The verification of safety properties for concurrent systems often reduces to the coverability problem for Petri nets. This problem was shown to be ExpSpace-complete forty years ago. Driven by the concurrency revolution, it has regained a lot of interest over the last decade. In this paper, we propose a generic and simple approach to solve this problem. Our method is inspired from the recent approach of Blondin, Finkel, Haase and Haddad [3]. Basically, we combine forward invariant generation techniques for Petri nets with backward reachability for wellstructured transition systems. An experimental evaluation demonstrates the efficiency of our approach.The one-step binary relation → is the union of these t-step relations. Formally, m → m ′ ⇔ ∃t ∈ T : m t − → m ′ . The many-step binary relation * − → is the reflexivetransitive closure of →.Example 2.1. Figure 1 depicts a simple Petri net N = (P, T, F, m init ) with places P = {p 1 , p 2 , p 3 }, transitions T = {t 1 , t 2 , t 3 } and flow function F such

Abstract. The verification of safety properties for concurrent systems often reduces to the coverability problem for Petri nets. This problem was shown to be ExpSpace-complete forty years ago. Driven by the concurrency revolution, it has regained a lot of interest over the last decade. In this paper, we propose a generic and simple approach to solve this problem. Our method is inspired from the recent approach of Blondin, Finkel, Haase and Haddad [3]. Basically, we combine forward invariant generation techniques for Petri nets with backward reachability for wellstructured transition systems. An experimental evaluation demonstrates the efficiency of our approach.

Driven by the concurrency revolution, the study of the coverability problem for Petri nets has regained a lot of interest in the recent years. A promising approach, which was presented in two papers last year, leverages a downward-closed forward invariant to accelerate the classical backward coverability analysis for Petri nets. In this paper, we propose a generalization of this approach to the class of well-structured transition systems (WSTSs), which contains Petri nets. We then apply this generalized approach to lossy channel systems (LCSs), a well-known subclass of WSTSs. We propose three downward-closed forward invariants for LCSs. One of them counts the number of messages in each channel, and the other two keep track of the order of messages. An experimental evaluation demonstrates the benefits of our approach.