Information Technology Reference
In-Depth Information
Definition 2 (Weak Order Relation). Let P =( A, s, e, C, F, T ) be a process
model and
T P its set of traces. The weak order relation
P
A
×
A contains all
pairs ( x, y ) , such that there is a trace σ = n 1 ,...,n m in
T P with j
∈{
1 ,...,m
m for which holds n j = x and n k = y .
Two activity nodes of a process model might be related by weak order in three
different ways, which define the characteristic relations of the behavioural profile.
1
}
and j<k
Definition 3 (Behavioural Profile). Let P =( A, s, e, C, F, T ) be a process
model. A pair ( x, y )
A
×
A is in one of the following relations:
The strict order relation
P ,if x
P y and y
P x .
The exclusiveness relation + P ,if x
P y and y
P x .
The interleaving order relation
|| P ,if x
P y and y
P x .
The set
B P =
{ P , + P ,
|| P }
of all three relations is the behavioural profile of P .
We illustrate the relations of the behavioural profile for the lower model in Fig. 1.
For instance, it holds (1)
(5) as both activities are ordered whenever they oc-
cur together in a trace. It holds (8) + (9) as both activities will never occur in
a single trace, and (5)
(6) due to the potential concurrent execution. The rela-
tions of the behavioural profile are mutually exclusive. Along with the reverse
strict order
||
1 =
, the relations partition the
Cartesian product of activities. An activity is either said to be exclusive to itself
(e.g., (1) + (1) in Fig. 1) or in interleaving order to itself (e.g., (6)
{
( x, y )
( A
×
A )
|
( y, x )
}
(6)). The
former holds, when an activity cannot be repeated, whereas the latter implies
that the activity may be executed multiple times. The behavioural profile is a
behavioural abstraction. Hence, there may be process models that show differ-
ent trace semantics but have equal behavioural profiles. Moreover, behavioural
profiles may be lifted from activities to labels of activities. Still, this may re-
sult in serious information loss as two occurrences of activities with equal la-
bels would not be distinguished. The behavioural profile
||
B P of a process model
P =( A, s, e, C, F, T ) may be restricted to a subset of activity nodes A
A by
to A ×
A .We
restricting the definition of the three relations
P ,+ P ,and
|| P
refer to this restricted profile as the behavioural profile over A .
Computation of the behavioural profile of a process model is done eciently
under the assumption of soundness. Soundness is a correctness criteria that guar-
antees the absence of behavioural anomalies, such as deadlocks [25]. It has been
defined for WF-nets, so that it can be directly applied to our notion of a pro-
cess model. We are able to reuse techniques for the computation of behavioural
profiles introduced for sound free-choice WF-nets [16]. Using these results, be-
havioural profiles are computed in O ( n 3 ) time with n as the number of nodes of
the respective WF-net.
3 A Set Algebra for Behavioural Profiles
To introduce a set algebra for behavioural profiles, we first discuss a notion of
strictness for profile relations in Section 3.1. Then, Section 3.2 introduces set-
theoretic relations, while Section 3.3 turnsthefocusonset-theoretic operations.
 
Search WWH ::




Custom Search