Process Modeling Umberto Nanni Enterprise Information Systems 1 - - PowerPoint PPT Presentation

process modeling
SMART_READER_LITE
LIVE PREVIEW

Process Modeling Umberto Nanni Enterprise Information Systems 1 - - PowerPoint PPT Presentation

D IPARTIMENTO DI I NGEGNERIA INFORMATICA AUTOMATICA E GESTIONALE A NTONIO R UBERTI Master Degree Programme in Manage gement ment Engin inee eeri ring En Enterprise erprise In Information formation Sys ystems ems Umbe mberto to Nan


slide-1
SLIDE 1

1 Enterprise Information Systems Umberto Nanni

Master Degree Programme in

Manage gement ment Engin inee eeri ring

En Enterprise erprise In Information formation Sys ystems ems

Umbe mberto to Nan anni ni

DIPARTIMENTO DI INGEGNERIA INFORMATICA AUTOMATICA E GESTIONALE ANTONIO RUBERTI

Process Modeling

slide-2
SLIDE 2

2 Enterprise Information Systems Umberto Nanni

Processes to be modeled

service

primary activities

support

the primary activities are the customers

management and control

regulate relations between directional and operational component

physical

manipulation of physical

  • bjects

information

information management

business

business process in the value chain

slide-3
SLIDE 3

3 Enterprise Information Systems Umberto Nanni

Some models for processes

  • Hierarchic
  • State Diagram (finite state machine)
  • DFD - Data Flow Diagram
  • WIDE - Workflow on an Intelligent and Distributed

database Environment

  • Action Workflow
  • Petri Nets
slide-4
SLIDE 4

4 Enterprise Information Systems Umberto Nanni

model:Hierarchical Model

SALES Warehouse management Sales management Product development Planning Order processing Sales Forces management Order reception Shipment Material preparation Invoice preparation Pricing Load Packaging Document preparation Macroprocess Process Phase Activity Operation

slide-5
SLIDE 5

5 Enterprise Information Systems Umberto Nanni

model:DFD – Data Flow Diagram

label label

process Data collection interface (external agent) Data flow

label label label label

slide-6
SLIDE 6

6 Enterprise Information Systems Umberto Nanni

Data Flow

A data flow (a directed link) represents any kind

  • f flow (e.g., material or information):
  • a first endpoint must be a process
  • a second endpoint can be a process, or an

interface, or a data collection

  • a data flow can be elementary or structured
  • a data flow can be an element of a collection
slide-7
SLIDE 7

7 Enterprise Information Systems Umberto Nanni

Sample DFD – Plan of Study (PoS)

PoS compo- sition Plan of Study student check PoS exam PoS PoS

  • utcome
slide-8
SLIDE 8

8 Enterprise Information Systems Umberto Nanni

Considerations on DFD

  • accompany the DFD

– a DATA DICTIONARY – the textual description of each process

  • the first DFD (single process, with interfaces

and interchange flows) is called “CONTEXT DIAGRAM”

  • hierarchical decomposition of processes

(consistent with numbering)

  • limitations of DFD ...
slide-9
SLIDE 9

9 Enterprise Information Systems Umberto Nanni

DOT, USA, 1998

slide-10
SLIDE 10

10 Enterprise Information Systems Umberto Nanni

DOT, USA, 1998

slide-11
SLIDE 11

11 Enterprise Information Systems Umberto Nanni

model:WIDE – Workflow on an Intelligent and

Distributed database Environment

  • PROCESS model

– description of the activities and control structure – similar to Activity Diagram (UML)

  • INFORMATION model

– data and documents necessary for the execution

  • ORGANIZATION model

– structure and agents

slide-12
SLIDE 12

12 Enterprise Information Systems Umberto Nanni

Elements of the WIDE model

<name> task business process fork/join cycle conditional fork begin/end control flow supertask

slide-13
SLIDE 13

13 Enterprise Information Systems Umberto Nanni

Travel Agency

[BS01, SI per PA: Metodologie e Tecnologie, CNIPA]

Travel selection Travel cancellation Travel confirmation !Approved Approved Preparation of documents Shipment of documents Request of payment

slide-14
SLIDE 14

14 Enterprise Information Systems Umberto Nanni

model:Action Workflow

  • activity

– working task

  • action

– requesting or making commitments

  • commitment

– interaction CUSTOMER - SUPPLIER

  • process

– structured set of activities

slide-15
SLIDE 15

15 Enterprise Information Systems Umberto Nanni

Action Workflow – Process modeling

  • 1. Request (Preparation)
  • 2. Acceptance (Negotiation)
  • 3. Production (Execution)
  • 4. Supply (Satisfaction)

CUSTOMER PROVIDER 1 2 3 4

TRANSACTION

slide-16
SLIDE 16

16 Enterprise Information Systems Umberto Nanni

Schemas and sample modeling

Source:

http://www.cnipa.gov.it/site/it-IT/ /La_Documentazione/In_archivio/Quaderni_AIPA/

Quaderno 1 (Introduction):

“Reingegnerizzazione dei Processi e Workflow Management” http://www.cnipa.gov.it/site/_contentfiles/00464500/464560_bpr.pdf

slide-17
SLIDE 17

17 Enterprise Information Systems Umberto Nanni

Basic Structure

customer provider TRANSACTION (primary workflow) (secondary workflow) (secondary workflow) (secondary workflow) (secondary workflow)

slide-18
SLIDE 18

18 Enterprise Information Systems Umberto Nanni

Contextualization to public administration

Citizen Public Administration

slide-19
SLIDE 19

19 Enterprise Information Systems Umberto Nanni

Process: land registry office

slide-20
SLIDE 20

20 Enterprise Information Systems Umberto Nanni

model:Petri Nets

place transition Petri net: bipartite graph G=(PT,A), A  (P×T)  (T×P)

Structure

slide-21
SLIDE 21

21 Enterprise Information Systems Umberto Nanni

Petri nets: a model for processes [Carl Adam Petri, 1962]

A Petri net is a 4-tuple P=(P,T,A,M0) where: P set of places T set of transitions A set of arcs A(PT)(TP) M0 is the initial marking M: PN 3 )

slide-22
SLIDE 22

22 Enterprise Information Systems Umberto Nanni

Some notation and terminology

  • each transition tT has an input set °tP and an output set

t°P (also called input/output places for t)

  • the same notation applies to places, set of places, set of

transitions; e.g.: °p3={t1} p°3={t2} {t1, t2}°={p1, p3, p4, p5}

  • when a transition t has at least one token in each input place,

it is enabled (or firable): p°t, M(p)1 t1 Input set for t1: °t1={p1, p2} Output set for t1: t1°={p3, p4} p1 p2 p3 p4 t2 t3 In the current marking: t1 is enabled, while t2 and t3 are not. We write: p5 M0 

t1

slide-23
SLIDE 23

23 Enterprise Information Systems Umberto Nanni

Evolution of a Petri net

An enabled transition can fire, deleting a token in each input place and creating a token in each output place

t1 t2 t3

A firable sequence is a sequence of transitions s = t1,t2,...,tn, where M0  M1  M2 ...  Mn i.e., M0  Mn

s

t1 t2 tn p1 p2 p3 p4 p5 M0=(2,1,0,0,0) M1=(1,0,1,1,0) M0  M1

t1

slide-24
SLIDE 24

24 Enterprise Information Systems Umberto Nanni

Potential evolution of a Petri net

Given a net P=(P,T,A,M0):

  • a potentially firable transition tT is such that there exists a

prefix sequence tT*, where tt is firable: M0 

  • a potentially firable sequence of transitions sT* is such that

there exists a prefix sequence tT*, where ts is firable: M0 

  • a reachable marking M is such that there exists a firable

sequence s such that M0  M

  • R(M0) is the set of reachable markings
  • PR is the set of reachable places:

PR = {pP |  MR(M0), M(p)>0}

tt ts

s

slide-25
SLIDE 25

25 Enterprise Information Systems Umberto Nanni

A special case: Finite State Machine

A (Finite) State Machine is a Petri net where, for each transition t, both the input and the output places have cardinality one:  tT, |°t|=|t°|=1 Example:

slide-26
SLIDE 26

26 Enterprise Information Systems Umberto Nanni

Workflow nets [Wil van der Aalst 1997]

A workflow net is a Petri net where:

  • 1. there exist two special places

a source i with |°i|=0 a sink o with |o°|=0

  • 2. for each node xPT, x is reachable from the source,

and the sink is reachable from x

i a b c d

  • 2

1 3 4 5 6 7

slide-27
SLIDE 27

27 Enterprise Information Systems Umberto Nanni

Importance of WF-nets

  • robust, since they are based on a well formalized and

studied model – Petri nets

  • strong properties which reduce the cost of evaluating

properties which would be impractical to test on general Petri nets

  • some constructs have been added in order to comply

more practical needs in modeling business processes (sometimes reintroducing complexity and unclear semantics)

  • established formalism to represent Business Processes,

with high impact in industry, with translation tool from/to main industrial models (e.g., BPEL)

  • a pragmatic consideration: van der Aalst has collected

more than 51000 citations – 34000 in the last 5 years