A Multipatch Metapopulation Model with
“Human” Agents Injected
Bruce Edmonds
Centre for Policy Modelling
Manchester Metropolitan University
A Multipatch Metapopulation Model with “Human” Agents Injected. Manchester Complexity Seminar, 18th Oct 2013. slide 1
Motivation
Combined
Complex
Complex
Individual-based
Simple Model
Model
System-dynamics
Model
Simple
System-dynamics
Model Complex
Individual-based
Model
Socio-Ecological
Social Model
Model
Ecological Model
A Multipatch Metapopulation Model with “Human” Agents Injected. Manchester Complexity Seminar, 18th Oct 2013. slide 2
Model of Ecology
Each
A well-mixed individual
represented
patch
separately
Slow
random rate
of migration
between
• A wrapped 2D grid of
patches
well-mixed patches with:
– energy (transient)
– bit string of characteristics
• Organisms represented
individually with its own
characteristics,
including:
–
–
–
–
bit string of characteristics
energy
position
stats recorders
A Multipatch Metapopulation Model with “Human” Agents Injected. Manchester Complexity Seminar, 18th Oct 2013. slide 3
How Dominance is Decided
A Multipatch Metapopulation Model with “Human” Agents Injected. Manchester Complexity Seminar, 18th Oct 2013. slide 4
Model sequence each simulation tick
1.
2.
3.
4.
5.
6.
7.
8.
9.
Input energy equally divided between patches.
Death. A life tax is subtracted, some die, age incremented
Initial seeding. until a viable is established, random new individual
Energy extraction from patch. energy divided among the
individuals there with positive score when its bit-string is evaluated
against patch
Predation. each individual is randomly paired with a number of
others on the patch, if dominate them, get a % of their energy, other
removed
Maximum Store. energy above a maximum level is discarded.
Birth. Those with energy > “reproduce-level” gives birth to a new
entity with the same bit-string as itself, with a probability of mutation,
Child has an energy of 1, taken from the parent.
Migration. randomly individuals move to one of 4 neighbours
Statistics. Various statistics are calculated.
A Multipatch Metapopulation Model with “Human” Agents Injected. Manchester Complexity Seminar, 18th Oct 2013. slide 5
Key advantages of this
• Produces a complex space of changing food webs
• Model creates endogenous shocks and unpredictable
adaptions (to, for example, man)
• Species arms-races, predator-prey dynamics, waves
of species spread etc. all exhibited
• Levels of environmental variety, size, granularity etc.
are determined by paramters
• Mutation and migration happen in parallel
• “Equilibrium” almost never observed
• Diversity of ecologies are emergent outcomes
• “Human” agents can be embedded in food chains and
the combined complexity of this explored
A Multipatch Metapopulation Model with “Human” Agents Injected. Manchester Complexity Seminar, 18th Oct 2013. slide 6
Non-Viable Ecology
Typical Non-Viable Ecology the world state (left) Number
of Species (centre) Log, 1 + Number of Individuals at
each trophic level (right)
A Multipatch Metapopulation Model with “Human” Agents Injected. Manchester Complexity Seminar, 18th Oct 2013. slide 7
Dominant Species Ecology
Typical Dominant Species Ecology the world state (left)
Number of Species (centre) Log, 1 + Number of
Individuals at each trophic level (right)
A Multipatch Metapopulation Model with “Human” Agents Injected. Manchester Complexity Seminar, 18th Oct 2013. slide 8
Rich Plant Ecology
Typical Rich Plant Ecology the world state (left) Number of
Species (centre) Log, 1 + Number of Individuals at each
trophic level (right)
A Multipatch Metapopulation Model with “Human” Agents Injected. Manchester Complexity Seminar, 18th Oct 2013. slide 9
Mixed Ecology
Typical Mixed Ecology the world state (left) Number of
Species (centre) Log, 1 + Number of Individuals at each
trophic level (right)
A Multipatch Metapopulation Model with “Human” Agents Injected. Manchester Complexity Seminar, 18th Oct 2013. slide 10
Longer-Term Trends in Num. Species
A Multipatch Metapopulation Model with “Human” Agents Injected. Manchester Complexity Seminar, 18th Oct 2013. slide 11
Neutral patches, random migration, plants only
• As predicted by
Hubble’s
“Neutral
Theory”
• “skewed
s-shaped”
relative species
abundance
curve
• “Multinomial
distribution” of
log2 species
distribution
• Except, maybe,
species-area
scatter chart
A Multipatch Metapopulation Model with “Human” Agents Injected. Manchester Complexity Seminar, 18th Oct 2013. slide 12
Neutral Patches, local migration, plants only
• unchanged
“skewed
s-shaped”
relative species
abundance
curve
• “Multinomial
distribution” of
log2 species
distribution but
with more
lower
abundance
species
• Pretty flat
species-area
scatter plot
A Multipatch Metapopulation Model with “Human” Agents Injected. Manchester Complexity Seminar, 18th Oct 2013. slide 13
Selective patches, random migration, plants only
• “Linear” drop
in log 2
species
abundance
• “Exponential”
shaped
relative
species
abundance
graph
• Flatter
species area
scatter plot
A Multipatch Metapopulation Model with “Human” Agents Injected. Manchester Complexity Seminar, 18th Oct 2013. slide 14
Neutral Patches, random migration, predators
• Lots of new
species plus
flatter log 2
species
distribution
• flat or curved
species
abundance
distribution
• a variety of
species-area
curves
A Multipatch Metapopulation Model with “Human” Agents Injected. Manchester Complexity Seminar, 18th Oct 2013. slide 15
Other combinations
A Multipatch Metapopulation Model with “Human” Agents Injected. Manchester Complexity Seminar, 18th Oct 2013. slide 16
Adding “Human” Agents
• The agents representing humans are “injected” (as a
group) into the simulation once an ecology of other
species has had time to evolve
• The state of the ecology is then evaluated some time
later or over a period of time
• These agents are the same as other individual in
most respects, including predation but “humans”:
– can change their bit-string of skills by imitating others on the
same patch (who are doing better than them)
– might have a higher “innovation” rate than mutation
– might share excess food with others around
– might have different migration rates etc.
• Could have many other learning, reasoning abilities
A Multipatch Metapopulation Model with “Human” Agents Injected. Manchester Complexity Seminar, 18th Oct 2013. slide 17
Extinction due to Consuming all Others
A Multipatch Metapopulation Model with “Human” Agents Injected. Manchester Complexity Seminar, 18th Oct 2013. slide 18
Waves of (Human) Predator-Prey Patterns
A Multipatch Metapopulation Model with “Human” Agents Injected. Manchester Complexity Seminar, 18th Oct 2013. slide 19
Example run with: selective environment, local
migration, predators, and “humans” @ tick 2000
@tick2000 – before “humans”
@tick5000 – after “humans”
A Multipatch Metapopulation Model with “Human” Agents Injected. Manchester Complexity Seminar, 18th Oct 2013. slide 20
Some of the dynamics of the example run
A Multipatch Metapopulation Model with “Human” Agents Injected. Manchester Complexity Seminar, 18th Oct 2013. slide 21
Effect of humans vs. environmental complexity
proportion of ecology types,
red=plant, blue=mixed,
purple=single species,
green=non-viable
diversity of ecology,
blue=with humans,
red=without
A Multipatch Metapopulation Model with “Human” Agents Injected. Manchester Complexity Seminar, 18th Oct 2013. slide 22
Effect of humans vs. general migration rate
proportion of ecology types,
red=plant, blue=mixed,
purple=single species,
green=non-viable
diversity of ecology,
blue=with humans,
red=without
A Multipatch Metapopulation Model with “Human” Agents Injected. Manchester Complexity Seminar, 18th Oct 2013. slide 23
Effect of humans vs. food input to world
proportion of ecology types,
red=plant, blue=mixed,
purple=single species,
green=non-viable
diversity of ecology,
blue=with humans,
red=without
A Multipatch Metapopulation Model with “Human” Agents Injected. Manchester Complexity Seminar, 18th Oct 2013. slide 24
Num. Species vs. Num. Variants (all with humans)
A Multipatch Metapopulation Model with “Human” Agents Injected. Manchester Complexity Seminar, 18th Oct 2013. slide 25
Innovation vs. migration rate (all with humans)
A Multipatch Metapopulation Model with “Human” Agents Injected. Manchester Complexity Seminar, 18th Oct 2013. slide 26
Migration vs. food rate (all with humans)
A Multipatch Metapopulation Model with “Human” Agents Injected. Manchester Complexity Seminar, 18th Oct 2013. slide 27
Human migr. rate vs. diversity (all with humans,
other entities having 0.1 migration rate)
A Multipatch Metapopulation Model with “Human” Agents Injected. Manchester Complexity Seminar, 18th Oct 2013. slide 28
Starting from plant ecology (with humans)
Starting from a Plant Ecology the effects of innovation
rate/mutation rate (left) and people migration rate/entity
migration rate (right) , red=plant, blue=mixed,
purple=single species, green=non-viable
A Multipatch Metapopulation Model with “Human” Agents Injected. Manchester Complexity Seminar, 18th Oct 2013. slide 29
Starting from mixed ecology (with humans)
Starting from a Mixed Ecology the effects of innovation
rate/mutation rate (left) and people migration rate/entity
migration rate (right) , red=plant, blue=mixed,
purple=single species, green=non-viable
A Multipatch Metapopulation Model with “Human” Agents Injected. Manchester Complexity Seminar, 18th Oct 2013. slide 30
Starting from single species ecology (with
humans)
Starting from a Single Species Ecology the effects of
innovation rate/mutation rate (left) and people migration
rate/entity migration rate (right) , red=plant, blue=mixed,
purple=single species, green=non-viable
A Multipatch Metapopulation Model with “Human” Agents Injected. Manchester Complexity Seminar, 18th Oct 2013. slide 31
The End!
Bruce Edmonds:
http://bruce.edmonds.name
Centre for Policy Modelling:
http://cfpm.org
A paper describing some of this, the model, these slides at:
http://cfpm.org/cpmrep222.html
A Multipatch Metapopulation Model with “Human” Agents Injected. Manchester Complexity Seminar, 18th Oct 2013. slide 32