| As the guarantee of national scientific research,the "infrastructure" of high performance computing cluster has become a national strategy.High performance computing is widely used,especially in the field of financial engineering,and is an indispensable tool.At present,high-performance financial computing platform for providing high-quality graphics,vision and remote financial research users has become a breakthrough in high-performance computing research.In financial markets,especially in financial transactions,any time and delay of information may bring huge economic losses.Therefore,the option pricing problem needs the high real-time performance of the algorithm.Backward stochastic differential equation(BSDE)has been widely studied in financial computation in recent years and applied to the problem of option pricing algorithm.Compared with Black-Scholes formula,BSDE calculation is more accurate when the probability model is uncertain.The BSDE-Theta numerical scheme is mainly obtained by combining PDE high-order numerical scheme with BSDE backward stochastic differential equation's own characteristics.Theta scheme discrete BSDE is used in time range,and the conditional numerical variance is calculated by Monte Carlo.The interpolation method is used to obtain the values of random meshless points in operation.Through this method,the result is very accurate.Therefore,this paper proposes a high-performance computing platform system based on Option Pricing in financial markets.The system is developed based on Python language and uses B/S architecture to provide services for users.The integration of BSDE high precision numerical calculation method and platform is realized.Designing user-friendly user interface,providing convenient access methods,and combining with CPU and GPU heterogeneous computing framework,providing platform users with cross-node computing needs,ensuring more accurate and efficient computing resources in option pricing applications. |
PDF Full Text Request