Design And Implementation Of A Finance Quantitative Strategy Development Platform

As financial tool is innovated, financial derivatives, like future and option, bring a lot of arbitrage opportunities. As compute science develops, quantitative investment model and program trade performs better on analyzing massive and complex finance information than human being.Quantitative Investment takes advantages of statistics analysis base on computer science. Quantitative investment researcher firstly proposes a hypothesis based on observed principle on the finance market, then extracts predictive factor and a logical model, writes a strategy program and validate it based on history finance market data, optimize the strategy program and test it based on finance market simulation.As computer science develops, quantitative investment takes obvious advantages. The concept of quantitative investment spreads widely in investing funds. In developed countries quantitative investment is highly practiced. The investing model and software infrastructure have developed for more then 30 years. Quantitative investment is accepted by domestic funds and security companies in recent years. Quantitative investment software platform develops on high speed. A quantitative strategy development platform on Windows system is developed on demand of local hedge fund, which offers functions.This article analyses the research situation of quantitative investment both in Local and abroad first and compare the difference. Based on the development requirement of local future company, it provides a suggested quantitative investment analysis platform schema, which provides for investors and analyst a demo strategy with a series of functions to develop, back test, optimize and run a strategy. GPU and CUDA is introduced to users as a component of the platform. So it costs less than massive parallel cloud quantitative platform and is more efficient than application client.Finance quantitative analysis platform provides basic functions for strategy developing, strategy test, virtual market data connection, real trade connection and other functionality. The platform provides a demo strategy program project for users to develop agilely. Strategy is imported into the platform after implementation.Strategy is validated and tested by history market. The platform provides functionality to store and access history market data. Analyzing huge size of history market data costs massive time. Optimizing history market data analysis computing model is a key point of the platform. The platform provides a friendly GPU program interface to enhance history market data analyzing efficiency up to 20 times faster. The platform uses history market data to simulate finance market to trigger strategy to validate its performance and profitability. Real future market connection and future transaction connection is provided for users to run strategy on real market.