Verification of Data-Aware Processes via Array-Based Systems (Extended Version)

Abstract

We study verification over a general model of artifact-centric systems, to assess (parameterized) safety properties irrespectively of the initial database instance. We view such artifact systems as array-based systems, which allows us to check safety by adapting backward reachability, establishing for the first time a correspondence with model checking based on Satisfiability-Modulo-Theories (SMT). To do so, we make use of the modeltheoretic machinery of model completion, which surprisingly turns out to be an effective tool for verification of relational systems, and represents the main original contribution of this paper. In this way, we pursue a twofold purpose. On the one hand, we reconstruct (restricted to safety) the essence of some important decidability results obtained in the literature for artifact-centric systems, and we devise a genuinely novel class of decidable cases. On the other, we are able to exploit SMT technology in implementations, building on the well-known MCMT model checker for array-based systems, and extending it to make all our foundational results fully operational.

SMT-based Safety Verification of Parameterised Multi-Agent Systems

“…In this paper we present a verification technique based on an SMT [8] approach for ABS [26,27,6,5,15], characterising its soundness and completeness. This is a very well-understood SMT-based theory for which a number of results of practical applicability already exist, and research is active [27,15,28,12,16,18,13,19,14,25,17]. This is the first paper to establish a formal connection between verification of PMASs and the long-standing tradition of SMT-based model checking for ABS.…”

“…• the formula in (9) makes sure that the environment and at least one concrete agent with id j and type t can execute a synchronisation action, which is then written in a global variable env act as well as in the array position arr At [j]; • then, concrete agents can select the same action, updating their action array as specified by (10); • in a further step, that is only enabled when for every concrete agents it is true that either env act is not executable or was indeed selected, the phase is progressed to S 2 . This is captured by formula (11); • when the current phase is S 2 , a bulk update is finally performed as in formula (12), which also updates the environment. In both cycles above, when the initial phase 0 is reached again, the action arrays arr At of each template T t are reset to contain nop values.…”

“…The input of the software is a simple, textual representation of an ABS, including (i) the declaration of individual and array variables, (ii) the initial state formula, (iii) the goal state formula, (iv) the list of transition formulae (i.e. either formulae (1)-( 5) or formulae ( 6)- (12) in Sec. 4).…”

Petri Nets with Parameterised Data: Modelling and Verification (Extended Version)

“…This MCMT transition can be executed only if flag init_fl, denoting whether the initial marking assignment has taken place, is TRUE. 5 It works as follows: Note that the flag should be previously declared using the MCMT statement :global init_fl BOOLE. Same holds for the boolean constants TRUE and FALSE: they are declared using the respective statements :smt (define TRUE ::BOOLE) and :smt (define FALSE ::BOOLE).…”

“…On the other hand, such transitions can inject data into tokens by extracting relevant values from the database or by generating genuinely fresh ones. We systematically encode CLog-nets into the most recent version of MCMT 1 [15], one of the few model checkers natively supporting the (parameterised) verification of data and processes [5,8,9]. We show that fresh-value injection is a particularly complex feature to handle, and discuss strategies to tame it.…”