Model Checking ofWorkflow Nets with Tables and Constraints

TL;DR

This paper introduces WFTC-nets, an enhanced model for workflow systems with database constraints, and develops a model checking approach using a new state reachability graph to verify complex design requirements.

Contribution

It redefines WFT-nets as WFTC-nets with guard constraints, proposes a new SRG generation method to avoid pseudo states, and introduces DCTL for verifying database-related properties.

Findings

The new SRG method effectively avoids pseudo states.

DCTL can express complex database-related requirements.

Experiments demonstrate the approach's practical usefulness.

Abstract

Many operations in workflow systems are dependent on database tables. The classical workflow net and its extensions (e.g., worflow net with data) cannot model these operations so that some related errors cannot be found by them. Recently, workflow nets with tables (WFT-nets) were proposed to remedy such a flaw. However, when the reachability graph of a WFT-net is constructed by their method, some pseudo states are possibly generated since it does not consider the guards that constrain the enabling and firing of transitions. Additionally, they only considered the soundness property that just represents a single design requirement, so that many other requirements, especially those related to tables, cannot be analyzed. In this paper, therefore, we re-define the WFT-net by augmenting constraints of guards to it and re-name it as workflow net with tables and constraints (WFTC-net). We propose a new method to generate the state reachability graphs (SRG) of WFTC-nets such that SRG can avoid pseudo states, due to the consideration of the guards in it. To represent design requirements related to database operations, we define database-oriented computation tree logic (DCTL), to represent more design requirements. We design the model checking algorithms of DCTL based on the SRG of WFTC-nets and develop a tool. Experiments on a number of public benchmarks show the usefulness of our methods.