Option Pricing Model And Its Variance Reduction Techniqu Research

With the rapid development of global financial industry, option pricing has got more and more attention. This paper focuses on the numerical methods in option pricing model, trying to discuss some option pricing issues through different ways.Firstly, for the error problems from the option pricing, we put forward the composite variance reduction technique.1, For American put options, we put forward the composite variance reduction technique which is the combination of the control variable method and the dual variables method. By simulation, it is concluded that the results of the composite variance reduction technique are better than using only the control variable method. Futher, for the Monte Carlo simulation method, the quasi-Monte Carlo simulation method is introduced. By simulation calculation, we find that in the case of high dimension, quasi-Monte Carlo simulation method is more stable than the Monte Carlo simulation method.2, For discrete barrier options, we put forward the composite variance reduction technique which is the combination of the conditional sampling method, the importance sampling method and the dual variable method. Through simulation calculation, we find that the results got from the composite variance reduction technique with the three kinds variance reduction technique are more accurate than using a single variance reduction technique.Secondly, according to the existing binary tree option pricing model, we put forward a new type of trigeminal tree option pricing model. Through new trigeminal tree option pricing model, we can predict the price of the option in the short term. The light media data from April 1, 2012 to March 29,2013 are selected according to the 3 days,5 days and 8 days of interval period of division. Then we give the empirical analysis to the short term datas. By comparison, we find that results predicted from trigeminal tree option pricing model are more consistent than binary tree option pricing model.Finally, we give the random permutation quasi-Monte Carlo simulation method, in which case, the random number are not uniformly distributed in the high dimension. By working with the ordinary Monte Carlo simulation method and the quasi Monte Carlo simulation method of Halton sequence, Faure sequence and Sobol sequence pricing options in different dimensions, the quasi-Monte Carlo simulation of random permutation method is better in higher dimension.