Algorithmic Trading Methods, 2nd Edition

Chapter 19: Machine Learning and Trade Schedule Optimization

Introduction

In this chapter we provide a technique to improve multiperiod trade schedule optimization solution times. The approach is based on a machine learning and neural network (NNET) methodology and provides nonlinear optimizers with a better initial parameter value that can be used as a starting point for the algorithmic trade schedule optimization problem. This technique results in 30%–75% faster optimization speeds. The usage of machine learning and NNETs in conjunction with the trade schedule optimization problem was previously studied by Kissell and Bae (2018).

In the chapter on Portfolio Algorithms and Trade Schedule Optimization, we presented different techniques ...