View
224
Download
0
Category
Preview:
Citation preview
4
Overview
Learning objectives
1. Introduction
2. Features of a system
3. Types and typologies of systems
4. The qualities of systems
5. System dynamics
6. Self-organisation
7. Summary
5
Learning objectives
• Define the main features of a system (input, output, transformation, feedback, boundary, environment) and identify them for a given system
• Distinguish between systematic and systemic
• Explain how feedback, regulation and learning can occur in systems
• Identify some ways in which systems have been classified
6
Learning objectives
• Explain the concept of holons and holarchies
• Describe some of the qualities of systems: closed and open, emergence, subsystems
• Explain how systems can be modelled using system dynamics techniques such as simulation
• Differentiate systems thinking from reductionist thinking
7
Introduction
• Any tools for describing the world can fail if they apply only to a static picture of the world
-The world changes, the representation obsolesces and the description fails
• The ideas we use to organise the world and its dynamic relationships are interrelated in a system
• Systems concepts help us to move our picture of the world from a static to a dynamic one
1. Introduction2. Features of a system3. Types and typologies of systems4. The qualities of systems5. System dynamics 6. Self-organisation7. Summary
9
Systems
• In general, systems refer to a purposive organisation of parts and how these behave and interact dynamically through time.
10
Systems theory
• Systems theory or systemics is concerned with the whole system and the qualities that relate to the system’s:
-purposes
-activity
-selfhood
-connectedness
-environment
11
Systems theory
• Systemics has its own literature, theories and approaches and includes such ideas as:
-connectedness-feedback and loops-balance and stability-subsystems and autonomy-behaviour in time-boundary and environment-control- learning and transformation.
12
Systemic and systematic
Systematic – characterised by order and method
Systemic – relating to the whole as a unity
13
Systemics is concerned with
the holistic qualities that
emerge dynamically from the
interactions among
components of a system
Recap
14
Systems
• In systems thinking virtually anything can be seen as a system …
1. Introduction2. Features of a system3. Types and typologies of systems4. The qualities of systems5. System dynamics 6. Self-organisation7. Summary
16
Systems and informatics
• In informatics, business analysts look at the patterns of activity and work processes in an organisation and see how information flows to the different parts, or how it might be directed to flow. They see an organisation as a system.
• Identifying the parts from the viewpoint of the whole system and seeing ways in which they are or could be related allows design choices to be made.
17
Features of a system
• All systems have certain attributes and the presence of these allows something to be defined or seen as a system.
• A system exists within an environment, or within a larger system. Systems generally operate within the larger environment, interacting with it.
• A system usually will take something in from outside, process it in some way and produce something that goes back to the outside again, usually in a different form.
18
Inputs
• Inputs are what we can identify as going into a system. These are typically materials or data.
• Everything within a system was either present there from the start, or arrived via the inputs. Logically these are the only sources of anything found in a system – nothing new can be created within a system boundary.
• In information systems, data inputs come from sources.
21
Transformations
• Nothing new can be created within a system boundary, although transformations within the system are possible.
• Most systems, particularly organisational, have the purpose of transformation of some sort.
• The diagrams show simple transformation processes. Normally a lot more detail of how this gets done would be modelled for a system.
24
Outputs
• Outputs are what we perceive to be leaving the system.
• Logically everything that leaves a system leaves via the outputs and as nothing new can be created, the output is some transformation of the input.
• Outputs go to sinks
25
Outputs
A more complete model might have ‘wheat husks and chaff ’ as by-products, which would require another output box to be drawn. If this exhausted all the input materials, the diagram would then be complete at an overview level of description.
29
Feedback
• In some systems all or part of the output from a system can be re-routed to the inputs and create a cycle. Adjusting the feedback to a system is the standard way of controlling it.
• Feedback is generally either positive or negative, reinforcing or inhibiting behaviour.
30
Positive feedback
• Positive feedback encourages more of the same behaviour – it amplifies its inputs, which, in a cyclic loop, can escalate out of control.
-A snowball rolling down a hill may gather more snow, the more snow it gathers the bigger it gets and the bigger it gets the more snow it gathers.
31
Positive feedback – Virtuous circle• An example of positive feedback is the virtuous circle,
where a pattern of behaviour reinforces itself leading to desirable consequences.
• A loop of linked causes can be identified and this can be modelled using a causal loop diagram
32 32
Positive feedback – Virtuous circle
• A plus means a change in the same direction (more enjoyment means more swimming).
• Overall the loop is changing in the same direction – positively reinforcing the pattern. This is shown by a plus sign in the middle of the diagram.
33 33
Positive feedback – Vicious circle
• The pattern can also be undesirable – in which case it is called a vicious circle
• A minus means a change in the opposite direction (less friends means more depressed).
• There is an even number of minus signs, so they cancel out, and the overall pattern is positively reinforcing.
34
Negative feedback
• Negative feedback discourages the behaviour that led to it.
- In many countries there are penalties for exceeding the speed limit. Therefore, drivers often watch their speedometers as they drive through speed limit zones, easing the pressure on the accelerator as the car starts to go too fast and increasing pressure if it slows down too much.
35
Negative feedback
• In a thermostat, more heat causes more room temperature, that causes the thermostat to heat up too, which at a certain point acts to reduce the heat and therefore room temperature.
• There is an odd number of minus signs, so the loop overall is a negative feedback or balancing loop, shown by the minus sign in the middle.
36
Negative feedback
• Negative feedback is often used to maintain equilibrium, a balanced state which is more or less constant and optimal
- Artificial systems, such as cruise control or room thermostats, can be designed to work on this principle to maintain a desirable equilibrium.
37
Feedback is re-routing some of the
output of a system to its input,
creating a cycle that enables
regulation.
Feedback can be positive (reinforcing
the same behaviour) or negative
(inhibiting the behaviour)
Recap
38
Regulation and learning - homeostasis
• When a system maintains itself through a feedback mechanism, it is called homeostasis.
• Homeostasis is simply the idea that a system will try to stay stably balanced, at equilibrium
-A thermostat keeps room temperature within a balanced range.
• A steady state is also constant over time and may or may not be at equilibrium.
• These ideas are applied in economics, health, astronomy and many other fields, including to the Earth itself
39 39
Homeostasis and the human body
• Homeostasis is one of the most fundamental concepts in health care: it is a principle that the human body will try to maintain homeostasis:
- If our energy is low, we eat.
- If our temperature is cold, we shiver.
- If we have been poisoned, we may develop a fever.
• Although a healthy body is more or less at equilibrium, dynamic processes that maintain that state are occurring all the time, without our conscious intervention.
40
Systems interacting with other systems – Coupling
• Coupling is where systems are interdependent through the output of one system becoming the input of another
- Tightly coupled
-where the output of one system is directly linked to the input of another.
- Loosely coupled
-where the link may not be direct or complete, or may be able to be overridden.
41
Tightly coupled and loosely coupled systems
• In a car, the output of the accelerator is directly linked to the input of the tachometer – tightly coupled
• The input for the accelerator depends on human judgment, which is informed by the tachometer’s output. These systems are more loosely coupled, there is less direct dependency and there is a human in the loop, allowing discretion to be exercised.
• With cruise control, the system is fully automated, though it can be overridden.
- The distinction between fully automated and partly automated systems is important in information systems design
42
Regulation through feedback is
called homeostasis, and works
to keep the system at
equilibrium.
Recap
43
Regulation and learning
• When a system uses feedback to adapt to a new situation, it is a type of learning
• Feedback is used to reduce or lessen the difference between the actual state and the intended state to which the system is purposefully orienting itself
• This can be modelled with more complex systems of loops
• Machines that can learn in this way have been around for many years now
44
Deutero-learning (Bateson)
• Learning by generalising from primary experience can be modelled by a simple feedback loop.
- Avoiding obstacles or steering a car between the edges of its traffic lane uses feedback from visual cues to allow error correction. Once we can do this, we can drive on roads we have not seen before.
• Deutero-learning contextualises this using a second feedback loop.
- We are aware of ourselves driving and knowing how to drive allows us to steer off the road, perhaps to avoid causing an accident.
• The ideas of deutero-learning, or ‘learning to learn’, have been applied in many different fields, including the theory of learning in organisations
45
Double loop learning (Argyris and Schön)
• Single loop learning
-When painting a factory wall, feedback from the painter or supervisor ensures it is error free, efficiently done and completed to a schedule.
-The feedback loop involves only action and monitoring its consequences.
• Double loop learning
-The whole activity of painting walls might be seen as a waste of resources ,so that system gets eliminated, replaced with an easily washed laminate wall or open brickwork.
-The whole system loop itself is monitored and feedback applies to that system as a whole.
46
Single and double loop learning
• The difference is between simple detection and correction of errors within some task or procedure (single loop learning), and questioning the procedure itself, particularly in relation to the organisation’s goals and objectives (double loop learning)
Deutero-learning – “learning to learn” by using a second feedback loop to contextualise the feedback from the first level loop.
Single loop learning – when a feedback loop involves only monitoring the consequences of an
action.
Double loop learning – when a whole system loop is itself monitored and feedback applied to the system
as a whole.
48
Self-fulfilling prophecies
• Some patterns of behaviour are self-reinforcing but dysfunctional
• Starting with a small misunderstanding, things can escalate into a situation both parties believe is actually the case
• Moreover, since each acts as if it were the case, it becomes true in the experience of each
49 49
viewI just know she doesn’t
like me
I’m not going to answer her
call if she feels like that
about me
How rude!!! He never returns my
calls
I’m not inviting him to the BBQ
then
50
A system can learn through using
feedback to adapt to a new
situation.
Higher-order loops allow
monitoring and reflection on the
whole system
Recap
51
Cybernetics and systems
• Many systems concepts have been developed in the systemics field of cybernetics.
• Cybernetics differs from conventional sciences by its focus on communication and control relationships in a system and regulating or steering the system on course to its goal.
• Cybernetics is often associated with control mechanisms
• However Parsegian argues that its principles and concepts also apply in understanding social and human interests
52
Purpose and trajectory
• Most systems have a purpose and informatics systems are generally goal or function oriented.
• These functions usually imply that some definite input occurs at a given moment in time and a definite output occurs at a second moment in time, e.g.
-A ‘request for payment’ comes in at moment 1
-The system then (typically) checks that ‘goods have been received’ and ‘records are kept on file’
-Then a payment authorisation is issued at moment 2.
-Alternatively, a denial of payment may be issued at moment 2, if it turns out that the records don’t tally.
53
Purpose and trajectory
• The calendar refers to the moments, intervals or time steps where a state of the system is noted.
• The path between the two moments is called the trajectory
• These moments need not be linked to actual dates and times, but instead show an ordering in time of relevant events and states. These are generally in line with the purpose or a goal of the system, such as to transform payment requests into resolved payments.
• Transition between states is a concept that can be applied very widely
54
State transition diagrams
• Used to represent the states and transitions in a system
• Various systems modelling languages have versions and extensions of state transition diagrams
- Graphical forms, matrix forms
• Commonly used in business modelling and associated software design, such as Unified Modelling Language (UML)
Informatics 101
Informatics 201
Informatics 301
Graduation
States in the trajectory of a three-year degree course
States Transitions
56 56
A three-year degree• A student’s progress through a three-year degree course may not
take three years:- May be faster if summer semesters are taken
- May be slower if subjects are repeated, or a gap year is taken.
• Modelling is better done in terms of the states along the trajectory, aiming at a completed degree
- First level studies
- Second level studies
- Third level studies,
- Graduated
• Entry to each state requires the student to have passed the previous state (eg. Informatics 101 is a prerequisite for Informatics 201) so the states must be passed through in sequential order as their requirement thresholds are met.
57 57
Equifinality and multifinality
• Many purposes can be achieved in different ways:
- If you are too hot you can swim, take a cold shower or put on air conditioning or a fan.
• The end result is the same – equifinality
• The opposite is multifinality where similar initial conditions lead to different outcomes.
-For example one child in a family may become a responsible citizen but their sibling becomes a bit wild, despite the same home environment
58
Multifinality – a situation where similar initial conditions lead to different outcomes.
Equifinality – the principle that the same given end state can be reached by various means.
59
Boundary and environment
• Systems are identified by their boundary and exist in an environment, which, by definition, lies outside the boundary
• The system interfaces with its environment at its boundary. This is the point at which exchange takes place into, or out of the system
60
Boundary and environment
Boundary of system Wheat and sugar are within environment of system, but outside boundary
61
Boundary and environment
Interface point for receiving wheat
Interface point for outbound biscuits
Interface point for receiving sugar
62
Input, transformation and output; boundary
and environment, interface and (optionally)
feedback, purpose and event or state
histories are the main general features of
any system.
The essential quality of a system, however,
lies in its function as a whole.
Recap
63
Types and typologies of systems
• There are many definitions of ‘system’, emphasising different aspects. The definition we use here is:
-A group of interacting, interrelated, or interdependent elements or parts that function together as a whole to accomplish a goal
• There have also been many attempts to categorise different types of systems:
-Jordan’s taxonomy
-Bánáthy’s systems continua
-Checkland’s systems map
-Boulding and Minger’s taxonomy
1. Introduction2. Features of a system3. Types and typologies of systems4. The qualities of systems5. System dynamics 6. Self-organisation7. Summary
64
Jordan’s taxonomy of systems
• Uses three pairs of constructs:-structural or functional
-Whether the static state is of most interest or its dynamism in time is of most interest
-purposive or non-purposive
- the system’s existence or action serves a purpose, or not
-mechanistic or organismic
-whether or not a component can be simply replaced without affecting the rest
• These three dimensions give eight types -Jordan was able to fit 15 different definitions of ‘system’ into one or other type
66
Bánáthy’s systems continua• Béla Bánáthy identified five dimensions along which systems
differ as part of a bigger map of system types
- This gives similar (but richer) space than Jordan’s
• The left hand side qualities together produce more rigidly controlled systems (for example, a clock), while those on the right tend to be creative and ideal-seeking (for example, some communes), with a range of types in between.
67
Checkland’s systems map
• Peter Checkland proposed ‘a systems map of the universe’ which includes five classes of system:
- Natural systems
- Designed physical systems
- Designed abstract systems
- Human activity systems
- Transcendental systems
68
Checkland’s systems mapatoms, weather systems, star systems, rain forests, bacterial colonies…
buildings, bridges, racing cars, traffic light systems, space stations, gyroscopes…
mathematics, poetry, literature and the language Esperanto…
a family, the army, a city or a political system…
70
Systems can be typified in different ways:
for example by whether they are natural or
relate to human activity, by their complexity, or
by characteristics such as the degree to which
they are closed or open, mechanistic or
organismic, and so on.
Typologies include those by Jordan, Balanthy,
Checkland and Boulding/Mingers.
Recap
71
The qualities of systems
• All systems are wholes in some sense – the organised parts that make them up function with an integrity that makes them a useful level to identify by name and to build an understanding around
• Systems are themselves often parts of bigger systems, and have subsystems of their own
• At each level the system is an integral whole, or holon, within a holarchy.
1. Introduction2. Features of a system3. Types and typologies of systems4. The qualities of systems5. System dynamics 6. Self-organisation7. Summary
73
Holarchy or hierarchy?
• In a holarchy each level contains parts that are whole systems in their own right and each part is part of a greater whole, in a hierarchical structure.
• The understanding of the word hierarchy has become corrupted over time to have, for some, a connotative meaning of authoritarian or bureaucratic power structures.
• But true hierarchies consist of increasing orders of wholeness: and the philosopher Ken Wilber considers, along with Koestler, that the correct word for hierarchy is holarchy, using them interchangeably.
76
Holarchies
• Cells, organs, bodily systems of individual humans, contagion in a population and pandemics may all be appropriate levels at which to consider disease.
• Some activity at a given level can impact on the rest of the structure, through connectedness.
- Poisoning a tissue may make an organ unviable, killing an organism.
• Levels have unique properties and applicable concepts of their own, which do not apply to other levels, but which can nonetheless result from the activity at other levels.
- The ugly red skin rash from poison-ivy can be reduced by an ingested drug.
• Activity at the level of the part has an effect on a property of the whole
- A pandemic may be averted by immunisation of individuals
77
Holarchies
• A fuel filtering system is part of a propulsion system is part of a submarine is part of a military deployment
• While it is sensible to talk of the efficiency of the fuel filter and the quietness of the submarine, it is less sensible to talk of the quietness of the fuel filter and the filtering efficiency of the submarine.
• The success of a military deployment may emerge thanks to the efficiency of a propulsion system.
78
Connectedness and relationship
• Many systems ideas are probably familiar from the model of an ecology or ecosystem. An ecosystem is basically a community of interacting organisms and its environment.
• Ecology has supplied many ideas to informatics, such as
- Interdependent webs (of organisations)
- Niche (niche marketing)
- Symbiosis
- Perturbation
Birds eat insects
Birds fertilise plants Rabbits eat
plants
Foxes eat rabbits
Insects eat dead foxes
80
Niche
• In ecology, a niche is a place within the grander scheme that is especially suited to an organism or species
• Organisations use the concept in the idea of niche marketing where they provide a specific product or service to consumers with those specialised tastes
81
Symbiosis
• Symbiosis is where two different organisms work together for mutual benefit
• Organisations can work symbiotically
- The waste chaff from the biscuit factory might be recycled into garden mulch by another organisation
82
Perturbation
• Where some disruption is introduced into an ecology or system, changing the dynamics of the system
• Systems are interconnected, so a change in one part of a system can affect the whole system
-Pesticides targeting insects go through the food chain, increasing concentration, to threaten the survival of bird species, or ultimately human health. This may solve an immediate agri-business problem, but cause an environmental one
• Understanding the impacts of change and how to manage them is a key skill in informatics
83
Closed and open systems
• Systems can be closed or open:
- Closed system
-does not exchange materials or information with its environment
- Open system
- interacts with its environment, receiving inputs and producing outputs
• Usually parts of a system can be considered as closed for practical purposes and other parts can be considered as open
84
Closed systems• A closed system does not exchange materials or information with
its environment• Examples of closed systems:
- The Earth’s water cycle
- Inside a spaceship
- The water in the pipework of a central heating or a cooling system
- The coolant in a car engine
- The blood circulating in your body
- The formal languages of mathematics and computing
- The numbers in a Sudoku puzzle
• Very few systems, other than formal ones, are completely closed • Usually they operate in the context of other systems, with which
they interact in some way
86
Open systems
• Open systems are those that interact with their environment. They receive inputs and produce outputs.
• Organisations are a good example of open systems: they take inputs and do something with them, to produce outputs.
- The biscuit-making organisation takes wheat as input and produces biscuits as output.
- In between it has processes that transform wheat (and other inputs) into biscuits, which can then become an input to a biscuit-retailing system.
• Organisations thus transform the inputs, usually to add value to them, or to make them useful as inputs to another system.
87
Time and change
• Time is essential for an understanding of change and the dynamics of change
• Delay in systems or lag time refers to how quickly information is provided after something happens
• Informatics often tries to reduce or eliminate lag time between the world and our picture of it, to ensure the picture remains accurate
89
Delay and negation systems
• Delay systems have no transformation, just reproduce inputs directly as outputs one time unit later
• Negation systems have outputs that are opposite to their inputs
• These systems can be used as memory devices for storing information
90
Stasis
• The converse of change is stasis, a state where nothing seems to happen outwardly, though there may be a dynamic equilibrium underlying it
• Dynamic equilibrium is like a tug of war between equally matched teams, there is no directional movement overall, but they are individually expending energy!
91
Recursion
• in systemics, refers to the phenomenon that every viable system contains, and is contained, within a viable system.
• More generally in informatics, recursion means applying the same process self-referentially until some limit or terminator is reached.
Recursion: see recursion
92
Emergence
• The counterpart to the concept of holarchies is the notion of emergence.
• Although systems may sometimes be broken down into parts, they have the properties of organised wholes. These wholes have properties that are not the properties of their parts, nor are they simply derived from them. Such properties are known as emergent properties and result from the quality of the system design
• Emergence often occurs through interactions between parts, to produce something that was not there before.
94
Emergence• A racing car is made up of, among other things, wheels, an engine,
a fuel system and perhaps red paint.
- None of these components is ‘fast’ or especially ‘attractive’.
• Put together in the appropriate way, though, the car as a whole is fast, attractive and aerodynamic. These are emergent properties of the whole, not of the parts and the car, as a whole system, defines the parts and their role in the design.
- The red paint is functionally no different from pink paint, but some believe red cars look as though they are faster. In systems thinking the whole is greater than the sum of its parts.
95
Components and subsystems
• Much of information systems development uses components that can be assembled in various ways and reused, like Lego™ bricks or tangram pieces
• Often a good way to model systems involves identifying subsystems, which can be looked at as replaceable components
• When subsystems can be isolated like this, it allows for different combinations to be tested by replacing parts of the system and observing performance
96
Black box and white box testing
• Both black box and white box testing are used in informatics, particularly software testing
• In black box testing you provide an input and get a functional output, but don’t know what goes on inside the thing being tested – if the required outcome is achieved that’s all that matters
• In white box testing the internal functioning is known and explicitly tested.
97 97
Systems thinking
• In systems thinking, the interactions, as much as the quality of the components themselves, are a key feature of the systems view.
• When components can act upon one another, as humans do, new outcomes can emerge and these may not be computable or simply predictable
• It is therefore an art (or a special type of science) to design a netball team (or a racing car/driver combination) from whose interactions particular outcomes emerge, some of which are unpredictable from just analysing the properties of the components.
98
Systems thinking• Analytic or reductionist thinking concentrates on breaking
things down into parts -But a disassembled racing car misses its essence!
• Systems thinking is much more interested in the qualities of the whole racing car –
-speed, aerodynamics, safety, racing fast without falling apart…
-And in this the concepts and activities of interaction, purpose, function and emergence are needed to complement the analysis of components and structure.
• A system’s behaviour in its environment must also be considered
- If the racetrack is wet, different tyres are needed.
-Adjustments are made to the car depending on who is driving
99
Systems thinking
• Whereas classical, discipline-specific science “tried to isolate the elements of the observed universe … now we have learned that for an understanding not only the elements but their interrelations as well are required”(Ludwig von Bertalanffy)
• Bertalanffy was an early systems researcher who investigated the principles that are true for systems in general, whether these are biological systems, social, engineered or other.
• His work was described as General Systems Theory and it provides one of the foundations for contemporary system sciences, or systemics.
100
Some ideas in systems thinking include
holons and holarchies,
connectedness and relationships,
open and closed systems,
time and change,
emergence,
and subsystems
Recap
101
System dynamics
• The systems tradition provides concepts relevant to theorising about processes, functions, changes, control mechanisms and other dynamic aspects of systems as they act, adapt and transform in time
• System dynamics is a modelling approach that allows prediction of the states of a system and given points in time, and whether it is tended towards stable equilibrium or catastrophe
1. Introduction2. Features of a system3. Types and typologies of systems4. The qualities of systems5. System dynamics 6. Self-organisation7. Summary
102
Rabbits and foxes + increase in one causes an increase in the other- increase in one causes a decrease in the other
(–) negative feedback loop
103
Rabbits and foxes
• The causal loop model shows several negative feedback loops (-)
• Thus the populations of both rabbits and foxes is under control
-An overpopulation of rabbits impacts on their food supply, which in turn means a decrease in the birth rate
• Forrester extended these causal loop systems to include quantities and time
• Systems dynamics can then simulate what happens to the stocks of rabbits and foxes over a period of time
Possible initial assumptions for the rabbits and foxes model (top) and possible output from a simulation model using these assumptions
105
Dynamical systems
• Tipping points are phenomena of crowd behaviour when at a point in time, their behaviour shifts from one state to another
-Sudoku becomes an international craze
-A few people have a disease; suddenly there is an epidemic
-Suddenly everyone seems to be saying “awesome”
-The last straw breaks the camel’s back
• Below a certain threshold nothing changes, but only a tiny change in the input then leads to the behaviour switching catastrophically
106
The fifth discipline and the learning organisation• Peter Senge and others call systems thinking ‘The Fifth
Discipline’, and have applied it to organisational settings
• Learning effective ways to interact and avoiding dysfunctional patterns of behaviour is a motivation for organisational learning
- It can remember its experience, reflect upon it, learn lessons and pass these on, avoiding making the same mistakes over again.
- It can make decisions in situations similar to previous ones more efficiently and effectively, process information more appropriately and, as a group, make sense of its activity.
107
System model archetypes
• Senge identified several commonly occurring system model patterns or archetypes relevant to organisational settings
108
Tragedy of the commons• Initially, the traditional
influences of drought and disease kept numbers of cattle under control in the Sahel
• When deeper wells and modern medicine were introduced, however, there were fewer deaths, expansion of herds – and overgrazing and desertification
After Senge
110
Self-organisation
• Many systems adapt intelligently to their environments. Others may select their actions to change towards an internal equilibrium. This can mean changing how their parts are organised to be closer to where they want to be, a bit like trying to find a comfortable sleeping position when camping.
• The original principle of self-organising systems is due to W. Ross Ashby, a major figure in systemics, who originated many of the field’s core ideas.
1. Introduction2. Features of a system3. Types and typologies of systems4. The qualities of systems5. System dynamics 6. Self-organisation7. Summary
111 111
Self-organisation
• Self-organising systems display the intelligence to reorganise their components into forms that allow them to adapt to changes in the environment
-Artificial neural networks can change the weighting within the system as they learn and incorporate feedback
-The human brain is plastic and its constructions of reality are adaptable
-Insects can rapidly generate new forms of body more suited to changed circumstances
112
Self-organisation
• One of the requirements that helps a system to reorganise itself is a mutual awareness among components, which have some autonomy over their activity
• Traffic flows are a good example of how a self-organising system can work
- Every day in our cities we see traffic moving along, more or less at the speed of surrounding traffic and usually without crashing into other vehicles.
113
Variety and complexity
• The idea of variety is important in systems and is due to Ross Ashby.
• This refers to the number of different possible states a system can be in and as such gives a measure for its complexity.
• Ashby’s law of requisite variety states that the system for managing or controlling a situation should have as many response actions available as possible states that can arise in that situation.
• Because the number of possible states to manage can escalate to huge numbers in practice managers will try to reduce these somehow.
114
Self-organising systems can
reorganise their components and
adapt to changes in the
environment
Recap
115 115
Strategy and management cybernetics• A strand of cybernetics work has always been
concerned with the societal and the organisational, and this body of knowledge now informs theories of management in complex situations, such as large organisations and even countries.
• Because they provide the theories and tools to deal with complexity and systems of systems, such approaches can support information requirements and strategic decision-making at very high levels.
116
Strategy and management cybernetics
• Some examples:
-The viable systems model-StratAchieve-Variety and complexity
117
The viable systems model
• Developed by one of the greatest of systems thinkers, Stafford Beer, founder of the field of management cybernetics
• The model addresses the issue of what a system must be to be capable of independent existence, that is, viable in its environment.
• Like the holarchy, it rests on recursion such that every viable system contains and is contained, within a viable system. It does not require an authoritarian order however; being concerned with systemic order not management politics.
• Its most famous application to date was in successfully designing the organisation of the Chilean economy as an integrated information system.
118
StratAchieve
• Strategic planning method that integrates:
• ‘Hard’ aspects-resource allocation, performance monitoring, activity based-costing
• ‘Soft’ aspects -related to general but vague mission statements, different viewpoints and feelings about different courses of action
• Iteration between planning and project management aspects
• Strategy development is holistic and the strategic rationale for particular activities is explicitly explored.
119
Managing variety and complexity
• Management is a challenge because there are so many ways that subordinates can generate information and complexity.
• Variety can quickly escalate and the possible number of different states can become astronomically (if not meaninglessly) large as a new element is introduced to a system.
• Raul Espejo suggests putting self-regulation into the situation, leaving only residual complexity to manage
Number of members Number of interpersonal relationships
Number of possible relationships
1 1
3 6
6 25
10 90
15 301
21 966
124
Summary
• Systemics is concerned with the holistic qualities that emerge dynamically from the interactions among components of a system
• Systems take in inputs from their environment, and transform them into outputs
• Systems are identified by their boundary and exist in an environment
• Feedback loops in systems permit regulation and control, and higher order loops allow learning
• Systems are themselves often parts of bigger systems, and have subsystems of their own
125
Summary
• There are different types and classifications of systems, but all share certain features
• System dynamics techniques such as simulation can be used to model systems
• Self-organising systems can reorganise their components and adapt to changes in the environment
• Management cybernetics has informed theories of management in complex situations
Recommended