A Taxonomy of Adaptive Workflow Management

1
A Taxonomy of Adaptive Workflow Management

• Yanbo Han
• Amit Sheth

Christoph Bussler
ART The Boeing Company Seattle, WA
LSDIS Lab University of Georgia Athens,
GA http//lsdis.cs.uga.edu
2
Example 1 Changes in Business Rules
New rule says any trip costing 3000 must be
approved by financial controller
Migrating Workflows Cichocki Rusinkiewicz,
NATO ASI, Istanbul, August 1997
3
Example 2a

• In a hospital, inpatient treatment of a patient
  may follow one of many possible work procedure.
  Standard plan expands constantly in response to
  medical advancements, changes of insurance and
  hospital policies, discovery of new diseases, or
  new treatment methods, drugs, devices, etc.

4
Example 2b

• If initial exam identifies needs for additional
  procedure in the blood test, the task/subprocess
  component is obtained from the repository and
  added to the subprocess

standard test
get blood
write report
special test
This example is from Han et al.
5
Example 3

• Allocating new patient service request (or case)
  to a pool of nurses where number of nurses may
  change (change of shift-- fewer nurses in night
  shift, vacation/sickness-- fewer nurses than
  yesterday, etc)

Option 1 Assign new case to smallest
worklist. Option 2 Maintain a common case queue,
assign when nurse is available.
o o o
6
Example 4

• In a telephone service order provisioning
  environment, there are five LFACS systems. We
  know that a particular type of task is to be
  processed by a LFACS system, but the specific
  LFACS system can be determined only at run-time
  based on the telephone number in a service
  provisioning request.

7
Example 5

• A physician orders laboratory tests for a
  patient, but cannot wait for the results in case
  of an emergency. Thus, he/she may start a
  treatment process to treat the emergent case. As
  soon as the test results arrive, they should
  trigger an action to let the physician know they
  are available. Note that nobody knows exactly
  when the results will arrive. After receiving
  the results, the physician may need to modify
  his/her plan.

8
Evolution of WFMS

• Changing Environment
• Business activities and environments, as well as
  may engineering branches in general, are highly
  dynamic and subject to constant evolution.
• Technical Advances
• Technical advances often lead to systems
  reconfiguration, due to, for example, replacement
  and updating of software components, addition of
  new components, and changes in component
  interfaces.

9
Ad-hoc Derivation

• Dynamic refinement
• unavailability of a complete spec
• User involvement
• users decision-making needs to be considered
• Unpredictable events
• external stimuli, intervention of users, timeout
• Erroneous situations
• system failure, erroneous operations

10
Classification of workflow adaptation
Adaptation of WF system to a changing business
context (e.g., policy change)
Domain

• Process
• Schema
• Task

Model evolution ad-hoc changes to model instances

• Resource adjustment
• Components interface
• Human resources
• Data-related adaptation

• Resource
• Software component
• Organization model
• Data Model

Infrastructure
System re-configuration
11
Mechanisms for Adaptive WF 1 of 2

• Meta-model Approach
• use meta-model to determine the structure and
  types of constituent components define a set of
  primitives to change model or model instance
• Open-point approach
• define special points in a workflow model, where
  adaptation can be made

12
Mechanisms for Adaptive WF 2 of 2

• Synthesized approach
• Flexible composition and dynamic hierarchy of
  workflow models
• systematic management and dynamic binding of
  workflow resources
• local decision making and user involvement
• Exception handling and adaptive WFM
• infrastructure support

13
METEORs ORBWork supports a variety of adaptation
See NATO ASI paper
14
Future emphasisCoordination and Collaboration
for Virtual Teams
Current Environment
Proposed Environment