Parallel Monte Carlo Method To Study Option Pricing Applications

Financial analysis is the important application of calculation science at present. It receives the more and more widespread attention. But along with the technical development, the financial analysis proposed the more and more complex random question, so it is very difficult or even impossible to give a approximate result with the definite method. The Monte Carlo method is the most commonly used method to the financial analysis, sometimes even is the only method. At the same time it is a very effective special numerical analysis method for European option pricing. However, due to the huge number of experiments, sometimes on millions of simulation, there is a considerable amount of computation on the Monte Carlo method for an effective pricing process. Enormous computational cost has serious hindered the application of the Monte Carlo method, therefore it is urgent needs to solve the problem of major computation quantity.The parallel machine provided the parallel computing method which solves large amount of computation and long time for calculating problem.Through the study of the basic principle of the Monte Carlo method and the specialty of the European option pricing, this article carried on the simulation with Monte Carlo method for European option pricing. In view of the complexity of the financial analysis and the huge computation quantity question of Monte Carlo method, this article used the pseudo-random numbers of lognormal distribution instead of random numbers, handed over the huge number of pseudo-random to the computer completed. This article proposed the parallel Monte Carlo method to sovle the problem, based on the deep analysis of the simulation process and the characteristics of parallel computation. Through the research and analysis of communication time between cluster nodes in COW, the parallel algorithm hase been improved again and again. The best parallel algorithm used master-slave programming model, realized load-balancing, improved the utilization rate of processors in COW, effectively solved the problem that the running time was long on enormous computation for serial algorithms. Through the confirmation, the parallel algorithm obtained a high speedup and parallel efficiency, enhanced the efficiency of the computation greatly, reduced the time of execution, completed the complex and massive task of calculating at a lower cost.