Information Technology Reference
In-Depth Information
change patterns are applicable to traditional information/workflow systems
as well.
Sometimes, the control-flow structure of a process model can remain intact
but the behavioral aspects of a model could have been changed. For exam-
ple, consider an insurance agency that classifies claims as “high” or “low”
depending on the amount claimed. An insurance claim of
1000 which would
have been classified as high last year is categorized as a low insurance claim
this year due to the organization's decision to increase the claim limit. The
structure of the process remains intact but the routing of cases changes.
-
Data Perspective:
This class of changes refer to the changes in the require-
ment, usage, and generation of data in a process. Tasks may produce or
require information/data. An example of change in a data perspective is
enabling the execution of a task without the requirement of an otherwise
needed data element
d
.
-
Resource Perspective:
This class deals with the changes in resources, their
roles, and organizational structure, and their influence on the execution of
a process. For example, there could have been a change pertaining to who
executes an activity in what roles in a process. As another example, certain
execution paths in a process could be enabled (disabled) upon the availabil-
ity (non-availability) of resources. Furthermore, resources tend to work in
a particular manner and this bias may change over time. For example, a
resource can have a bias of executing a set of parallel activities in a spe-
cific sequential order. Such biases could be more prominent when a limited
number of resources are available; the addition of new resources can remove
this bias.
e
Based on the duration for which a change is active, one can classify changes into
momentary and permanent. Momentary changes are short-lived and affect only
a very few cases while permanent changes are persistent and stay for a while
[6]. In this paper, we consider only permanent changes. Changes are perceived
to induce a drift in the concept (process behavior). We identify four classes of
drifts as depicted in Fig. 1 based on how they manifest.
-
Sudden Drift:
This corresponds to a substitution of an existing process
M
1
with a new process
M
2
as depicted in Fig. 1(a).
M
1
ceases to exist from
the moment of substitution. In other words, all cases (process instances)
from the instant of substitution emanate from
M
2
. This class of drifts are
typically seen in scenarios such as emergency response planning. As an ex-
ample, airlines and airports changing their security processes due to a new
regulation.
-
Recurring Drift:
This corresponds to the scenario where a set of processes
reappear after some time (substituted back and forth) as depicted in Fig. 1(b).
It is quite natural to see such a phenomenon with processes having a seasonal
influence. For example, a travel agency might deploy a different process to
attract customers during Christmas period. The recurrence of processes may
be periodic or non-periodic. An example of a non-periodic recurrence is a
deployment of a process subject to market conditions. The point of deploy-
ment and duration of deployment are both dependent on external factors
(here, the market conditions).
