Conceptual modelling for the management of environmental · PDF file1 Territories,...

1Territories, Environment, Remote Sensing & Spatial Information Joint Research Unit Cemagref - CIRAD - ENGREF

Conceptual modelling for the Conceptual modelling for the management of environmental datamanagement of environmental data

Thérèse LIBOUREL [email protected]

ProfessorMontpellier Laboratory of Computer Science, Robotics and Microelectronics

CNRS – University of Montpellier 2

METIER Graduate Training Course no. 2 METIER Graduate Training Course no. 2 –– Montpellier Montpellier -- February 2007February 2007

Information Management in Environmental SciencesInformation Management in Environmental Sciences

2 / 71

Conceptual modelling Conceptual modelling for for the the management of management of environmental environmental datadata

OverviewOverview

• Introduction

• Object concepts and UML formalism

• Functional model

• Structural model

• Dynamic model

• EIS specificities

• Conclusion and perspectives

3 / 71


OverviewOverview

• Introduction




• Dynamic model



4 / 71


IntroductionIntroduction

Environmental information systems

Environment

Legislation, policy

Social

5 / 71



Physical organization of data

Data storage structures

Accelerating structures (index)

Dependent on the data model

Dependent on the DBMSPhysical

model


Independent of the DBMS

Logical

model

Independent of the data model


Conceptual

model

The real

6 / 71


The approach and the modelling are essential

…. for future extensions…. for exchange

A model is a simplification/abstraction of the reality

Constructing models allows us to better understand the systems we are developing …


7 / 71



Why methods?

A necessity: reduce the gaps

The real The computer-based world

• Evolutionary

• Ambiguity

• Codified languages

• Unique semantics

8 / 71


The methods = structuring guides

• Decomposition of work

• Organization into stages

• Founding concepts

• Semi-formal representations

Ensure a reproducible approach to obtain reliable results


9 / 71


An analysis and design method proposes an approach that distinguishes the

development stages of a software’s life cyclebases itself on a representation formalism that

facilitates communication, organization and verification

The modelling languageproduces documents (models) that facilitate feedback

on the design and evolution of applications

To summarize …


10 / 71


The trend towards EIS

• Systemic methods vs. object methods

• E/R vs. OO data modelling

• From closed formalisms to extensible formalisms (ex. UML)

• From aspatial and atemporal formalisms to formalism with spatialand temporal extension


11 / 71


• CONGOO [Pantazis and Donnay, 1996]• Geo-ER [Hadzilacos and Tryfona, 1997]• Geo-OM [Tryfona et al., 1997]• GeoOOA with its modelling software [Kösters et al.,

1997]• MADS with its modelling software [Parent et al., 1997]• Modul-R with its modelling software and its automatic

code generator [Bédard and Paquette, 1989; Pageau and Bédard, 1992; Caron et al., 1993; Bédard et al., 1996] …

PERCEPTORY • OMEGA and AIGLE [El Bath, 1997] • POLLEN [Gayte et al., 1997]


12 / 71


People Project

Product

Process

Tools

Artefact, result

participants

automatization

*

*

*

Keywords

The 4 P’s

Object-oriented approach for projects


13 / 71


UML: a formal description language

Accepted by the OMG – 1997A unique common language:

A meta-model A less ambiguous language A simple graphical notation,

Comprehensible by non-computer-specialists Allows communication between actors

Has become THE reference for object modelling


14 / 71


OverviewOverview

• Introduction




• Dynamic model



15 / 71


The perception of models The graphical views (diagrams)

Class diagramsObject diagrams

Sequence diagramsCollaboration diagrams

State-transition diagramsActivity diagrams

Use case diagramsComponent diagramsDeployment diagrams

Object concepts and ULM formalismObject concepts and ULM formalism

16 / 71


System functions fromthe user’s point of view.

Physicalcomponentsof an application.

Representationof behaviour interms of states.

Static structure of classes and relationships between these classes.

Schemas for installationof components on the hardware mechanisms.

Representation ofoperational behaviourin terms of actions.

Representation of objects,mutual links and potentialinteractions.

Objects and basic relationshipsbetween these objects.


17 / 71


Construction structure

Model orientation by use cases

Use cases

Logic Implementation

Processes Deployment

classes/dynamics components

use case + scenarios

scenarios on the componentsconcurrencedistributiontolerance to breakdowns

Components ‘projected’on the hardware


18 / 71


• Descriptive models from the users’ points of view

• Functional scenarios

The manner of using the system

The ‘USE CASEs’

Functional model Functional model

19 / 71


Two concepts

Actor

Use case

Actor (Producer)

Actor (Decider)

<<communicate>>

<<communicate >>

Displayindicators

Managedata


20 / 71


Actor (role 1)

Actor (role 2)

‘use’‘extend’

Use cases can be linked by relationships– ‘use’ usage (decomposition)

– ‘extend’ refinements (exceptions processing)


21 / 71


Diagram of the ‘static context’

system

Actor (role 2)

Actor (role 1)

‘actor’ role

association

0..1

0..1

0..*


22 / 71


In UML, the structural or static model is described using two types of diagrams

Class diagramsdescription of all or part of a system in an abstract way, in terms of classes, structure and associations.

Object diagramsdescription of configuration examples of all or part of the system, in terms of objects, values and links.

Structural model Structural model

23 / 71


Objects in the real world Computer-based objects

Hidden internalstate

Visiblebehaviour

The objects


24 / 71


Behaviour influences stateState reflects past behaviours

Object

State evolves over time

Behaviour actions and reactions

Identity essence


25 / 71


Sophie

Alain

System

DB

Luc

: Professor

: Discipline

Two objects orinstances


26 / 71


Discipline

LabelnbhCoursenbhPracticals

Describe ()Inscribe (e: Student)

Operations and methods

Methods

Implementations

name of the

class

attributes

operations


27 / 71


Class

Attributes (properties)

Discipline


: Discipline

Label =ACOnbhCourse=12nbhPracticals=15

Instance

Attribute values (State)

‘is-instance-of ‘


28 / 71


Association/Link(analogy Class/Instance)

Discipline Teacher

label name

Is-taught

Association

: Discipline

label=ACO:Teacher

name = Louis

Is-taught

Link


29 / 71


Association in general is binary (degree = 2) but...

Member copyborrow

ReadingMechanism

read

Association name

Binary association

Ternary association


30 / 71


Multiplicity and an association’s roles

Company Individual

Nameaddress

nameBirth dateSS no.address

employs

works-for

employer employee

head

worker

staff

* 1..*

1..*

0..1


31 / 71


Enterprise Individual

nameaddress

namedate of birthaddress

quantity

Possesses-shares

capital shareholder

Association classAssociation class

PossessionLine of

portfolio

* 1..*


32 / 71


Other ‘abstractions’

• Specific associations (composition/aggregation)

• Specialization/generalization


33 / 71


Specific association All/part

Title bar B’ground Scroll barBorder

IndicatorTitle Close button Arrow

Window

0..2

2

CompositionStructural model Structural model

34 / 71


Aggregation

Semantic Collection/Item

Tree

State

Forest1

1..n

Region

Country

1

1..n

1

1..n


35 / 71


Composition/Aggregation

Constraints

- Exclusivity/Sharing

- Dependence/Independence

Propagation/Distribution


36 / 71


Generalization/Specialization

Mechanism for intellectual inference of characteristicsEither we refine (specialization)Or we abstract (generalization)

SemanticSet-theoretic point of viewLogical point of view


37 / 71


Individual

name address

Teacher

grade address

teach

{disjoint}

Student

ID no. address



38 / 71


Pump Heat exchanger

Tank

Centrifugal pump

Submersiblepump

Pressure tank

Equipment

...

Type of equipment

...

Type of pump

...

Type of tank

Generalization/SpecializationStructural model Structural model

39 / 71


Terrestrial vehicle Aquatic vehicle

Car Amphibious vehicle Ship

Vehicle

Multiple generalization/Specialization


40 / 71


Composition/Aggregation or Generalization/Specialization?

• Aggregation � link between instances�an aggregation tree is composed of objects

that are part of a composite object

• Generalization � link between classes


41 / 71


The constraints

• The constraints are predicates, that can apply to several items of the static model, which have to be verified at all times.

• The constraints allow details to be taken into account at a very fine granularity in a class diagram. They can express conditions or restrictions.

• In UML, the constraints are expressed in textual form, between brackets and, preferably, in OCL (Object Constraint Language).

• Constraints are inherited.


42 / 71


The constraints

Route

Edge

*

1..*

Individual Committee

presides *1

memberOf**

{subset}

{ordered} Constraint on

association extremity

Constraint between two associations


43 / 71


active: Real {value ≥ 0}passive: Real

The constraints

Constraint on class

Individual

head

subordinate

<commands

Company

{ active = passive }

{ Individual.employer =Individual.head.employer }

employer

* 1..* 0..10..1

Constraint on attribute

Constraints on 2 associations


44 / 71


• Collaboration diagrams

• Sequence diagrams

• State-transition diagrams

• Activity diagrams (not covered)

Describes interactions between objectsand changes that take place over time

Dynamic model Dynamic model

45 / 71


Communication

Region

State

message

Country

Collaboration diagram


46 / 71


Poitou:Region

France:Country

4: m4

Bretagne:Region

LR:Region

Gard: State

Hérault:StateAude:State

Country

Region

State

….

1:Population()

1.2:Population()

1.3:Population()

1.1:Population()

1.3.1:Population() 1.3.2:Population()

1.3.3:Population()


47 / 71


Sequence diagram

B CA

M1

M2

M3

M4

M6

M5

TIME


48 / 71


Life line

‘create’

Creation by ‘create’ message

Activation of object thatexecutes an operation op

Destruction by another object

:C1

‘destroy’

op


49 / 71


France Poitou Bretagne LR

t im

e

A H G

Popul()

Popul()

Popul() Popul()

Popul()

Popul()

Popul()


50 / 71


Event and State

• State of an object

– values of its attributes and of its links

– over time an object can change state

• Event

– stimulus from an object towards another object


51 / 71


Initial

Final

Simple

Complex

Creditor

State name

entry/op1exit/ op2

on evt1/ op3on evt2/ op4do/ activity

At the beginning

At the end

During the event

All the time

Internal activities

Notation of states


52 / 71


Notation of arcs

label

label• event(parameters)• [condition]• /action


53 / 71


State diagramsStates of a bank account

Open Close

CreateRequest()

DebtorWe withdraw/premiumsWe deposit/increase balance

Close()

CreditorWe withdraw/debit balanceWe deposit/increase balance

[Balance >=0]

[Balance < 0]

open


54 / 71


OverviewOverview

• Introduction




• Dynamic model



55 / 71


EIS specificities

Spatiality

Temporality

Which representations?...

56 / 71


Taking the geometry into account

Roadname: stringspeed: integer

Create()Display()ChangeSpeed(v)

Line

Point

2..*

0..*

0..1

1

x: realy: real

Highway

EIS specifities EIS specifities

57 / 71


Representations of space

Euclidean geometry

point (dim 0) line (dim 1) area (dim 2) volume (dim 3)

Segment Point

XY

Has-as-extremities>

0..1 2..2

Is-on: boolean


58 / 71


Polyline Point

XY

0..1 2..n

Is-on: boolean

closed

Segment

1..n

0..1

0..1

2..2

{order}

Euclidean geometry



59 / 71


Polygon Point

XY

0..1 3..n

Belongs: bool.

Segment

3..n

0..1

2..2

2..2

{order}

Euclidean geometry



60 / 71


Polygon Point

XY

0..11..1

Belongs: bool.

Segment

3..n

1..2

0..n

1..1

orientation

0..n

start

end

Euclidean geometry



61 / 71


Node Edge

2..2 1..n

Is-linked-to

0..n

delimits

0..n


Graphs

Node, Edge


62 / 71



Topology…


63 / 71



Topology…


64 / 71


Representation of time …

Instant Interval

1 0..n

0..n1

end

start

operators

duration


65 / 71


And the dynamics …

Objects’ life line

Movement

displacement

growth

deformation

Life

creation

stability

destruction

fission fusion


66 / 71



:Country

:Country

evolution

fission

‘create’

:Country

t im

e


67 / 71


OverviewOverview

• Introduction




• Dynamic model



68 / 71


Conclusion

Approach having ‘proven’ itself

Example: Aggregative perceptions of the landscape

Landscape Parcel

Type of land use

….

1..*

Landscape unit

surface

Consists of

1..*

combining

[Forest, heath, cultivated area,…]

[Lardon & al., 2000]

69 / 71


Conclusion

OpenGIS standard

70 / 71


Perspectives

• Taking the ‘cognitive’ into account

ST pictograms

PVL (Perceptory)

0D 1D 2D …

Mads

71 / 71


Perspectives

Complex projects

• towards the GIS’s

(geodatabase)

• towards object-oriented DBMS extensions

thematic and spatial integration

Instrumentation: AGL

1Territories, Environment, Remote Sensing & Spatial Information Joint Research Unit Cemagref - CIRAD - ENGREF

Conceptual modelling for the Conceptual modelling for the management of environmental datamanagement of environmental data

Thérèse LIBOUREL [email protected]

ProfessorMontpellier Laboratory of Computer Science, Robotics and Microelectronics

CNRS – University of Montpellier 2

METIER Graduate Training Course no. 2 METIER Graduate Training Course no. 2 –– Montpellier Montpellier -- February 2007February 2007

Information Management in Environmental SciencesInformation Management in Environmental Sciences

2 / 71


OverviewOverview

• Introduction




• Dynamic model



3 / 71


OverviewOverview

• Introduction




• Dynamic model



4 / 71



Environmental information systems

Environment

Legislation, policy

Social

5 / 71



Physical organization of data

Data storage structures

Accelerating structures (index)


Dependent on the DBMSPhysical

model



Logical

model

Independent of the data model


Conceptual

model

The real