QUANTUM ALGORITHMS IN DERIVATIVES PRICING MODELS FOR ENHANCED PRICE DISCOVERY AT MCX INDIA

Abstract The purpose of this study is to explore the potential of quantum algorithms in enhancing derivatives pricing models on the Multi Commodity Exchange of India (MCX), focusing on improving the accuracy, efficiency, and reliability of price discovery. By examining the applicability and challenges of quantum computing in a real-market environment, this research aims to contribute to the evolving field of quantum finance and its implications for complex financial instruments. This study employs a hybrid methodology that combines classical linear regression for price prediction and the Quantum Approximate Optimization Algorithm (QAOA) for parameter optimization to analyze commodity prices in the base metals market. The approach involves data collection and preprocessing of trading records, implementation of a linear regression model using scikit-learn, and optimization through QAOA using Qiskit, culminating in a comprehensive evaluation of model performance and integrated visualization of results. The analysis of trading data for base metals from January 3 and 4, 2022, reveals significant trading volume and values, with copper and nickel showing notable market activity. The Quantum Approximate Optimization Algorithm (QAOA) produced optimization parameters that highlight the complexity of accurately predicting market prices, resulting in a high-cost value of 144,460,619,823.28, indicating a considerable discrepancy between predicted and actual trading values. The study underscores the potential of quantum algorithms to enhance price discovery in commodity trading, although challenges remain in achieving precise price predictions due to inherent market complexities and fluctuations. The variations in trading activity across different commodities suggest that while quantum techniques may provide insights, their current application requires further refinement for practical use in real-time trading environments. This analysis is limited by the dataset's relatively short time frame and the number of trading records, which may not fully represent market dynamics over longer periods.