Lesson 16: Advanced Topics in Behavior Analysis

1
Lesson 16Advanced Topics in Behavior Analysis
Object-Oriented Modeling
Applications

2
Lesson Objectives

• Describe the characteristics of Real-Time Systems
• Understand action tables
• Understand picket fence representation
• Learn about timing diagram
• Understand leveling behavior models
• Learn about concurrency

3
State-Transition Diagrams

• Related Terms state machine, state chart, finite
  automaton
• Definition
• A STD shows a sequence of states an object (a
  class, a package, or a system) can assume during
  its lifetime (execution), together with the
  stimuli that cause changes of state.
• A STD describes a hypothetical machine (finite
  automaton) which at any given time is found in a
  set of finite states.
• It consists of
• A finite, non-empty set of states
• A finite, non-empty set of events
• Functions (actions) which describe the transition
  from one state to the next
• An initial state
• A set of final states

4
State-Transition Diagrams Events/Conditions/Opera
tions

• A Transition is triggered by one or more events.
• An Event consists of a name and a list of
  possible argument.
• Conditions can be attached to the event or
  conditions may be formulated independently of a
  specific event.
• Events/Conditions can trigger actions inside the
  state that are realized through appropriate
  operations (actions).
• Three special triggers are predefined
• entry fires automatically when entering a state
• exist fires automatically when leaving a state
• do fires repeatedly as long as the state is
  active

5
State-Transition Diagrams Events/Conditions/Opera
tions

• An event is an occurrence which has a particular
  significance in a given context, can be localized
  in space and time, and needs to be taken into
  account because it triggers a state transition.
• An Event may have the following causes
• A condition (defined for a transition) is
  satisfied.
• The object receive a message.
• Transitions are usually triggered by events which
  are noted on the arrows that connect the states.
• Transitions without event specifications are
  triggered automatically as soon as the actions
  belong to a state (inside) or previous state (on
  transition) are terminated.
• 0 or more Event(arguments) 0 or more conditions
  /operations (arguments)

6
Characteristics of Real-Time Systems

• Problems are formulated in scientific or
  engineering terms
• RTSs contain sensors and actuators to interact
  with their environment instead of with humans
• A RTS has sensors to sense important physical
  parameters, such as temperature, pressure, or
  speed
• A RTS has actuators, such as valves and
  thermostats that can make direct changes to the
  environment without going through a human
  intermediary.
• Require a degree of coordination similar to the
  eye-hand coordination of people.
• RTSs often require concurrent processing of
  multiple inputs - Examples
• Temperature, pressure, acidity may have to be
  sensed and used together to calculate how to
  control a chemical reaction.
• Valves and heaters may have to be adjusted
  simultaneously to maintain the desired flow rate,
  temperature, and reaction rate.

7
Characteristics of Real-Time Systems

• RTSs operate at very short time scales
• Systems like a nuclear reactor may responses in
  milliseconds or even microseconds.
• Such rapid response times may sometimes be at the
  limit of the available technology.
• As a result, there may be several technological
  constraints in the design of the system.
• RTSs require higher precision than ordinary
  systems.
• Constraints are imposed by the task, not the user
• These constraints guarantee safety, liveness, and
  timeliness
• Safety means that responses match specifications
  and the system is fault-tolerant
• Liveness means that the system responds to all
  events
• Timeliness means that the system responds within
  time constraints imposed by the environment

8
Comments on the Characteristics of RTSs

• The list of previous RTS characteristics is not
  mandatory
• A method for designing RTSs must be able to deal
  with all of the RTS characteristics.
• There are no method deals with all the
  characteristics
• Methods for RTS developments
• ROOM
• Berard
• O-ET

9
STD for the Control Temperature Process
Start
Stop
Idle State
Cold Enable Gas()
Cold or Cool Enable Electric()
Hot Enable Refrign()
Gas Heating
Cold Enable Gas()
Cool Disable Gas()
Electric Heating
OK Disable Electric()
Hot Disable Electric() Enable Refrign()
Cold oe Cool Disable Refrign()
Refrigeration
10
An Action Table Equivalent to previous STD
SIGNALS
Cold Cool OK Hot
STATUS
Enable Electric()
Enable Gas() Enable Electric()
Enable Refrigeration()
Idle Gas Heat Electric Heat Refrigeration
Disable Gas()
Disable Electric() Idle-State
Enable Gas()
Disable Electric() Enable Refrign()
Disable Refrign() Idle-State
Disable Refrign() Idle-State
11
Timing Diagram
Idle Warming Running Cooling
12
Timing Diagram (contd)

• Timing Diagrams help in understanding a complex
  sequence of related actions and activities within
  a thing or between instances
• A Timing Diagram represents the relative time at
  which an action or activity occurs
• Each row in the diagram denotes a separate
  action or activity

13
Timing Diagram (contd)

• When the lines are drawn
• Low gt the action or activity is considered not
  to be taken place
• Highgt the action or activity is considered to be
  taken place
• When one activity must precede another, a
  freehand line can be drawn to show the precedence
  relationship
• Timing Diagram can be annotated to show the
  birth and death of object instances, and the
  formation and breakup of relationships, etc.

14
Leveling Behavior Models
Idle
On
O F F
Stopped
Start
O N
Accelerating
Decelerating
Off
Done
At Speed
Running
15
Leveling Behavior Models (contd)

• Behavior models can be leveled
• Leveling suppresses details to increase
  understanding
• Allows the representation of very large behavior
  models, in manageable pieces
• Two popular approaches
• Sub-class/Super-class Class-based
• Composite object structure Instance-based

16
Concurrency

• Aggregation Concurrency
• A dynamic model describes a set of concurrent
  objects, each with its state and STD.
• The objects in a system are inherently concurrent
  and can change state independently
• The state of the entire system cannot be
  represented by a single state in a single object
  it is the product of the states of all its
  objects.
• Aggregation implies concurrency.
• The aggregate state corresponds to the combined
  states of all the components STDs
• Aggregation is the and relationship
• The aggregate state is one state from the first
  STD, and a state from the second STD, and a state
  from each other STD.

17
Concurrency (contd)

• Concurrency within an object
• Concurrency within the state of a single object
  arises when the object can be partitioned subsets
  of attributes and links, each o which has its own
  sub-STD. The state of an object comprises one
  state from each sub-STD

18
Discussion Questions
Define An action table Concurrency Timing
Diagram What are the differences between
aggregation and concurrency? What are the
characteristics of RTSs?