Solving Black-Scholes PDE associated with Local Volatility via Physics-Informed Neural Network

Author(s)
이무현
Advisor
배형옥
Department
일반대학원 금융공학과
Publisher
The Graduate School, Ajou University
Publication Year
2021-08
Language
eng
Keyword
Black-Scholes PDEDeep LearningDupire’s equationLocal volatilityOption Pricingparametric PDE
Alternative Abstract
In this article, we solve the Black-Scholes Partial Difference Equation(BS PDE) under the local volatility model by using an artificial neural network, and obtain the price and the greeks of derivatives in forms of function simultaneously. Dupire’s local volatility model is one of the most successful for equity models. In practice, it is important to price and hedge derivatives under the local volatility model. We provide an artificial neural network scheme for efficiently solving parametric PDEs. We adopt local volatility models, especially, constant elasticity of variance(CEV) model and volatility surface. The price function of European vanilla options under the local volatility model satisfies Dupire's equation. We solve the parametric PDE of the European put option under the local volatility model with an artificial neural network and show that solution and Dupire's equation are approximated.
URI
https://dspace.ajou.ac.kr/handle/2018.oak/20430
Fulltext

Appears in Collections:
Graduate School of Ajou University > Department of Financial Engineering > 3. Theses(Master)
Files in This Item:
There are no files associated with this item.
Export
RIS (EndNote)
XLS (Excel)
XML

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.

Browse