The calibration of quantitative models - as the famous Black-Scholes model or extensions, for instance, the constant elasticity of variance (CEV) model - to real market data with the ambition of a productive use in the area of option pricing is of great interest in mathematical finance. Due to the inverse character of the underlying problem, computational cost is a crucial factor. Hence, we are interested in efficient and reliable numerical methods for the pricing of European options in the CEV model. This thesis thoroughly considers analytical closed formulas as well as the finite element method. First, we discuss implementations of the former, the so-called CEV formula. Second, we develop a novel, well-defined framework for the variational solution of the pricing PDE for European options in the CEV model, which may not suffer from convection dominance; addresses all possible values of elasticity of variance; as well as incorporates a symmetric and coercive bilinear form. Third, to capture the temporal inhomogeneity of the latter, we generalize existing stability and convergence results regarding the finite element method. Fourth, we present numerical outcomes, including convection-dominated as well as generic calibration setups. Fifth, we remark on a potential future application of the reduced basis method within the developed framework.
«
The calibration of quantitative models - as the famous Black-Scholes model or extensions, for instance, the constant elasticity of variance (CEV) model - to real market data with the ambition of a productive use in the area of option pricing is of great interest in mathematical finance. Due to the inverse character of the underlying problem, computational cost is a crucial factor. Hence, we are interested in efficient and reliable numerical methods for the pricing of European options in the CEV...
»