1Spring 2005 Specification and Analysis of Information Systems Specifying State-based Behavior With UML Statechart Diagrams Eran Toch erant

1Spring 2005Specification and Analysis of Information Systems

Specifying State-based BehaviorWith UML Statechart Diagrams

Eran Toch

http://www.technion.ac.il/~erant

Spring 2007

2

Outline

• Introduction to Statecharts

• Statechart building blocks– States– Transitions

• Advanced Characteristics – Composite States– Parallel States

3

Modeling Process in UML

Intro | Building Blocks | Advanced

PhaseActionsOutcome

InitiationRaising a business needBusiness documents

RequirementsInterviewing stakeholders, exploring the system environment

Organized documentation

SpecificationAnalyze the engineering aspect of the system, building system concepts

Formal specification

DesignDefine architecture, components, data types, algorithms

Formal Specification

ImplementationProgram, build, unit-testing, integrate, documentation

Testable system

Testing & Integration

Integrate all components, verification, validation, installation, guidance

Testing results, Working sys

MaintenanceBug fixes, modifications, adaptationSystem versions

4

What’s Missing in Behavior Modeling?


p : Product : ShooppingCart

addProduct (p)

customer

display)(

getPrice)(

checkout ()

In Sequence diagram, we do not really know how the state of the shopping cart changes.

5

Modeling States

• How can we model the state of the shopping cart?

• Depends on:– Object state )e.g. created, destroyed(– Variable assignment– Relation status )i.e. number of items(– Operating methods & processes– History

addNewProduct)p : Product, quantity : int(removeProduct)p : Product(

dateCreated : DatedateUpdated : Date

ShoppingCart

productId : Stringname : Stringphoto : Picture

Product

**

quantity

Item


6

The Automata

• A machine whose output behavior is not only a direct consequence of the current input, but of some past history of its inputs.

• Characterized by an internal state which represents this past experience.

If the phone is ON, then clicking will turn it off

If the phone is OFF, then clicking will turn it on


7

Automata – Mathematical Foundations

• An Automata is a 5-tuple: S,, ,S0– S is a set of states is an alphabet - finite set of symbols is the transition function: : S S – S0 is an initial state

Many uses:– Stochastic processes )x, y & z are probabilities(– Formal and natural languages )x, y & z are characters(– Complex systems )x, y & z are…(


offoff ononclick

clickGiven a state and an input, the automata jumps to a new stateS0

8

Automata Modeling with Statecharts

• A Statechart augments deterministic automata, making them suitable for specifying behavior of complex systems, using:– Visual formalism– Rich alphabet model– State Modularity – Parallel behavior

• Developed by David Harel )Weizman Inst.(


9

Outline




10

States & Classes

Order

status: {created, sent, received}hasProblems : boolean

gatherItems)(updateInventory)(sendOrder)(received)customer(sendReciept)(

createdcreated

sentsent

State view


Class view

in gatheringin gathering

is received

State

Transition

in problem

11

Activities

• When holding the state, the component can perform activities.– Can be continuous, taking a

specific, or unbounded, time– Activities within a state are

sequential– Activities can be interrupted

• There are special types of activities: enter, exit and event classified activities

in gathering

do: gatherItems()do: updateInventory()

in gathering

do: gatherItems()do: updateInventory()


created

enter: set hasProblems to falsedo: init gathering listexit: clean log file

created

enter: set hasProblems to falsedo: init gathering listexit: clean log file

12

Transitions

xx yyevent [guard] | action

The event that triggers the transition

Conditions that must be met for the transition to take place

The action that takes place when the transition is taken


13

Events

• General Events– Method calls: received)customer(

– Event signals: inventoryReady

• Time event– interval expiry;

tm)3 days(, tm)1000 ms(

– calendar/clock time; when)11:50(

• Change Event:– Change in value of some entity;

false)hasProblems(

true)hasProblems(


createdcreated

sentsent


is received

inventoryReady

tm)3 days(

false)hasProblems(

received)(

in problem

14

Guards (Conditions)

• Boolean expressions.• Evaluated when the transition is triggered• Types of guards:

– Simple predicate: [hasProblems], [x > 0]– Combined predicates:

[hasProblems )hasProblems order.sum < 100]– Guards on activities: [active)gatherItems(]– State related )we’ll get back to it later(


sentsent


gatheringFinished [all items were gathered]

in problemin problemgatheringFinished [items are not found]

15

Guards - Example

SellingSelling

UnhappyUnhappy

bid [)value >= 100( AND )value < 200(]

HappyHappybid [value >= 200]

bidbid [value < 100]


16

Static Conditional Branching

• A graphical shortcut for convenient rendering of decision trees

[)value >= 100( & )value < 200(]

[value >= 200]

[value < 100]

SellingSellingSellingSelling

UnhappyUnhappyUnhappyUnhappy

HappyHappyHappyHappy

bid


17

Empty Transitions

• A transition can have any combination )including none( of the events, guards and actions

• When a transition does not have an event, it is taken after all the activities were ended

Cleaningdo: put waterdo: put soap

do: wash soapdo: drain

doing nothing

Employee

rest[break]

work

Workingdo: shovel

Empty Transition

work


18

Guards and Events

• What’s the difference between the two machines?

S1 S2E1

S3true(C)

S1 S2E1

S3 [C]

What happens if C changes to True before E1?


19

Actions

• An executable atomic computation

• Types of actions– Variable assignment:

– Throwing a signal:

– Start, or stop activities )and concatenation of actions(:

sentsentsentsent is receivedis receivedreceived)( | status := received

sentsentsentsent is receivedis receivedreceived)( | throw)InventoryUpdate(

sentsentsentsent is receivedis received| start)sendBill(; stop)delivery(; x := x+1


20

Transitions - advanced

• Self-transitions: Transitions can be directed to the same state:

• Un-deterministic states – when two transitions are taken in the same time, one of will be taken in an un-deterministic fashion:

S2 E1 / c:=c+1

/ c=0

S1 S2E1

S3

[C1]


21

How does a Washing Machine Works?

• On / Off button. Start button• )No stop button.(• Feedback is given on the

current stage )soaking, rinsing, draining, drying(

• Three plans:– Regular– Delicate )no soaking(– Super delicate )no soaking, no

drying(

• Off can be clicked only before starting, or after finishing


22

Washing Machine


off

idle

Do: light)soak(Do: pump)in(

soak

Do: light)rinse(Do: stir)(

rinse

Do: light)drain(Do: pump)out(

drain

Do: light)dry(Do: stir)(

dry

click)power( click)power(

click)start( [plan=regular]

click)start( [plan=delicate or super delicate]

tm)30 mins(

tm)30 mins(

Tm)5 mins( [plan=super delicate] | light)off(

tm)5 mins([plan=not super

delicate]

Tm)10 mins( | light)off(

23

Outline




24

State Explosion: An Example

What is the off button can be clicked at any time?

What if we want to show how many minutes left to the end of the cycle?


What if we want to come back to the same state we left?

off

idle


soak


rinse


drain


dry

click)power( click)power(



tm)30 mins(

tm)30 mins(



delicate]


25

Abstraction in Statechart

Finding Common Behavior Separating Independent Regions

Composite States Parallel States


26

The State Explosion Problem

• Let– n: Num of Classes– m: Num of variables )assume equals among classes(

• Number of possible states = 2)nm(

• And...– What if the state space of each variable > 2– What about association between objects?

Class

v1 : {t, f}v2 : {t, f}v3 : {t, f}...

Class

v1 : {t, f}v2 : {t, f}v3 : {t, f}...

Class

v1 : {t, f}v2 : {t, f}v3 : {t, f}...

...


27

Composite States


On

off

idle


soak


rinse


drain


dry



tm)30 mins(

tm)30 mins(



delicate]


click)power(click)power(

28

Composite + History

On

off

idle


soak


rinse


drain


dry



tm)30 mins(

tm)30 mins(



delicate]


click)power(

click)power(

H

29

deep and shallow

DiagnosingDiagnosingDiagnosingDiagnosing

Diagnostic1Diagnostic1

setupsetupsetupsetup

Check blood Check blood pressurepressure

Check blood Check blood pressurepressure

Diagnostic2Diagnostic2

Find a veinFind a veinFind a veinFind a vein

Draw BloodDraw BloodDraw BloodDraw Blood

H*H*


HH

StopStopStopStoppause

return

return-full

Shift ended | clh)(

clh)( – clears the history

30

Completion Transitions

• Triggered by a completion event– generated automatically when an immediately nested

state machine terminates

CommittingCommittingCommittingCommitting

Phase1Phase1Phase1Phase1

Phase2Phase2Phase2Phase2Commit DoneCommit DoneCommit DoneCommit Done

completion transition )no trigger(


31

LampFlashingLampFlashing

off/off/

FlashOffFlashOff

FlashOnFlashOn

Triggering Rules

• Two or more transitions may have the same event trigger– inner transition takes precedence– if no transition is triggered, event is discarded

on/on/

on/on/


32

Order of Actions: Complex Case

• Same approach as for the simple case

S1S1exit:exS1exit:exS1




S2S2entry:enS2entry:enS2




| initS2| initS2E/actEE/actE

Actions execution sequence:

exS11 exS1 actEenS2 initS2 enS21


33

Parallel States

• Sometimes, some aspect of an entity are independent of other aspects

single

married

divorced

Martial

StaffMember

Manager

employee


34

Parallel States Annotation

• Combine multiple simultaneous descriptions

StaffStaffMemberMember

ManagerManager

employeeemployee

StaffStaffMemberMember

ManagerManager

employeeemployee

singlesingle

marriedmarried

divorceddivorced

MartialMartial

singlesingle

marriedmarried

divorceddivorced

Martial Martial


35

Interactions Between Parallel States


On

idle

soak

rinse

drain

idle

Do: pump in water

in

Regular

Delicate

click)start( [Plan in Regular]

Entered)soak(

Tm)30([Plan not in Super Delicate]

Tm)30(

click)plan(

Pump PlanController

Super Delicate

click)plan(

click)plan(

click)start([Plan not in Regular]

Tm)30(

Tm)30([Plan in

Super Delicate]

idle

Rotor

Do: spin)150rpm(

Spin Slow

Do: pump out water

out

Entered)rinse(

Entered)drain(

Do: spin)400rpm(

Spin Fast

Exited)drain(

Exited)rinse(Tm)entered)rinse(,15(

Entered)rinse( [Plan not in Regular]

off

click)power(

click)power(

H

36

“Flat” and Parallel Machines

Use of Orthogonal

Regions

No Use of Orthogonal

Regions

B

C

E

G

F

A D

(in G)

C,F

C,E C,G

B,GB,E

B,F

• Every parallel machine can be transformed into a sequential machine:


37

Transition Forks and Joins

• For transitions into/out of orthogonal regions:

Credit Card verificationCredit Card verification

chargingcharging

ProcessingProcessing SentSent ConfirmedConfirmed

shippingshipping

Receipt SentReceipt Sent


In problemIn problem

[ok]

[not ok] Joinfork

38

Summary

Statechart:– State-based modeling– Based on Automatas

Statechart building blocks– States– Transitions

Advanced Characteristics – Composite States– Parallel States

Documents

1Spring 2005 Specification and Analysis of Information Systems Specifying State-based Behavior With UML Statechart Diagrams Eran Toch erant