[Paper Review] Simulation Over One-counter Nets is PSPACE-Complete

Piotr Hofman, Sławomir Lasota+2 more|arXiv (Cornell University)|Oct 23, 2013

Simulation Techniques and Applications4 references4 citations

TL;DR

This paper establishes that both strong and weak simulation preorder checking between one-counter nets (OCNs) is PSPACE-complete. The authors provide a constructive proof of the Belt Theorem, using a PSPACE algorithm that verifies simulation conditions via local checks over periodic structures in the state space, and extend this to weak simulation by adapting techniques from prior semilinear approximant methods.

ABSTRACT

One-counter nets (OCN) are Petri nets with exactly one unbounded place. They are equivalent to a subclass of one-counter automata with just a weak test for zero. Unlike many other semantic equivalences, strong and weak simulation preorder are decidable for OCN, but the computational complexity was an open problem. We show that both strong and weak simulation preorder on OCN are PSPACE-complete.

Motivation & Objective

• To close the gap in computational complexity for simulation preorder on one-counter nets (OCNs), which were known to be decidable but lacked complexity bounds.
• To provide a constructive proof of the Belt Theorem, showing simulation relations in OCNs can be described by finitely many periodic partitionings of the state grid.
• To develop a PSPACE algorithm for checking strong simulation preorder on OCNs using local verification and a shifting window technique over periodic structures.
• To extend the PSPACE algorithm to weak simulation preorder by adapting the semilinear approximant method from prior work, ensuring polynomial-size representations of intermediate relations.
• To establish PSPACE-completeness for both strong and weak simulation preorder on OCNs, resolving an open problem in infinite-state system verification.

Proposed method

• Constructively prove the Belt Theorem by characterizing simulation relations as finitely many partitionings of the grid ℕ×ℕ induced by parallel lines, ensuring periodic structure.
• Design a PSPACE algorithm that verifies simulation conditions locally by maintaining a polynomial-sized sliding window over the state grid, checking transitions between neighboring counter values.
• Use a binary counter to track the number of window shifts, ensuring termination within exponential bounds while using only polynomial space.
• Adapt the semilinear approximant method from [4] to weak simulation, showing that each approximant relation over modified OCNs can be computed in polynomial size.
• Leverage the fact that strong simulation over OCNs admits semilinear representations to compute and verify approximants inductively.
• Prove that the entire simulation relation cannot be represented in polynomial size, implying that PSPACE is optimal for the problem.

Experimental results

Research questions

• RQ1 What is the computational complexity of strong simulation preorder on one-counter nets?
• RQ2 Can a constructive proof of the Belt Theorem be used to derive an effective PSPACE algorithm for strong simulation on OCNs?
• RQ3 Is weak simulation preorder on OCNs decidable within PSPACE, and does it admit a polynomial-space algorithm?
• RQ4 Can the semilinear approximant method for weak simulation be adapted to ensure polynomial-size intermediate representations in the algorithm?
• RQ5 Is PSPACE the optimal complexity class for simulation preorder on OCNs, or could it be lower?

Key findings

• Strong simulation preorder on one-counter nets is PSPACE-complete, with a matching PSPACE lower bound established via reduction from [10].
• A new constructive proof of the Belt Theorem is provided, showing that simulation relations in OCNs are described by finitely many periodic partitionings of the state grid.
• The PSPACE algorithm for strong simulation uses a shifting window technique to verify local simulation conditions over periodic structures, using only polynomial space.
• Weak simulation preorder on OCNs is also PSPACE-complete, achieved by adapting the semilinear approximant method with polynomial-size intermediate representations.
• The simulation relation over OCNs cannot be represented as a semilinear set in polynomial size, as this would imply a collapse of the polynomial hierarchy.
• The results confirm that PSPACE is the optimal complexity class for both strong and weak simulation preorder on OCNs, under standard complexity-theoretic assumptions.