Aries: Risk-Averse Trading Agent for Colombian Energy Market

Overview

Aries is a sophisticated Reinforcement Learning-based trading agent designed for the Colombian wholesale energy market (XM). The system combines probabilistic forecasting with risk-averse decision-making to optimize energy trading strategies.

Features

• Probabilistic Forecasting: LSTM/Transformer-based models for price prediction with uncertainty quantification
• Risk-Averse RL Agent: PPO/SAC algorithms with CVaR-based risk management
• Market Simulation: Realistic trading environment using historical data
• Real-time Integration: XM API integration for live market data
• San Andrés Focus: Specialized for San Andrés and Providencia energy market dynamics
• Web Interface: Streamlit-based dashboard for monitoring and control

Architecture

Phase 1: Data Foundation

• Historical energy price data from XM API
• Weather and external factors
• San Andrés specific data integration

Phase 2: Probabilistic Forecasting

• LSTM/Transformer models for price prediction
• Quantile regression for uncertainty bounds
• Mixture Density Networks for distribution modeling

Phase 3: Risk-Averse RL Agent

• State representation with forecast distributions
• Action space: buy/sell/hold decisions
• Reward function: Profit - λ × CVaR

Phase 4: Integration & Deployment

• Real-time market simulation
• Backtesting framework
• Performance monitoring dashboard

Installation

# Clone the repository
git clone https://github.com/your-repo/aries.git
cd aries

# Install dependencies
pip install -r requirements.txt

# Set up environment variables
cp .env.example .env
# Edit .env with your configuration

Quick Start

from aries.agent import RiskAverseTradingAgent
from aries.environment import EnergyMarketEnv
from aries.forecaster import ProbabilisticForecaster

# Initialize components
forecaster = ProbabilisticForecaster()
env = EnergyMarketEnv()
agent = RiskAverseTradingAgent(env, forecaster)

# Train the agent
agent.train(episodes=10000)

# Run backtesting
results = agent.backtest(test_data)

Usage

Web Interface

streamlit run app.py

Command Line

python -m aries.cli train --episodes 10000 --risk-aversion 0.5
python -m aries.cli backtest --model-path models/best_model.zip

Configuration

The system can be configured through environment variables or config files:

• RISK_AVERSION_LAMBDA: Risk aversion coefficient (default: 0.5)
• FORECAST_HORIZON: Prediction horizon in hours (default: 24)
• TRADING_FREQUENCY: Trading frequency in minutes (default: 60)
• MAX_POSITION_SIZE: Maximum position size (default: 1000)

Performance Metrics

• Financial: Total return, Sharpe ratio, Maximum drawdown
• Risk: CVaR, Value at Risk (VaR)
• Forecasting: CRPS, MAE, RMSE
• Trading: Win rate, Average trade duration

Contributing

1. Fork the repository
2. Create a feature branch
3. Make your changes
4. Add tests
5. Submit a pull request

License

MIT License - see LICENSE file for details