faculteit technologie management workflow mining: current status and future directions ana karla a....

/faculteit technologie management

Workflow Mining: Current Status and Workflow Mining: Current Status and Future DirectionsFuture Directions

Ana Karla A. de Medeiros,

W.M.P van der Aalst and A.J.M.M. Weijters

Eindhoven University of Technology

Department of Information and Technology

[email protected]


Outline

• Motivation

• Workflow Mining: -algorithm

• Workflow Mining: limitations

• Extensions to Mining Algorithms

• Discussion and Future work


Workflow Mining – Motivation

– Workflow Mining (What is the process?)

– Delta analysis (Are we doing what was specified?)

– Performance analysis (How can we improve?)


Workflow Mining – Process log

ABCDABCD

ACBDACBD

EFEF

case 1 : task A case 1 : task A case 2 : task A case 2 : task A case 3 : task A case 3 : task A case 3 : task B case 3 : task B case 1 : task B case 1 : task B case 1 : task C case 1 : task C case 2 : task C case 2 : task C case 4 : task A case 4 : task A case 2 : task B case 2 : task B case 2 : task D case 2 : task D case 5 : task E case 5 : task E case 4 : task C case 4 : task C case 1 : task D case 1 : task D case 3 : task C case 3 : task C case 3 : task D case 3 : task D case 4 : task B case 4 : task B case 5 : task F case 5 : task F case 4 : task D case 4 : task D

• Minimal information in noise-free log: case id’s and task id’s

• Additional information: event type, time, resources, and data

• In this log there are three possible sequences:


Workflow Mining – Ordering Relations >,,||,#

• Direct succession: x>y iff for some case x is directly followed by y.

• Causality: xy iff x>y and not y>x.

• Parallel: x||y iff x>y and y>x

• Unrelated: x#y iff not x>y and not y>x.

case 1 : task A case 1 : task A case 2 : task A case 2 : task A case 3 : task A case 3 : task A case 3 : task B case 3 : task B case 1 : task B case 1 : task B case 1 : task C case 1 : task C case 2 : task C case 2 : task C case 4 : task A case 4 : task A case 2 : task B case 2 : task B ......

A>BA>BA>CA>CB>CB>CB>DB>DC>BC>BC>DC>DE>FE>F

AABB

AACC

BBDD

CCDD

EEFF

B||CB||CC||BC||B

ABCDABCD

ACBDACBD

EFEF


Workflow Mining – -algorithm

Let W be a workflow log over T. (W) is defined as follows.

1. TW = { t T W t },

2. TI = { t T W t = first() },

3. TO = { t T W t = last() },

4. XW = { (A,B) A TW B TW a Ab B a W b a1,a2 A a1#W

a2 b1,b2 B b1#W b2 },

5. YW = { (A,B) X (A,B) XA A B B (A,B) = (A,B) },

6. PW = { p(A,B) (A,B) YW } {iW,oW},

7. FW = { (a,p(A,B)) (A,B) YW a A } { (p(A,B),b) (A,B) YW b

B } { (iW,t) t TI} { (t,oW) t TO}, and

8. (W) = (PW,TW,FW).



A

B

C

D

E F

ABCDABCD

ACBDACBD

EFEF

AABB

AACC

BBDD

CCDD

EEFF

B||CB||CC||BC||B



• If log is complete with respect to relation >, it can be used to mine SWF-net without short loops

• Structured Workflow Nets (SWF-nets) have no implicit places and the following two constructs cannot be used:


• Problematic constructs– Short loops – Invisible Tasks– Synchronization of OR-join places– Duplicate Tasks– Implicit Places – Non-free Choice

Workflow Mining – -algorithm limitations


-algorithm limitations – Short Loops

One-length

Two-length

B>B and not B>B implies BB (impossible!)

A>B and B>A implies A||B and B||A instead of AB and BA


• Problematic constructs– Short loops – Invisible Tasks– Synchronization of OR-join places – Duplicate Tasks – Implicit Places– Non-free Choice



-algorithm limitations – Invisible Tasks

Nets are behaviorally equivalent!


• Problematic constructs– Short loops – Invisible Tasks– Synchronization of OR-join places – Duplicate Tasks – Implicit Places– Non-free Choice



-algorithm limitations – Synchronization of OR-join places

But...


• Problematic constructs– Short loops – Invisible Tasks– Synchronization of OR-join places – Duplicate Tasks– Implicit Places – Non-free Choice



-algorithm limitations – Duplicate Tasks


-algorithm limitations – Implicit Places


-algorithm limitations – Non-free Choice

But...


• Some problematic constructs relate because– Can have same complete workflow log and/or– Same set of ordering relations

Workflow Mining – Trade-offs between Problematic Constructs

XYXY

XAAYXAAY

XXAA

A A YY


• Process mining = 3-phase process1. Pre-processing

2. Processing

3. Post-processing

• Example for 1-lenght loops in SWF-nets

Workflow Mining – Extensions to Mining Algorithms


Workflow Mining – Extensions to Mining Algorithms (Example)


-algorithm cannot mine correctly – Short loops– Invisible tasks– Duplicate Tasks– Implicit Places– Non-free Choice– Synchronization of OR-join Places

• Extensions are possible, but relations between problematic constructs imply in trade-offs

Workflow Mining – Discussion and Future Work


• Future steps– Extend the class of WF-nets -algorithm can correctly

mine– Create mining algorithms to handle workflows beyond

WF-nets– Develop mining heuristics to deal with noisy or

incomplete workflow logs

Workflow Mining – Discussion and Future Work


Questions?

faculteit technologie management workflow mining: current status and future directions ana karla a....

Documents