Pricing early-exercise and discrete barrier options by Fourier-cosine series expansions (Q849055)

This paper generalizes the COS (method based on Fourier-cosine series expansion) pricing method to Bermudan and discrete-monitored barrier options. This method can be used whenever the characteristic function of the underlying price process is available, such as regular affine diffusion processes and, in particular, for exponential Lévy processes. The COS method exhibits an exponential convergence for density functions in \(C^{\infty}[a,b]\) and an impressive computational speed. It compares favorably to the CONV(convolution) method.