Skip to content

Latest commit

 

History

History
234 lines (181 loc) · 5.39 KB

README.md

File metadata and controls

234 lines (181 loc) · 5.39 KB

Trading Bot System Documentation

Overview

The Trading Bot System is a sophisticated algorithmic trading platform written in C++. It provides a comprehensive suite of tools for backtesting and optimizing trading strategies using various technical indicators and optimization algorithms.

System Architecture

Core Components

  1. Strategy Module

    • Base strategy interface (IStrategy)
    • Implemented strategies:
      • Mean Reversion
      • Momentum
      • Bollinger Bands
      • RSI (Relative Strength Index)
    • Each strategy implements signal generation and state management

  2. Optimization Module

    • Parameter optimization framework
    • Implemented algorithms:
      • Genetic Algorithm
      • Hill Climbing
    • Parallel evaluation support via OpenMP
    • Configurable fitness functions

  3. Backtesting Engine

    • Historical data processing
    • Performance metrics calculation
    • Trade execution simulation

  4. Data Management

    • Market data handling
    • CSV data import/export
    • JSON configuration and results storage

Technical Details

Strategy Implementation

Mean Reversion Strategy

  • Uses z-score for identifying overbought/oversold conditions
  • Parameters:
    • z_score_threshold: Signal threshold
    • window_size: Lookback period
    • take_profit: Profit target
    • stop_loss: Loss limit

Momentum Strategy

  • Tracks price momentum over a lookback period
  • Parameters:
    • lookback_period: Momentum calculation window
    • threshold: Signal threshold
    • take_profit: Profit target
    • stop_loss: Loss limit

Bollinger Bands Strategy

  • Uses standard deviation bands for volatility-based signals
  • Parameters:
    • window_size: Moving average period
    • num_std_dev: Standard deviation multiplier
    • take_profit: Profit target
    • stop_loss: Loss limit

RSI Strategy

  • Implements Relative Strength Index indicator
  • Parameters:
    • period: RSI calculation period
    • overbought: Upper threshold
    • oversold: Lower threshold
    • take_profit: Profit target
    • stop_loss: Loss limit

Optimization Algorithms

Genetic Algorithm

  • Population-based optimization
  • Features:
    • Tournament selection
    • Crossover and mutation
    • Elitism
    • Parallel evaluation
  • Parameters:
    • population_size: Number of individuals
    • generations: Number of iterations
    • mutation_rate: Probability of mutation
    • crossover_rate: Probability of crossover

Hill Climbing

  • Local search optimization
  • Features:
    • Random restart
    • Adaptive step size
    • Parallel neighbor evaluation
  • Parameters:
    • max_iterations: Number of iterations
    • step_size: Initial step size
    • step_decay: Step size reduction rate

Performance Metrics

The system calculates various performance metrics:

  • Total Profit
  • Sharpe Ratio
  • Maximum Drawdown
  • Win Rate
  • Sortino Ratio
  • Maximum Trade Duration

Fitness Function

The optimization process uses a weighted fitness function:

fitness = profit_weight * total_profit +
          sharpe_weight * sharpe_ratio -
          drawdown_weight * max_drawdown +
          win_rate_weight * win_rate +
          sortino_weight * sortino_ratio -
          max_trade_duration_weight * max_trade_duration

Usage

Compilation

# Create build directory
mkdir build && cd build

# Configure with optimization enabled
cmake -DENABLE_OPTIMIZATION=ON ..

# Build
make

Running Optimization

# Basic usage
./tradingbot_optimize data.csv

# With specific strategy and algorithm
./tradingbot_optimize --strategy=rsi --algorithm=hill_climbing data.csv

# With custom weights
./tradingbot_optimize --weights=profit:0.6,sharpe:0.3,drawdown:0.1 data.csv

# With parallel processing
./tradingbot_optimize --threads=4 data.csv

Command Line Options

  • --strategy=<name>: Select strategy (default: mean_reversion)
  • --algorithm=<name>: Select optimization algorithm (genetic or hill_climbing)
  • --weights=<config>: Configure fitness weights
  • --threads=<n>: Number of threads for parallel processing
  • --generations=<n>: Number of generations/iterations
  • --population=<n>: Population size

Development

Adding New Strategies

  1. Create strategy header file:
#pragma once
#include "IStrategy.hpp"

class NewStrategy : public IStrategy {
public:
    struct Config {
        // Strategy parameters
    };
    
    // Implement required methods
    virtual Signal generateSignal(const MarketData& data) override;
    virtual void updateState(const MarketData& data) override;
    virtual void reset() override;
};
  1. Implement strategy:
#include "NewStrategy.hpp"

// Implement methods
  1. Add to strategy factory in main_optimize.cpp

Adding New Optimization Algorithms

  1. Create optimizer header:
#pragma once
#include "ParameterOptimizer.hpp"

class NewOptimizer {
public:
    // Constructor and methods
    Individual optimize();
};
  1. Implement optimizer
  2. Add to main optimization logic

Dependencies

  • C++17 or later
  • OpenMP
  • nlohmann/json
  • CMake 3.10 or later

License

This project is licensed under the MIT License - see the LICENSE file for details.

Author

Contributing

  1. Fork the repository
  2. Create your feature branch
  3. Commit your changes
  4. Push to the branch
  5. Create a new Pull Request