Systems

1
Systems

• Chapter 4 of the book
• Human Information Thing
• Interaction and technology

2
Overview

• What is a system?
• Characteristics of a system
• Complexity
• What is a model?
• Different kinds of models
• Mathematics
• FSM
• UML

3
Why introduce systems?

• A useful view of reality
• Divides a complex entity into manageable
  components.
• - Does this in practise by modelling from a
  specific view and for a specific purpose
• Find and reuse common principles
• Think engineering, think design (DOIT)
• ? Aesthetics, art, humanism (Holistic thinking)

4
System

• A unified whole made up of subsystems.

Environment
Interface
System (struktur , beteende)
Ex Car, University (student in, ? out)
5
Is H a system?

• What do we loose by viewing H as a system?
• Separate part from the rest, but is this true?

6
System behaviour described as states
Fill up
Many many states on a detailed level of a
complex system (such as a human)
7
Memory and feedback
Delay(memory)
O
I
S

O
I
I
Ex Driving a car, feedback through senses
Alternatives Open loop, feed-forward
(predictive)
8
Adaptive system (survival of the adaptive!)
E.g. Autofocus E.g. University, Goverment
controls, supports parameters (number of
students), labour market
9
Systems that learn
Controlled system
Controller
Critic
Reinforced (critic informs if choice by
controller was good or bad, e.g. learning to
walk) Supervised (extern expert provides right
answers and suggests behaviour, e.g. University
lecture)
10
System characteristics

• Sequential , parallel system
• queue of cars in one lane only,
• Several parallel lanes (still queue ?)
  (Stockholm)

Efficient Redundancy Specialisation
11
Linear, Stable, Equilibrium

• Linear in 0 out, out proportional to in
• Cmp salary per hour (overtime, bonus non linear)
• Stability Small change in no change out
• cmp H-H conversation
• Equilibrium No energy exchange with environment
• cmp sleeping in the sofa (nothing happens)

12
System characteristics, ctd

• Time invariant, Static, Dynamic
• H? Dog? Typewriter?
• Deterministic , Stokastisc
• Sun rises tomorrow or not?
• Measurable, Controllable, Observable
• Internal states can be reached, observed
• How do you measure an idea?
• How do you control a country?
• Is it possible to observe internal states of a
  dog?

13
System characteristics, ctd

• Distributed, central control
• Internet, Nuclear power station
• Heterogenous, Autonomous, Emergenta
• Autonomous Behaviour determined by experience
• H, dog, PC autonomous?
• Emergent if behaviour of whole not trivially
  follows from behaviour of components
• cmp bil -gt förorter, Arena
• !!! Most of these characteristics can be dressed
  up in mathematics

14
Intelligence

• When? Open question!
• Autonomous, Adaptive, Learn
• Solve problem
• Quickly, complex problem, creative solution
• Creator of memory
• Self awareness
• Complexity of goals and wishes
• Struktural plasticity
• Unpredictable

15
Complexity

• Well known?
• Examplify Open question
• Why is it complex?
• Universe, Internet, H, PC, Tree, Grain of sand,
  Atom
• Complexity estimation? Measures? Open question
• Internet vs a Dog
• Fler än 7 tillstånd/ element/
• Algoritmisk komplexitet
• Feedback, non linear, delay, stochastic, parallel

16
How reduce complexity?

• Open question Group, Sort, Order
• Group by similarity, length, width
• Group by characteristics family, grafical object
• Hierachy, cmp bil (Group by attachment)
• Sort 1,2,3,4, use symmetries (.e.g face)
• Simplifies search
• Sort by functionality (layers)

Vision
All senses
Organisation (worker to executive)
17
Manage complexity

• Group by approximation

• Group in vector spaces (multiple dimensions)

Matematiskt(x, y, z)(1, 1, 1)
18
Gestalt theory

• Tries to find common human behaviour by studying
  how images are interpreted

X X X X O O O O X X X X O O O O
xxx xxx xxx
19
xxx xxx xxx
X X X X O O O O X X X X O O O O
20
(No Transcript)
21
System modeling

• A model is a mapping of a systemor a design onto
  something formed, natural or artificial.

22
Deductions, reasoning,calculations
Model world
Forecasting, comprehension
Abstract model
Abstraction, modelling
Application, evaluation
System world
Anticipation, understanding
Questioning
Observations
Example Weather observations
23
Modelling, Why?

• Simplification of reality
• Understanding, e.g. visualisation
• Specification
• Templatefor system design
• Communication, documentation
• Modelling to reduce complexity

24
Modelling, How?

• No single model fully describes a non trivial
  system,
• H Physiologi, Psychologi, Sociologi,Linguistic
• Type of model affects how you workand the
  solution you get
• Analytic modell (formal, pen paper)
• Numeric model (computer program)
• Clay model, 3D CAD model

25
Modelling, How?

• Different abstraction levels
• Transparens (jfr observability)
• Low -gt too many details (physics for boxing
  match)
• High -gt not enough details (psychology for
  generating speech)
• Phenomenology
• Low transparency, system observed from outside
• HICKs law tklog2(n1) t time to find right
  alternative among n similar
• Webers law dI/Ik, I intensity of stimuli, dI
  change of stimuli 10 for amplitude of sound

26
Models

• Metaphore
• (I) Taste an idea, digest information
• (T) Computer died, life length of battery
• (H) ?
• Internet as a web,

27
Models, ctd

• Structure
• Maps
• OH
• Behaviour
• State machine
• OH
• UML for both structure and behaviour

28
System behaviour described as states
Fill up
Many many states on a detailed level of a
complex system (such as a human)
29
Communikation (Shannon)
Transmitter
Source code
Channel code
Bit
Image
Modulation
Analogue signal
Channel
De Modulation
Receiver
Source decode
Channel decode
30
H as a finite state machine?
Open question
31
Human model processor(Cognitive model)
32
Data flow model of a boxning match!
Interactors, data?
Open question
33
H-H conversation (behaviour)
34
Model of context?
Many dimensions, represent as a spaceOpen
question
Interna tillstånd (Self)
Fysisk omgivning (Environment)
Aktivitet
35
Types of models

• At every abstraction level we have three
  different views
• Intentional (purpose
• Conceptual (function, logic of the system)
• Physical, how it is implemented and affects
  context
• Cmp alarm clock built into bed

36
Types of models, ctd

• Formal, informal
• Stochastic, deterministic
• Deterministiskt Salary (HG)
• Stochastisc Salary (x)
• Text based, graphical
• Executable

37
Types of models, ctd

• Static, dynamic
• model(t),
• How to represent t in simulations (e.g. Sims)?
• Continous, discrete (time)
• Cmp looking in the rear mirror of a car
• Parameterised
• Different kinds of face models, size of mouth

38
Data strukture and algorithm
39
System architecture (strukture)
40
Multimedia system
Out
In
Computer
Computer
SwitchX
A/D
D/A
Storage
41
Protocol (structure, lagered)
42
H
T
5p (Kulturella perpektiv)
5p (Sensorer)
H 10p
T 5p
DoItDesign 22pInt. Tekn. 18p
I 15p
Hum. profil 15p
5p (Datakommunik.)
Tek. profil 30p
I
Entreprenöriell Affärsutveckling 5p
Programmering 10p
Matematik 15p
43
JAVA

• public class site_desciption
• String name
• String url
•  
• public void setName(String s) name s
• public void setURL(String s) url s
• public void print_url()
• System.out.println(name " at " url)
•  

44
Model or real system

• Experiment with a nuclear power station?
• Prototype
• OH prototyper
• Concep telephone-gtBreadboard-gtConsumer test
• Simulaton
• Battle at Ratan
• How to test a mars lander?
• Adapt model to reality
• How to model an abstract idea, e.g. democracy?

45
Protype for examination of mobile phone (Anglo
design)
46
(No Transcript)
47
Models are necessary!
48
User adaptation/modeling (modell of user)