Describes a single simulation.
config = { # * = Required, ! = supports default arg only
'agents': { # *
'rice': {
'amount': 10, # *
'id': 1,
<currency>: 10,
'connections': [<Agent>],
'delay_start': 0,
},
#...
},
'termination': [{
'condition': 'time', # !
'value': 100,
'unit': 'day' # 'min', 'hour', 'day', 'year'
}],
'priorities': [ # by agent class; sub-agents step randomly
'structures', 'storage',
'power_generation',
'inhabitants', 'eclss',
'plants'
],
'seed': 12345, # Must be an integer
'global_entropy': 0, # 0 = no variation, 1 = max variation
'location': 'mars' # !
'minutes_per_step': 60, # !
'single_agent': 1, # !
}The data returned by AgentModel.get_data(). To return all fields use with
debug=True. All lists have one entry for each step of the simulation.
data = { # * = default fields
'rice': {
'name': 'rice', # *
'full_amount': 10, # *
'lifetime': 720,
'reproduce': True,
'initial_variable': 0.98765,
'capacity': {<storage>: 100},
'amount': [10, ...], # *
'age': [0, 1, ...],
'step_variable': [1.2, 2.3, ...],
'storage': {<storage>: [0, ...]},
'storage_ratios': {<storage>: [1, ...]},
'flows': {
'in': {<currency>: {<storage>: [1, ...]}},
'out': {<currency>: {<storage>: [1, ...]}},
},
'buffer': {<currency>: [8, ...]},
'deprive': {<currency>: [720, ...]},
'growth': {
'total_growth': [0, ...],
'growth': [0, ...],
'grown': [False, ...],
'agent_step_num': [0, 1, ...],
},
'events': [{
<event>: [{
'magnitude': 0.8,
'duration': 10
}]
}],
'event_multipliers': {<event>: [0, ...]}
}
}currency_desc = {
'food': { # Currency class
'radish': { # * = Required, ^ = food only
'label': 'Radish', # * Display name
'description': 'Radishes, fresh',
'source': <url>, # ^ Nutrition data reference
'unit': 'kg', # ^
'nutrition': { # ^ Grams per <unit>
"kcal": 180,
"water": 941,
"protein": 10,
"carbohydrate": 25,
"fat": 2
}
},
# ...
}
# ...
}Currency classes: atmosphere, nutrients, food, water, energy
agent_desc = {
'plants': { # Agent class
'radish': { # Agent name
'description': '', # Text description
'data': {
'inputs': [...<Input>], # Currencies consumed
'outputs': [...<Output>], # Currencies produced
'characteristics': [...<Char>] # Misc params
}
}
# ...
}
# ...
}
<Input/Output> = { # * = Required
'type': 'co2', # * Currency name, must be in currency_desc
'value': 0.006534, # * Amount exchagned
'flow_rate': { # * Units applied to currency exchanged
'unit': 'kg',
'time': 'hour'
},
'required': 'mandatory', # 'mandatory' = if unavailable, skip step()
# 'desired' = if unavailable, continue step()
'deprive': { # If unavailable, how long to survive
'value': 72,
'unit': 'hour'
}
'growth': { # Map value across the hours in a day
# and/or hours in agent's lifetime such
# that mean hourly value is as defined.
"lifetime": {
"type": "sigmoid" # 'sigmoid' = greater and end-of-lifetime
# 'normal' = greater at mid-life
},
"daily": {
"type": "normal" # 'normal' = greatest in middle of day
# 'clipped' = reduced early/late values
# 'switch' = boolean for 'is daylight'
}
},
'requires': ['h2'], # If input is missing, skip flow
'weighted': 'current_growth' # Multiply value by agent storage amount or attribute
'criteria': { # Activate flow based on view of a connected agent
'name': 'co2_ratio_in', # '<currency>_<view>_<direction>'
'limit': '>', # '=', '>', '<'
'value': 0.001, # What the returned value is compared to
'buffer': 2 # Wait until valid for N steps before activating.
}
}
<Char> = {
'type': 'capacity_o2', # Characteristic type
'value': 10000, # Supports bool, int, float or string
'unit': 'kg' # Optional
}Agent classes: inhabitants, eclss, plants, isru, structures,
fabrication, power_generation, mobility, communication, storage
Characteristic types:
capacity_<currency>: The maximum amount of a particular currency that can be stored.lifetime: Length of one growth cyclecarbon_fixation: 'c3' or 'c4', determines if/how plant responds to ambient co2.volume: m**3mass: kgcategory: sub-class, e.g. 'habitat'reproduce: boolean; whether lifecycle ends or is repeatedcustom_function: two are included in the SIMOC repo:atmosphere_equalizerandrate_finder.threshold_lower_<currency>: Agent is killed if ambient currency falls below
Connections are directional links between agents which determine the source of inputs or destination of outputs.
The to/from fields specify an agent and currency. For the agent field,
two additional options, habitat and greenhouse, are used; when a model
is initialized, those options are replaced with the agent that includes the
word 'habitat' or 'greenhouse' (e.g. 'greenhouse.o2' -> 'greenhouse_medium.o2')
The priority field is optional. If present, when the first connection
(priority=0) is empty, the initiating agent will change to the second
(priority=1) connection, and so on.
agent_conn = [{
'from': '<agent>.<currency>',
'to': '<agent>.<currency>',
'priority': 0
}, ...]Agent variation is off by default. To activate, set the global_entropy
parameter in config to a number 0 < N <= 1.
When active, all currency exchange values are scaled up or down when
initialized and/or every step. Scalars are a random number from a
defined probability density function. The upper and lower parameters
specify the maximum absolute distance up or down from 1 (no effect).
agent_variation = {
'plants': { # Can be agent or agent class
'initial': { # Applied to values on initialization
'upper': 0.5, # Multiplier upper bound
'lower': 0.5, # Multiplier lower bound
'distribution': 'normal' # Probability: 'normal' or 'exponential'
},
'step': {
'upper': 0.1,
'lower': 0.1,
'distribution': 'normal'
}
}Alternatively, upper and lower values can be defined for each individual currency.
agent_variation['humans'] = {
'initial': {
"upper": {
"o2": 0.045,
# ...
},
"lower": {
"o2": 0.025417,
# ...
},
"distribution": "normal",
"stdev_range": 1.65,
"characteristics": ["mass"]
}
# ...
}Agent events are off by default. To activate, set the global_entropy
parameter in config to a number 0 < N <= 1.
agent_events = {
"solar_pv_array_mars": [
{
"type": "duststorm",
"function": "multiplier", # 'multiplier': apply to all flows
# 'termination': kill agent
"scope": "group", # 'group': affects all instances
# 'agent': affects a single instance
"probability": { # Per group/individual based on scope
"value": 0.0004566210046, # Likelihood per step (if not active)
"unit": "hour"
},
"magnitude": {
"value": 1,
"variation": {
"upper": 0, # Maximum remains 1x, no effect
"lower": 0.9, # Minimum is 0.1x
"distribution": "normal" # Mean is 0.55x
}
},
"duration": {
"value": 24, # How long the effect lasts
"unit": "hour",
"variation": {
"upper": 60, # "From 1 to 60 days"
"lower": 1,
"distribution": "exponential" # Likely a low number
}
}
},
# ...
]
}