scikit-agent

A scientific Python toolkit for agent-based economic modeling: build models from modular blocks, solve them with deep learning or classical numerical methods, and simulate.

📖 Documentation  ·  Quickstart  ·  Examples  ·  API Reference

Quick example

import skagent as ska
from skagent.models.consumer import cons_problem, calibration

# `cons_problem` is a prebuilt consumption-saving model. Supply a decision
# rule for the control `c` and simulate a population of agents forward.
simulator = ska.MonteCarloSimulator(
    calibration=calibration,
    block=cons_problem,
    dr={"c": lambda m: 0.9 * m},
    initial={"k": 1.0},
    sample_count=1000,
    T_sim=50,
    seed=42,
)
simulator.initialize_sim()
history = simulator.simulate()

The Quickstart goes further, solving the model for an optimal policy instead of hand-coding a rule.

scikit-agent is a scientific Python toolkit for agent-based economic modeling and multi-agent systems design. It provides a unified interface for creating, analyzing, solving, and simulating models using neural, symbolic, and numerical techniques.

Our goal is for scikit-agent to be for computational social science what scikit-learn is for machine learning.

Key Features

• 🧱 Modular modeling system. Construct multi-agent environments from composable blocks of structural equations.
• ⚡ Solution algorithms. Solve models with deep-learning methods (following Maliar, Maliar, and Winant, 2021), value backwards induction, and reinforcement learning via Stable-Baselines3.
• 📊 Simulation tools. Generate synthetic data and run policy experiments with a Monte Carlo engine.
• 🐍 Built on Scientific Python and PyTorch for easy integration with the wider Python ecosystem.

Installation

uv add scikit-agent

Or, outside a uv project, uv pip install scikit-agent (or plain pip install scikit-agent).

For a development installation (uv required):

git clone https://github.com/scikit-agent/scikit-agent.git
cd scikit-agent
uv sync --extra test --extra docs

See the documentation for the user guide, a gallery of runnable examples, and the full API reference.