DevEX - reference for building teams, processes, and platforms
A Unified Framework for Collective Systems
1. A UNIFIED
FRAMEWORK FOR
COLLECTIVE SYSTEMS
Emma Hart, Edinburgh Napier University
Jeremy Pitt, Imperial College London
Ulle Endriss, University of Amsterdam
2. Grand Vi i
G d Vision
Applications
A Software Toolkit of
Design Patterns and
Components
p
A Unified Theory of
Operations for CAST
Systems
y
3. Why do
Wh d we need a new theory ?
d
h
• Existing engineering
approaches provide some
theoretical basis
• E.g. control theory –
ensure/prove stability
• But most methods don’t
account for defining
t f d fi i
properties of CAST systems
• Lead to systems that are
oscillatory or at worst unfit for
purpose
• Existing methods often
domain-driven (e g
(e.g.
telecoms, robotics)
• Not generalisable or
transferable
4. CAS i M l i Di i li
is Multi-Disciplinary
• Many theories from
individual disciplines
• Hard to compare
theories
• Theories address
different aspects of
CAS
• Don’t account for
Don t
engineering
constraints
5. Towards a unified theory
T
d
ifi d h
• Unifies concepts from
multiple disciplines into
a single framework
• Qualitative theory
represented in
a o at c o
axiomatic form
• Can be formalised and
analysed
• Operationalised via
design patterns
Biological
Systems
Computational
Social Choice
Organisational
Theory
6. Biological Systems
Bi l i l S
• Immune-neuro-
• Long-term stability
endocrine mechanisms
lead to homeostasis
• Cohen’s cognitive
immune system :
• Adapt over multiple
• Decision making via co-
respondence
• Swarm insects
• Coordination, partial info
• Symbiosis between
multiple species:
• C
Cooperation
ti
timescales
ti
l
• Coordinate multiple
heterogeneous
components
• Deal with limited and
partial information
• Decision making
• Conflict resolution
7. Social Choice Theory
S i l Ch i Th
• Originates in economics
and political science
• Concerns design &
analysis of methods for
aggregating preferences of
multiple agents into
collective decisions
• Social choice considers
formal aspects of
democratic decision
making (e.g electoral
systems)
t
)
• Computational Social
g
Choice add an algorithmic
perspective
•
•
•
•
•
•
Heterogeneous agents
Multiple objectives
Collective decisions
Open-ness
Fair division of resources
Stability
8. Organisational Th
O
i i
l Theory
• Elucidates principles
for stable resource
management
• Study of engineered
systems
• Insights into
engineering sociotechnical
‘organisations’ in a top
down manner
• Collective Action
• Trust
• Cooperation
• Stable and enduring
g
systems
9. A Unified Th
U ifi d Theory of O
f Operation
i
New CAST propertie
w
es…
Con
nflicts
Engineering Requirements of CAST Systems
D
Diverse Objective
O
es
Organisational
Theory
S
Social Interaction
ns
Biological
Systems
Noisy Inf
formation
n
Computational
Social Choice
Open
n-ness
A Unified Theory of Operations for CAST Systems
10. What d
Wh does synthesis give ?
h i i
Biological
g
Systems
Computational
Social Choice
Biological
Biological
Systems
Organisational
Theory
Engineering
Constraints
Computational
p
Social Choice
>>
Engineering
Constraints
• Addresses weaknesses in individual theories
• Addresses conflicts
• Respects engineering constraints
Organisational
Theory
11. Individual Weaknesses
I di id l W k
• Biological Systems:
• Tend to rely on homogeneous collectives
• Global rather than individual objectives
• Considerable physical differences
• Computational Social Choice
• Based on standard models from economics
• Abstracted from human decision making (different goals but same
model)
• Institutional Theories
• Easy to get locked into sub-optimal states due to path
dependencies
• Not clear how to evaluate ‘fitness’ of an institution
fitness
12. Conclusions
C
l i
• Unification addresses current fragmented approach to
inter-disciplinary research
• Diff
Different analysis t l currently hi d elucidating
t
l i tools
tl hinder l id ti
connections between fields
• Many existing theories don’t account for engineering
don t
constraints of CAS
• A unified theory will:
• Enable formal comparison between concepts from different
disciplines
• Drive innovation in field