Title:Pricing of options on stocks driven by multi-dimensional operator stable Levy processes

Abstract: We model the price of a stock via a Langévin equation with multi-dimensional fluctuations coupled in the price and in time. We generalize previous models in that we assume that the fluctuations conditioned on the time step are compound Poisson processes with operator stable jump intensities. We derive exact relations for Fourier transforms of the jump intensity in case of different scaling indices $\underline{\underline{E}}$ of the process. We express the Fourier transform of the joint probability density of the process to attain given values at several different times and to attain a given maximal value in a given time period through Fourier transforms of the jump intensity. Then we consider a portfolio composed of stocks and of options on stocks and we derive the Fourier transform of a random variable $\mathfrak{D}_t$ (deviation of the portfolio) that is defined as a small temporal change of the portfolio diminished by the the compound interest earned. We show that if the price of the option at time $t$ satisfies a certain functional equation specified in text then the deviation of the portfolio has a zero mean $E[ \mathfrak{D}_t ] = 0$ and the option pricing problem may have a solution. We compare our approach to other approaches that assumed the log-characteristic function of the fluctuations that drive the stock price to be an analytic function.