Research Of Option Pricing With Backward Stochastic Differential Equations Based On Maprduce

The option prices is the only option contracts with market supply and demand and change variables, it directly affects the buyers and sellers of the profit and loss situation, the core problem is options trading. Option pricing in financial application fields is mathematically one of the most complicated problems. MapReduce is a relatively simple programming model, and currently in the massive datasets widely used in parallel computing. The main idea is MapReduce borrowing from the functional programming language, the vast Numbers of not distributed parallel programming personnel in distributed system study on writing and operation procedures to reduce the threshold. MapReduce in many applications to be used extensively, such as web connection diagram reversal, web access log analysis, distribution, grep, distribution, reverse index build order document clustering, machine learning, based on the statistical machine translation etc.Along with the computer technology is applied to fields related to finance, looking for efficient method in modern architecture to realize option pricing model becomes more and more important. With the development of the technology, option pricing needs to seek appropriate methods to improve traditional option pricing efficiency. Cloud computing is getting hotter now of how to make financial calculation through cloud computing platforms completed, is that we need to do their job. Put option pricing numerical method through parallel computing to achieve a recent relatively new also hotter studies, in which backward stochastic differential equations using method of realization option pricing, is higher, with show calculation accuracy of financial market closest one way. Through the backward stochastic differential equation realize option pricing although there exists much advantage in accuracy, but it need large amount of calculation, computational complexity, the time required for very long. On the financial markets accurate expected a second time, early, may bring hefty gains. In order to shorten the calculation time, improve the timeliness, we are expected to realize option pricing through computing tasks with MapReduce, and make it parallelize quickly and efficiently. To our knowledge, this is the first time to use framework implementation option pricing on through MapReduce hadoop platform. We hope for financial calculations can be realized by clouds computing more and more efficiently.This article presents a method of option pricing implemented on MapReduce. It parallel option pricing with BSDES on MapReduce and improve the computing efficiency. In this article, it first presents the research background, domestic and international research status an. Secondly, it introduces the option pricing related concepts, commonly option pricing model, and detailed introduces the use of backward stochastic differential equation method option pricing model. In addition MapReduce framework is introduced in detail, the executive scheduling process. Thirdly, it is the focus of this article, it detailed introduces option pricing parallelization algorithm backward stochastic differential equations on MapReduce, including parallelize thought of stochastic differential equation a, the designing of model<key, value>, the designing of map function and reduce function, model architecture and model of data flow. In the experimental parts, it introduces the MapReduce environment deployment, the analysis of experimental results of model and performance evaluation. In addition MapReduce according to experiment is compared with other parallel tools, puts forward the advantages and disadvantages of the place can be improved. At last, it mainly summarizes this paper, based on the way that MapReduce stochastic differential equation option pricing parallelization model and the characteristics of the deficiencies, and summarizes the place can be improved, and discusses the future work.