli23
  • ...and 41 more