The General Vector Addition System Reachability Problem by Presburger Inductive Invariants

The reachability problem for Vector Addition Systems (VASs) is a central problem of net theory. The general problem is known to be decidable by algorithms exclusively based on the classical Kosaraju-Lambert-Mayr-Sacerdote-Tenney decomposition. This decomposition is used in this paper to prove that the Parikh images of languages recognized by VASs are semi-pseudo-linear; a class that extends the semi-linear sets, a.k.a. the sets definable in Presburger arithmetic. We provide an application of this result; we prove that a final configuration is not reachable from an initial one if and only if there exists a semi-linear inductive invariant that contains the initial configuration but not the final one. Since we can decide if a Presburger formula denotes an inductive invariant, we deduce that there exist checkable certificates of non-reachability. In particular, there exists a simple algorithm for deciding the general VAS reachability problem based on two semi-algorithms. A first one that tries to prove the reachability by enumerating finite sequences of actions and a second one that tries to prove the non-reachability by enumerating Presburger formulas.

💡 Research Summary

The paper tackles the classic reachability problem for Vector Addition Systems (VAS), a cornerstone of net theory and concurrent system verification. While decidability of VAS reachability has been known for decades through the intricate Kosaraju‑Lambert‑Mayr‑Sacerdote‑Tenney (KLMS‑T) decomposition, existing algorithms are notoriously complex and provide little insight when a target configuration is unreachable. The authors introduce a fresh perspective by linking VAS behavior to a newly defined class of sets—semi‑pseudo‑linear (SPL) sets—which strictly extend the well‑studied semi‑linear sets (the sets definable in Presburger arithmetic).

The technical core begins with a meticulous re‑examination of the KLMS‑T decomposition. By interpreting the decomposition in terms of Parikh images, the authors prove that the Parikh image of any language recognized by a VAS belongs to the SPL class. An SPL set can be described as a finite union of base vectors together with a finite set of “pseudo‑linear” generators that allow limited non‑linear growth, yet remain expressible by Presburger formulas. This result bridges the gap between the infinite state space of VASs and the decidable world of Presburger arithmetic.

Armed with the SPL characterization, the paper turns to inductive invariants. An inductive invariant I is a set of configurations that contains the initial configuration s₀ and is closed under all VAS transitions (i.e., for every transition τ, τ(I) ⊆ I). The authors show that if a configuration s_f is not reachable from s₀, then there exists a semi‑linear inductive invariant that includes s₀ but excludes s_f. Crucially, any semi‑linear set can be encoded as a Presburger formula, and the authors prove that the property “φ defines an inductive invariant for a given VAS” is decidable. Consequently, non‑reachability can be reduced to the existence of a Presburger formula satisfying a simple syntactic condition.

Based on these insights, the paper proposes a dual‑semi‑algorithm framework. The first semi‑algorithm is the traditional forward search: it enumerates finite sequences of actions (transition words) of increasing length, checking whether any such sequence leads from s₀ to s_f. This semi‑algorithm is complete for the positive case: if s_f is reachable, it will eventually find a witness. The second semi‑algorithm enumerates candidate Presburger formulas in a systematic way (e.g., by bounding the number of variables, coefficients, and quantifier depth). For each candidate φ, a decidable check verifies whether φ holds for s₀, fails for s_f, and is inductive. If such a φ exists, the algorithm halts, delivering a verifiable certificate of non‑reachability. Because both procedures are semi‑decidable and mutually exclusive, running them in parallel yields a simple, conceptually clean decision procedure for the general VAS reachability problem.

Beyond the theoretical contribution, the authors discuss practical implications. A non‑reachability certificate expressed as a Presburger formula can be fed to off‑the‑shelf SMT solvers (Z3, CVC5, etc.) for independent validation, making the approach attractive for safety verification of concurrent programs, Petri nets, and counter systems. Moreover, the SPL framework may inspire new abstraction techniques for other infinite‑state models, as it captures a richer class of behaviors while staying within a decidable logical fragment.

The paper concludes with several avenues for future work: refining the enumeration strategies for Presburger invariants to improve practical performance, extending the SPL analysis to richer models such as VAS with resets or transfer arcs, and conducting extensive experimental evaluation on benchmark VAS instances to compare the proposed dual‑semi‑algorithm against state‑of‑the‑art KLMS‑T‑based tools. In summary, the work delivers a theoretically elegant and practically promising method for VAS reachability, unifying classical decomposition techniques with modern logical invariants and providing checkable certificates for both positive and negative instances.