Mesa tutorial: plot of agent wealth is deterministic instead of random

Question

thanks for taking the time. (Python 3.7.0) I'm a beginner at python and doing the Mesa tutorial since I want to make an agent-based model for a research.

I have the following issue: when I run the following code, each time a randomized plot showing the wealth of 10 agents in the model should come up. The agents all start with wealth 1 and randomly trade (=give wealth) with each other. However, the plot is always the same and just shows a stack of value 10! I think there is something wrong in the definition of agent_wealth, but I took it straight from the tutorial.

from mesa_tutorial import * #import all definitions from mesa_tutorial
import matplotlib.pyplot as plt
model = MoneyModel(10)
for i in range(10):
  model.step()
agent_wealth = [a.wealth for a in model.schedule.agents]
plt.hist(agent_wealth)
plt.show()

Resulting in the following plot: non-random plot with stack 10

Here's the definition of the model

class MoneyModel(Model): # define MoneyModel as a Subclass of Model
'''A model with some number (N) of agents'''
  def __init__(self, N):
      #Create N agents
      self.num_agents = N
      self.schedule = RandomActivation(self) #Executes the step of all agents, one at a time, in random order.         
      for i in range(self.num_agents): #loop with a range of N = number of agents           
          a = MoneyAgent(i, self) # no idea what happens here, a = agent?            
          self.schedule.add(a) #adds a to the schedule of the model       

  def step(self):
      '''Advance the model by one step'''         
      self.schedule.step()

score 0 · Answer 1 · answered Oct 30 '18 at 10:09

Can you post your Moneyagent class in this class the agents should randomly exchange money. See the step function below.

# model.py
class MoneyAgent(Agent):
    """ An agent with fixed initial wealth."""
    def __init__(self, unique_id, model):
        super().__init__(unique_id, model)
        self.wealth = 1

    def step(self):
        if self.wealth == 0:
            return
        other_agent = random.choice(self.model.schedule.agents)
        other_agent.wealth += 1
        self.wealth -= 1

With this step function you should start getting a positively skewed distribution or the positive half of a normal distribution, if the agents can go negative.

I solved it, thanks! The issue was that the step method/function was a subfunction of the init function, i.e. it had one indentation too much. So the step function could only be run by calling the init function, instead of by agent.step(). Now that I solved that it works! Thanks for the help! — Thijs van der Plas, Oct 31 '18 at 12:34

Mesa tutorial: plot of agent wealth is deterministic instead of random

1 Answers1