Model

A Honeybees model consists of a model class, which contains an agent class, area class, and reporter class. The area class defines the study area through its bounds (xmin, xmax, ymin, ymax), and contains any additional features to be displayed, such as rivers.

The agent class is, of course, the hearth of the model. This class is initialized from the model itself, and itself contains all agent classes. In the example below the agent class is initialized as self.agents = Agents(self). Then, in the agent class the people class is initialized through self.people = People(model, self). Any additional agent classes could be initialized for example using self.government = Government(model, self). By passing the agent class (that contains all agent types) to the specific agent type by passing self, the agents can also interact with each other (see example in Introduction).

Then, the current time (start time), timestep and number of timesteps are set. Then, the reporter is initialized. This class is used to report agent data to disk, and can be configured in the specified config file (examples/config.yml in this example). See Reporting for more details.

from datetime import date
from dateutil.relativedelta import relativedelta

from honeybees.model import Model
from honeybees.area import Area
from honeybees.reporter import Reporter
from honeybees.agents import AgentBaseClass

class People(AgentBaseClass):
    def __init__(self, model, agents):
        self.n = 1000
        self.model = model
        self.agents = agents
        self.age = np.random.randint(0, 100, self.n)
        self.locations = np.zeros((self.n, 2), dtype=np.float32)
        self.update_locations()

    def update_locations(self):
        self.locations[:, 0] = np.random.uniform(self.model.xmin, self.model.xmax, self.n)
        self.locations[:, 1] = np.random.uniform(self.model.ymin, self.model.ymax, self.n)

    def step(self):
        self.age += 1
        self.update_locations()

class Agents:
    def __init__(self, model):
        self.people = People(model, self)

    def step(self):
        self.people.step()

class ABMModel(Model):
    def __init__(self, config_path, study_area, args=None):
        self.area = Area(self, study_area)
        self.agents = Agents(self)

        current_time = date(2020, 1, 1)
        timestep_length = relativedelta(years=1)
        n_timesteps = 10

        Model.__init__(self, current_time, timestep_length, config_path, args=args, n_timesteps=n_timesteps)

        self.reporter = Reporter(self)


config_path = None
study_area = {
    'name': 'TestArea',
    'xmin': -10,
    'xmax': 10,
    'ymin': -10,
    'ymax': 10
}
model = ABMModel(config_path, study_area)

The model can then be run by stepping through the model:

for i in range(10):
    model.step()

Or by performing all timesteps at once by calling the run method:

model.run()