Monitoring Object-Centric Constraints in Business Processes 3
Figure 1(a) are violated by the case. For instance, Constraint 1 is violated since
pick item is followed by send invoice in the case and Constraint 3 is violated
since pick item is preceded by send invoice.
However, in reality, the order and each item have different lifecycles as shown
in Figure 1(b). First, we place an order with two items. While the invoice is sent
and the payment is collected for the order, we check the availability of each item
and pick each of them. We finally deliver the order with two items after packing
two items together. In this object-centric setting, constraints should be defined
in relation to objects to provide accurate insights. For instance, Constraint 1*
extends Constraint 1 with the corresponding object type (i.e., item). Contrary
to Constraint 1,Constraint 1* is not violated since pick item is directly followed
by pack item for any items. Moreover, we can analyze more object-centric con-
straints by considering the interaction of different objects. First, we can analyze
if an execution of an activity involves (un)necessary objects (cf. Constraint 4
and Constraint 5 ). Also, we can analyze the cardinality of objects for executing
an activity (cf. Constraint 6 and Constraint 7 ).
In this work, we propose a technique for constraint monitoring in object-
centric settings. To this end, we first introduce object-centric behavioral metrics
that can be computed from Object-Centric Event Logs (OCELs), e.g., a metric
to measure the degree to which pick item precedes pack items in the lifecycle of
items. Next, we develop Object-Centric Constraint Graphs (OCCGs) to formally
represent constraints using such metrics. Finally, monitoring engine evaluates
the violation of the constraints represented by OCCGs by analyzing OCELs.
We have implemented a web application to support the approach. A demo
video and a manual are available at https://github.com/gyunamister/ProPPa.
git. Moreover, we have conducted case studies with a production process and a
Purchase-To-Pay (P2P) process supported by an SAP ERP system.
The remainder is organized as follows. We discuss the related work in sec-
tion 2 and present the preliminaries, including OCELs in section 3. In section 4,
we introduce object-centric behavioral metrics. Afterward, we present OCCGs
to formally represent constraints and the monitoring engine to evaluate the vio-
lation of the constraints in section 5. Next, section 6 introduces the implementa-
tion of the proposed approach and case studies using real-life event data. Finally,
section 7 concludes the paper.
2 Related Work
Many approaches have been proposed to monitor the violation of constraints
by analyzing event data. Weidlich et al. [11] propose a technique to abstract
process models to behavioral profiles and produce event queries from the profile.
Violated executions of events are monitored using Complex Event Processing
(CEP) engines with the event queries. Awad et al. [5] define a set of generic
patterns regarding the occurrence of tasks, their ordering, and resource assign-
ments and generate anti-patterns from the generic patterns to monitor event
executions. Maggi et al. [8] represent control-flow properties of a running pro-