2019/10/29 by Olivier Guéant, Guéant, Olivier, Iuliia Manziuk +1 · 3 citations
Economics, Econometrics and Finance · Engineering · #Stochastic processes and financial applications #Energy Load and Power Forecasting #Electric Power System Optimization
paper · pdf · doi:10.48550/arxiv.1910.13205
In corporate bond markets, which are mainly OTC markets, market makers play a\ncentral role by providing bid and ask prices for a large number of bonds to\nasset managers from all around the globe. Determining the optimal bid and ask\nquotes that a market maker should set for a given universe of bonds is a\ncomplex task. Useful models exist, most of them inspired by that of Avellaneda\nand Stoikov. These models describe the complex optimization problem faced by\nmarket makers: proposing bid and ask prices in an optimal way for making money\nout of the difference between bid and ask prices while mitigating the market\nrisk associated with holding inventory. While most of the models only tackle\none-asset market making, they can often be generalized to a multi-asset\nframework. However, the problem of solving numerically the equations\ncharacterizing the optimal bid and ask quotes is seldom tackled in the\nliterature, especially in high dimension. In this paper, our goal is to propose\na numerical method for approximating the optimal bid and ask quotes over a\nlarge universe of bonds in a model `a la Avellaneda-Stoikov. Because we aim at\nconsidering a large universe of bonds, classical finite difference methods as\nthose discussed in the literature cannot be used and we present therefore a\ndiscrete-time method inspired by reinforcement learning techniques. More\nprecisely, the approach we propose is a model-based actor-critic-like algorithm\ninvolving deep neural networks.\n