1. Preliminaries

The concept of order placement requires a brief understanding of the inner-workings of a financial market. Hence, this Section provides insight into the context of our problem domain including the artefacts Order book, Match Engine and Order Execution as well as the related evaluation technique generally known as Backtesting. We further make note of time series, from which type of data our problem domain is derived from. Lastly, this Section introduces Reinforcement Learning and in particular Deep Reinforcement Learning, an area of Machine Learning inspired by behaviourist psychology, which is the preferred method of this thesis.

1.1 Order book

1.2 Match Engine

1.3 Order Execution

1.4 Backtesting

1.5 Time Series

1.6 Reinforcement Learning