Automated financial trading platform

Abstract

Trading electronically via algorithms is a growing trend nowadays. This project is aimed to provide a platform for algorithmic research. There are various components involved in making an automated trading platform, such as data management, portfolio management, order management and algorithm management. These components hinder the development of an algorithm, because they require a lot of development work.

This project separates the algorithms from other components and provides an application user interface to develop algorithms without worrying about other components. Unlike other softwares, which provide in-built algorithms and do not allow any extension, this software enables any user to design and develop their algorithms while using the functionality offered by other components in this system. These algorithms can be added to the system very efficiently.

Using Interactive Brokers API for Java, this software was designed, implemented and tested to enable algorithmic trading. The design ensured the components remain independent and replaceable. Implementation was done while dealing with all the limitations of the broker as well as the market. Testing ensured that the system behaved as intended.

As a result, a software prototype was prepared for use. This software provided all the intended functionality and was also tested with a heuristic version of passive aggressive mean reversion algorithm. In future, this software may be expanded to include more functionality such as portfolio analysis and extensive backtesting.