Multi-asset allocation based on financial market microstructure model

Financial market microstructure model is a phenomenon model which describes financial markets based the microstructure theory. The initial values of the unknown parameters and/or states of the model have a great impact on the model identification, so an estimation method which combines genetic algorithm, Kalman filter and maximum likelihood method is presented to estimate the unknown parameters and/or states of the microstructure model. Based on the identified model and the indirectly obtained market excess demand instead of the prediction for price, a dynamic multi-asset allocation strategy is proposed. Case analysis for a combination asset of two stocks and currency shows that the total assets under the control of dynamic allocation strategy are much more than those without allocation control, which proves that the proposed parameter estimate and asset allocation method are feasible and effective.