In mathematical finance we are interested in pricing derivative contracts, especially option contracts. Therefore quantitative models like the famous Black-Scholes model or extensions like the constant elasticity of variance (CEV) model were developed. The calibration of these quantitative models to real market data for identifying the model parameter is of great interest in the area of option pricing. The computational cost is a very crucial factor herein because calibration is an inverse problem, hence, for a wide range of parameters the option price has to be evaluated. So we are interested in efficient and reliable numerical methods for the pricing of options in the CEV model. To this aim we present a reduced-basis method for the linear parabolic CEV option pricing partial differential equation (PDE). The essential ingredients of the reduced-basis method are the following: first, in a offline stage a rapidly convergent reduced-basis approximation is constructed through a Galerkin projection onto a space $\WN$ of very small dimension $N$ spanned by solutions of the concerning PDE at $N$ selected parameter-time samples. Secondly, a efficient online computational procedure is required, where for a given new parameter value the reduced-basis approximation is calculated at very low computational cost. The reduced-basis method is thus ideally suited for efficient and reliable evaluation of parametrized problems in the many-query or real-time contexts, i.e. perfectly fitting to our concerning task of calibration for parameter identification within the CEV model. We present the analytical formulation and numerical results for the application of the reduced-basis method for the CEV option pricing problem. Even if it suffices to consider the reduced-basis method for affine parameter dependent problems in our case, we also extend our presented reduced-basis method to nonaffine problem, since we are aiming to use more complex financial models in the future where nonaffine terms will occur. To this end, we introduce the empirical interpolation method for getting an affine approximation of the nonaffine term. Finally, we provide numerical results for the application of the reduced-basis method to the nonaffine CEV option pricing problem.
«
In mathematical finance we are interested in pricing derivative contracts, especially option contracts. Therefore quantitative models like the famous Black-Scholes model or extensions like the constant elasticity of variance (CEV) model were developed. The calibration of these quantitative models to real market data for identifying the model parameter is of great interest in the area of option pricing. The computational cost is a very crucial factor herein because calibration is an inverse probl...
»