程序化交易算法模型的研究

发布时间：2018-11-19 12:24

【摘要】：随着近几年证券市场和程序化交易平台的不断发展和完善,程序化交易在证券交易市场的扮演着越来越重要的角色。程序化交易是通过计算机执行设定的程序的过程。而这个计算机程序就是交易者的交易思想和模型的计算机程序化。因此,程序化交易是依赖于计算机交易平台和交易模型的。由于国内程序化交易平台的开发和发展已经比较完善,所以研究分析交易模型成了影响程序化交易的关键因素。程序化交易的最大的优势在于完全不受交易者主观情绪的影响。计算机会严格执行交易策略代码,不涉及任何犹豫,贪婪和恐慌等情绪。而交易者在交易中不可避免的要受到自身主观情绪的影响,从而错过了止损或盈利的最佳时机。在交易过程中,速度对操盘手的要求也是很严格的。市场是瞬息万变的,如若下单速度比其他人慢,最佳的交易机会有可能就丧失了。而程序化交易的交易速度非常快,而且还可以24小时工作。对进行短线交易或高频交易的投资者来说,程序化交易非常具有吸引力。算法交易,是指把一个具有指定交易量、交易时间区间和价格区间约束的交易指令通过计算机程序来完成,用算法决定交易指令的递交时间、价格、数量和指令的类型。算法交易可以算是程序化交易在不断发展中的一个分支。因此,将算法思想融入到程序化交易中,不仅可以优化资源配置,也可以减少交易成本,获得较高的收益。本文先从程序化交易的起源出发,最早出现在1975年的美国证券市场。接着给出了程序化交易的NYSE定义,由于程序化交易已经经历的30多年的发展,依据现在市场上的环境重新定义了程序化交易,更贴近交易者的使用习惯,更适用于市场化的发展意义。接着介绍了国内外程序化交易的发展状况,可以了解国内的发展趋势,对投资者进行程序化交易有一定的参考作用。通过介绍分析程序化交易的优点和应用领域可以了解到程序化交易深受交易者喜爱的原因。同时,本文着重介绍了程序化交易的交易模型,对每个模型的原理,应用规则和模型的不足,都做了介绍分析,这为下面加入算法模型思想做了铺垫。紧接着,本文介绍了算法模型的思想和常用的算法交易的模型。通过介绍算法交易的定义和应用领域,可以帮助投资者进一步的了解算法交易。最后是本文的创新部分。基于对常用的算法交易模型进行了改进,并将其加入到了程序化交易模型中,得到了一个程序化算法模型。这个模型终合了程序化交易模型的特点和算法交易模型的特点,是对程序化交易模型的一种改进,使程序化交易更好地捕捉到市场的信息和价格,获得较高的利润。
[Abstract]:With the continuous development and improvement of the securities market and the procedural trading platform in recent years, the programmed trading plays an increasingly important role in the securities trading market. A programmed transaction is the process of executing a set program through a computer. The computer program is the trading thought and model of traders. Therefore, programmed transactions depend on computer trading platforms and trading models. Since the development and development of the domestic procedural trading platform has been relatively perfect, the research and analysis of the transaction model has become the key factor affecting the procedural trading. The biggest advantage of programmed trading is that it is completely independent of traders' subjective emotions. The computer strictly executes the trading policy code without any hesitation, greed, or panic. Traders are inevitably influenced by their own subjective emotions, thus missing the best time to stop loss or profit. In the course of trading, the speed of the traders is also very strict requirements. The market is rapidly changing, and if orders are placed at a slower pace than others, the best trading opportunities are likely to be lost. Programmed transactions are very fast and can work 24 hours a day. For short-term trading or high-frequency trading investors, program trading is very attractive. Algorithmic transaction means that a transaction order with specified trading volume, transaction time interval and price interval constraints is completed by a computer program. The algorithm is used to determine the delivery time, price, quantity and type of the order. Algorithmic transaction can be regarded as a branch of programmed transaction. Therefore, integrating the idea of algorithm into programmed transaction can not only optimize the allocation of resources, but also reduce the transaction cost and obtain higher income. Starting from the origin of programmed trading, this paper first appeared in the American stock market in 1975. Secondly, the NYSE definition of programmed trading is given. Due to the development of programmed trading for more than 30 years, it redefines programmed trading according to the current market environment, which is closer to the usage habits of traders. More suitable for the development of market-oriented significance. Then it introduces the development situation of the domestic and foreign programmed transactions, and can understand the domestic development trend, which has a certain reference role for investors to carry out the programmed trading. By analyzing the advantages and applications of programmed trading, we can find out the reasons why programmed trading is popular with traders. At the same time, this paper mainly introduces the transaction model of programmed transaction. The principle of each model, the application rules and the shortcomings of the model are introduced and analyzed, which lays the groundwork for adding the idea of algorithm model to the following. Then, this paper introduces the idea of algorithm model and the commonly used algorithm transaction model. By introducing the definition and application field of algorithmic trading, we can help investors to know more about algorithmic transactions. The last part is the innovation of this paper. Based on the improvement of the commonly used algorithm transaction model, a program algorithm model is obtained by adding it to the program transaction model. This model is an improvement on the programmed transaction model, which can capture the market information and price better and obtain higher profit.
【学位授予单位】：山东大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：F224;F830.91

【参考文献】