Research on State-Owned Assets Portfolio Investment Strategy Based on Improved Differential Evolution

The state-owned assets portfolio problem is a nonlinear programming problem, and the traditional algorithm can not effectively find the optimal solution, so the effective solution has become a hot issue. Differential evolution algorithm has some disadvantages, such as slow convergence speed and easy to fall into local optimal solution. In this paper, a new differential evolution algorithm based on cluster analysis is proposed. The improved algorithm is used to make portfolio investment of state-owned assets. Firstly, the cluster analysis method is used to cluster the populations of the difference algorithm, extract the representative element individuals, replace the poor individuals in the original population with new individuals, remove the redundant information in the population, and optimize and update the population, so that the whole population can converge to the global optimal solution quickly and accurately. The experimental results show that the differential evolution algorithm with cluster analysis strategy can not only effectively inhibit premature convergence and improve convergence speed, but also has the characteristics of simplicity, efficiency and strong robustness. On the basis of satisfying the investment objectives and constraints, it also reflects the different kinds of income and risk needs of investors, and has good practice.