• market-making
• Avellaneda-Stoikov model
• limit order book

Market makers are fundamental intermediaries in financial markets. On
a regular basis they propose prices at which they are willing to buy and sell
a certain asset, playing a fundamental role in the functioning of financial markets.
The seminal Avellaneda-Stoikov model gives a simplified framework
of market's dynamics, in which the optimal market making problem can be analytically solved.

The objective of this thesis is to study if, and how, the assumptions of the
Avellaneda-Stoikov model adapt to Limit Order Books (LOBs), digital queuing systems in which orders are stored, with the aim to facilitate trades between market participants.

In order to avoid typical problems of testing trading strategies on historical data, we decide to implement a LOB simulator, capable of accurately reproduce real LOB behaviour and responde to our orders.

We start this thesis by presenting the Avellaneda-Stoikov model in its details.
We then proceed by presenting the Ratio model, a LOB simulator that overcomes some limitation of the ZI model by considering many regularities of the order flow, i.e. the arrival of the different types of order. We then present our own LOB simulator, that takes into account the correlations in the order flow and show that it outperforms the Ratio model in many aspects.
We conclude by simulating the market making strategy in our LOB environment, highlighting the difference in its effectiveness between the stylized Avellaneda-Stoikov model and our realistic model of LOB.