Options Trader Simulator

This is a fullstack project for simulating options trading. I hope to learn the basics of principals behind options pricing as well as apply efficient system design.

The project consists of a high-performance C++ pricing engine that implements option pricing models (currently European options using the Black-Scholes formula and American option using the Binomial Tree Model), and will later include a frontend UI (React) and full portfolio/trading management features.

---

Structure

• backend/: C++ pricing engine & (planned) simulation logic
• frontend/: (planned) UI for user interaction and effective visualization
• data/: (planned) sample historical data for testing

---

Status

Completed (Reverse Chronological Order):

• American Option speedup: BAW
• Consideration of dividends
• Data Flow (I/O): Market Data Feed
• Transition to Linux (WSL)
• Memory reuse for Binomial Tree (thread-local workspace)
• SIMD + OpenMP batch pricing (1M options in <20ms)
• Unified dispatcher for European & American pricing
• American Option pricing & Greeks via Binomial Tree
• Performance Benchmarking
• Template-separated European Call/Put
• European Options Greeks (delta, gamma, etc.)
• C++ pricing engine complete for European call/put

In Progress:

• American options speedup: Ju-Zhong Models

Planned:

• Look into saving previously calculated Binomial Tree Prices, avoid recalculation ever steps; have to fix ticks per unit time & shift tree instead of recalculate
• GPU acceleration (maybe)
• Portfolio simulation: track cash balance, option positions, etc. Methods like buy, sell
• Market Simulation: Simulate realistic market conditions: volatility, price movements
• Strategies/Backtesting
• React frontend

---

Performance Techniques

• SIMD Vectorization: Batch pricing leverages vectorized CPU instructions to price many options at once
• Multithreading (OpenMP): Workloads are parallelized across CPU cores to increase throughput
• Batch Processing: Grouping options into SoA format for high-throughput vectorized pricing per CPU
• Dispatcher Model: Dynamically chooses between models based on option style (European vs American)

---

Build & Run (C++ Backend on Linux with Boost.Asio + OpenMP)

# Enter your backend project directory
cd backend

# Create a build directory and run CMake
mkdir -p build && cd build
cmake ..

# Build with all cores
make -j$(nproc)

# Run the simulator
./options_simulator

Requirements (Ubuntu/WSL)

Make sure you have the following dependencies installed:

sudo apt update
sudo apt install build-essential cmake libboost-all-dev

This installs:

• GCC with OpenMP support
• Boost (including Boost.Asio)
• CMake for building the project