Agent-Based Economic Simulation

A dynamic agent-based simulation modeling economic interactions between autonomous agents and limited resources. The system explores resource allocation strategies, price dynamics, and wealth distribution under various economic parameters.

Key Features

Autonomous Agents:
50 agents with evolving resource needs, income/expense dynamics, and bankruptcy rules
Resource Ecosystem:
3 resources with price elasticity, regeneration patterns, and capacity adaptation
Economic Mechanics:
Dynamic pricing based on supply/demand
Tax system with wealth redistribution
Bankruptcy detection and agent removal
Resource capacity adaptation to economic output
Analysis Tools:
Gini coefficient calculation
Parameter experimentation framework
Key metric tracking (bankruptcies, balances, prices)

Getting Started

Requirements

Python 3.7+
NumPy

Basic Usage

Run full experimentation suite:

python main.py

Code Structure

File	Purpose
`constants.py`	Central configuration of simulation parameters
`models.py`	Agent/Resource class definitions with core behaviors
`simulation.py`	Main simulation loop and step-by-step execution logic
`helpers.py`	Economic calculations and system operations
`experimentation.py`	Parameter space exploration and result analysis
`main.py`	Entry point for running experiments and viewing results

Core Parameters (constants.py)

Parameter	Description	Default
`NUM_AGENTS`	Initial population size	50
`SIMULATION_STEPS`	Duration of each simulation run	100
`TAX_RATE`	Wealth redistribution percentage	2%
`PRICE_ELASTICITY`	Demand sensitivity to price changes	0.05
`RESOURCE_REGEN_RATE`	Base resource regeneration rate	1%
`BANKRUPTCY_THRESHOLD`	Balance level for agent removal	-50
`AGENT_INCOME_CEILING`	Maximum possible agent income	1.0

Experimentation Insights

The system tests four key parameters across ranges:

Price Elasticity (0.01-0.1)
Resource Regeneration (0.005-0.02)
Tax Rates (0-5%)
Expense Rates (10-50%)

Sample findings:

Tax rate that minimizes bankruptcies: 0.0278
Regen rate that maximizes average final balance: 0.0189

Key Metrics Tracked

Average agent balance
Wealth inequality (Gini coefficient)
Bankruptcy count
Resource price stability
System-wide economic output

Simulation Flow

Price updates based on resource utilization
Resource allocation through agent bidding
Income distribution and expense deduction
Tax collection/redistribution
Bankruptcy checks and agent removal
Resource regeneration and capacity adjustment
Agent behavior adaptation

Analysis Capabilities

Parameter sensitivity testing
Optimal policy identification
System stability evaluation
Emergent pattern detection
Equilibrium state analysis

Source Code

Browse the source repository

GitHub Repository