| In this thesis,dynamic asset allocation optimization is carried out by using model predictive control(MPC),which is based on non-parametric Bayesian method,i.e.,hierarchical Dirichlet process,to predict the future mean and variance of return.Because of its rolling optimization,the model predictive control method not only accords with the dynamic optimization investment idea of updating investment decision in each period,but also has the simplicity of calculation.The constraints of transaction cost and holding cost are added to the model to reduce the estimation error.At the same time,the maximum drawdown is used to control the risk aversion condition of investors.On the basis of further improving the maximum drawdown control function,the allocation strategy can be applied more flexibly by using the time-varying parameters.For the selection of individual stocks,asset allocation is influenced by many factors in reality,especially market and industry changes.Research on industry allocation reflects style investment trend,and the allocation of large-category assets can give investors a grasp of the whole market.In this paper,the simulated data estimation model are first used,and then ShangHai Industrial index are used for empirical tests,compared with the 1/N strategy、the selected actual benchmark index and strategy with HMM prediction.Adjusting the risk aversion coefficient of investors can control the actual drawdown value,and adding leverage can increase returns without increasing the maximum drawdown.From 2015 to 2016,the market experienced significant fluctuations.After studying the asset allocation around the period,the results obtained by the model are still consistent with the return and risk of the index,but the return of the model exceeds the benchmark return、equal-weighted allocation scheme and strategy with HMM prediction by far.In this thesis,by using the rolling optimization idea,model predictive control using hierarchical dirichlet prediction model on the basis of the optimal dynamic asset allocation plan,changing the condition of maximum drawdown,provides investors with a relatively simple and quick investment strategies,in real life we need to make informed decisions on the basis of the condition of the market macro to micro. |
