Civil Violence

Agents

class whynot.simulators.civil_violence.Agent[source]

Covariates for a single agent in the simulation.

Examples

>>> # An agent with low risk aversion and high hardship
>>> civil_violence.Agent(hardship=0.99, risk_aversion=0.01)
active_threshold = 0.1

Threshold above which agent starts to openly rebell

hardship = 0.5

How aggrieved is the agent by external circumstances.

legitimacy = 0.5

How legitimate does the agent perceive the current regime.

risk_aversion = 0.5

How likely the agent is to rebel for a fixed greviance level.

vision = 3

How many adjacent squares an agent sees and uses to determine probability of arrest.

Config

class whynot.simulators.civil_violence.Config[source]

Metaparameters for a single run of the civil violence simulator.

arrest_prob_constant = 2.3

A fixed parameter to calibrate arrest likelihood.

cop_fraction = 0.05

What fraction of the agents are police

cop_vision = 5

How many adjacent squares a cop can see when determining who to arrest.

grid_height = 50

Vertical grid size

grid_width = 50

Horizontal grid size

prison_interaction = 0.1

How strongly other agent grievances affect each agent in prison.

Experiments

Experiments for civil violence simulator.

whynot.simulators.civil_violence.experiments.RCT = <whynot.framework.GenericExperiment object>

RCT experiment that explores how violations of SUTVA impact inference.