Information Technology Reference
In-Depth Information
that for each possible and valid feature configuration there is a corresponding
valid process model that satisfies the well-formedness constraints.
More similar to our objective is the approach for process configuration from
van der Aalst et al. [16]. Their framework ensures correctness-preserving config-
uration of (reference) process models. In contrast to our work, they capture the
variability directly in the workflow net by variation points of transitions. Ac-
cordingly, a configuration is built by assigning a value to the transitions, while
our approach uses feature selections.
Weidlich et al. [8,12] derive behavior profiles to describe the essential be-
havior in terms of activity relations like exclusivity, interleaving and ordering
of activities. Weber et al. [17] extend process models by semantic annotations
and use them for the validation of process behavior correctness that captures
control-flow interaction and behavior of activities. In contrast to our work, their
focus is on behavioral constraints, while we consider structural well-formedness
constraints. Moreover, our particular emphasis is on the feature-oriented process
family representation.
In the context of SPLs several approaches have been introduced, in order to
ensure the well-formedness of solution space models. Czarnecki et al. [18] specify
constraints on solution space model configurations using OCL constraints. Prob-
lem space models, solution space models with OCL constraints, and mappings
between them are transformed to Binary-Decision Diagrams.
Thaker et al. [19] introduce an approach for the verification of type safety, i.e.,
the absence of references to undefined classes, methods, and variables, in solu-
tion space models w.r.t. all possible problem space configurations. They specify
the models and their relations as propositional formulas and use SAT solvers to
detect inconsistencies. Janota et al. [20] and van der Storm [21] introduce ap-
proaches to validate the correctness of mappings between feature and component
models. They use propositional logics too.
8Conluon
As shown in the related work section, our contribution is primarily related to the
validation of families of business processes. While the concept of business process
families was previously introduced and even covered in our own work [22], there
have been very limited (if any) attempts to propose a validation of such families.
Our proposal validates business process models w.r.t. their well-formedness
constraints; mappings to problem space models; and dependencies in the prob-
lem space models. Hence, unlike other approaches on validation of (model-driven)
software product lines, our approach also considers the very nature of business
process models through the set of business process practices encoded in con-
trol flow patterns. Even though, in this paper, we used BPMN for defining the
solution space of business process families, our approach is easily generalizable
to other types of business process modeling languages. This can be deduced
from control flow patterns used in this paper and control flow support analyzes
presented in the relevant literature [23].
 
Search WWH ::




Custom Search