11

CS/CPE 426Senior Projects

Chapter 21: State MachinesChapter 22:Advanced State Machines

[Arlow and Neustadt 2005]

March 24, 2009

22

Outline State machines:

IntroductionState machine diagramsStatesTransitionsEvents

Advanced state machines:Composite states

SimpleOrthogonal

Submachine statesSubmachine communicationHistory

33

Introduction

Both activity diagrams and state machine diagrams model system behavior

However, they have different semantics:Activity diagrams are based on Petri Nets and

usually model processes when several objects participate

State machine are based on Harel’s statecharts and typically used to model single reactive objects

44

Introduction

Reactive objects:Respond to external eventsMay generate and respond to internal

eventsHave a lifecycle modeled as a progression

of states, transitions and eventsMay have current behavior that depends

on past behavior State machines are used to model

behavior of classifiers such as classes, use cases, subsystems, systems

55

Introduction

There are two types of state machines: Behavioral state machines Protocol state machines

State machines are most commonly used to model dynamic behavior of classes

In UP, state machines can be used in: requirements, analysis, design

A significant challenge is testing state machines: manual walkthrough, simulation, code generation + test harnesses for state machines

66

State machine diagrams There are three main modeling elements in state

diagrams: states transitions, and events. Example of a simple state machine, Fig. 21.2 [Arlow &

Neustadt]

77

States

Summary of UML state syntax, Fig.21.4 [Arlow & Neustadt 2005]

88

TransitionsSummary of UML syntax for transitions in behavioral state diagrams, Fig.21.5 [Arlow & Neustadt 2005]

Where: event(s)= internal or external occurrence(s) that trigger the transition guardCondition = boolean expression, when true the transition is

allowed anAction = some operation that takes place when the transition fires

99

Transitions

Summary of UML syntax for transitions in protocol state

diagrams, Fig.21.6 [Arlow & Neustadt 2005]

Note that there is no action and guard conditions are replaced by preconditions and postconditions.

1010

TransitionsA junction pseudo-state represents a point where transitions

merge or branch, e.g. Fig.21.7 [Arlow & Neustadt 2005]

1111

TransitionsA junction pseudo-state may have more than one output

transition(protected by mutually exclusive guard conditions) e.g.

Fig.21.8[Arlow & Neustadt 2005]

1212

TransitionsChoice pseudo-state can also be used, e.g. Fig. 21.9 [Arlow &

Neustadt2005]

1313

Events

Events can be of four types:Call eventSignal eventChange eventTime event

1414

Events

Example of a call event, Fig.21.11 [Arlow & Neustadt 2005]

1515

Events

A signal is a package of information sent asynchronously between objects. Example of signal event Fig. 21.12 [Arlow & Neustadt 2005]

Example of sending a Signal Fig. 21.13 [Arlow & Neustadt 2005]

1616

EventsChange events are positive edge triggered.Example of a change event, Fig. 21.15 [Arlow & Neustadt

2005]

1717

Events

Time events are indicated by the keywords when and after. Example of a time event, Fig. 21.16 [Arlow & Neustadt,

2002]

1818

Example of astate machine[Dascalu 2001]

Stopped MoveWithSpeedOne

MoveWithSpeedTwo

Blocked

go1_command /speedOne()

target_reached /stop()

go2_command /speedTwo()

go2_command /speedOne()

limit_reached or otheremergency condition /

stop ()

permission_to_restart /reset ()

limit_reached or otheremergency condition /

stop ()

stateinitial state

transition

event action

change_direction_command /reverseDirection()

final state

off

off

From the stopped state, the 2-speed motor brings the probe in atarget position by "accelerating" with speed 1, "crusing" with speed2, "deccelerating" with speed 1, and then stopping. If out of limitsor in case of another emergency, the motor stops immediately andrequires the operator's permission to restart working.

1919

Composite states

A composite state contains one or more nested state machines (submachines), each existing in its own region, Fig 22.2 [Arlow & Neustadt 2005].

The composition icon is shown in Fig. 22.4

2020

Simple composite states A superstate that contains a single region is called a

simple composite state, e.g. Fig 22.5 [Arlow & Neustadt 2005]

2121

Orthogonal composite states Orthogonal composite states consist of two or more sub-

machines that execute in parallel. In Fig 22.6 [Arlow & Neustadt 2005] there are two such composite states, Initializing and Monitoring

2222

Orthogonal composite states

The composite state Initializing, Fig 22.7 [Arlow & Neustadt 2005]

2323

Orthogonal composite statesThe composite state Monitoring, Fig 22.8 [Arlow & Neustadt 2008]

2424

Submachine states

A submachine state is a special state that references a state machine recorded in a separate diagram, e.g. Fig. 22. 9 [Arlow and Neustadt 2005]

2525

Submachine states Fig. 22. 10 [Arlow & Neustadt 2005]. The notation

for a submachine state is

<state name : name of referenced state machine diagram>

2626

Submachine communication

Asynchronous submachine communication can be achieved via attributes:

The modeled reactive object has a set of attributes that can be used by submachines

The communication mechanism: one machine sets attributes and the other uses the attribute values in guard conditions of their transitions

2727

Submachine communication Example of communication via attributes, Fig. 22.11

[A&N 2005]

2828

HistoryExample of using the shallow history indicator, Fig 22.12 [Arlow &

Neustadt 2005]

2929

History

Example of using the deep history indicatorFig 22.13 [Arlow & Neustadt 2005]