Wealth Distribution Simulation

This project simulates the evolution of wealth and education in a population of agents over time. It models inheritance, education, interactions, and demographic processes, and provides tools for visualizing the resulting distributions.

Features

• Agent-based simulation of wealth, education, and demographic changes.
• Inheritance mechanism: Wealth is passed to children upon death.
• Education growth: Agents increase their education over time, affecting their income.
• Interactions: Agents interact and transact wealth based on proximity and probability.
• Demographic realism: Agents have parents, children, and age-dependent death probability.
• Metrics logging: Wealth and education statistics are logged at each iteration.
• Visualization: Jupyter notebook for plotting Gini coefficient, wealth/education percentiles, and more.

Getting Started

Prerequisites

• Rust (edition 2021)
• Python 3 with matplotlib and pandas for visualization

Build & Run

1. Build the simulation:

   cargo build --release

2. Run the simulation:

   cargo run --release

   By default, configuration is loaded from config/default.json. You can provide a custom config file as an argument.

   cargo run --release -- path/to/your/config.json

   If you want to run a simulation without a visualisation window, provide an environment variable VISUALISE = false:

   VISUALISE=false cargo run --release -- path/to/your/config.json

3. **View results:**cargo run --release

   • Simulation metrics are saved to visualisation/metrics.csv.
   • Open visualisation/visualisation.ipynb in Jupyter to plot and analyze results.

Configuration

Simulation parameters are set in config/default.json, including:

• Number of agents
• World size
• Interaction parameters
• Tax rate
• Age and death curve parameters
• Education and wealth parameters

Example:

{
  "num_iterations": 5000,
  "num_agents": 1000,
  "length": 1000,
  "width": 1000,
  "interaction_radius": 50.0,
  "max_movement": 15.0,
  "age_and_death": {
    "mean_age": 30.0,
    "stddev_age": 10.0,
    "mid_age": 80.0,
    "max_start_age": 90.0,
    "steepness": 0.02
  },
  "education": {
    "initial_adult_min": 4.0,
    "initial_adult_max": 10.0,
    "elemental_education_threshold": 4.0,
    "children_education_jitter": 2.0,
    "learning_rate_min": 0.005,
    "learning_rate_max": 0.05,
    "max": 10.0
  },
  "income_and_consumption": {
    "income_age_parameter": 0.05,
    "income_education_parameter": 2.0,
    "base_consumption": 10.0,
    "aditional_consumption_rate": 0.2
  },
  "transaction": {
    "transaction_probability": 0.3,
    "education_parameter": 1.0,
    "age_parameter": 0.001,
    "tax_rate": 0.05,
    "amount_rate": 0.05
  },
  "wealth": {
    "min_initial_wealth": 10.0,
    "max_initial_wealth": 100.0,
    "min_inheritance_at_birth_rate": 0.1,
    "max_inheritance_at_birth_rate": 0.3
  }
}

Visualization

The visualisation/visualisation.ipynb notebook provides plots for:

• Gini coefficient over time
• Wealth percentiles over time
• Total wealth and adult agent count
• Education mean and percentiles
• Death probability curve

File Structure

• src/agent.rs - Agent definition and behavior
• src/environment.rs - Simulation environment and logic
• src/metrics.rs - Logging of simulation metrics (wealth and education)
• src/environment_config.rs - Configuration structs and loading
• visualisation/metrics.csv - Output metrics for plotting
• visualisation/visualisation.ipynb - Jupyter notebook for analysis

License

MIT License

---

Authors:
Kacper Cienkosz - GitHub
Miłosz Dubiel - GitHub