Research And Application Of Portfolio Optimization Based On Improved Sine Cosine Algorithm

Investing is a risk and reward activity.In the financial securities investment market,many uncertainties make the market unpredictable and risky.An effective portfolio can help investors obtain stable investment returns and diversify risks,thus the investment portfolio has been widely studied in the field of finance and economics.Portfolio problems in real-world often contain multiple constraints and objectives,making it a highly complex,non-linear problem.Traditional mathematical optimization methods such as weighted sum method,linear planning method,Newton’s method,gradient method cannot effectively solve the problem,but simple,efficient and flexible intelligent optimization algorithm are widely used to solve such problems,as they often yield good results.In this paper,two types of portfolio optimization problems are investigated,and a new intelligent optimization algorithm,the Sine Cosine Algorithm(SCA),is improved and applied to solve the portfolio model.The main contents include:A Hybrid Beetle Antennae Search and Sine Cosine Algorithm(BASSCA)based on nonlinear inertia weight is proposed for solving the portfolio problem with base constraint.The nonlinear inertial weights are introduced to effectively balance the global exploration and local development of the algorithm.The convergence performance of the sine cosine algorithm is also improved by perturbing the individual optimal solution by the beetle antennae algorithm.The classical test function experiment shows that the proposed algorithm effectively improves the global optimization capability of SCA.The experimental results from real market data shows that the proposed algorithm has better optimization performance and applicability than other intelligent optimization algorithms in solving portfolio problems.For the portfolio problem with transaction fees,a Sine Cosine Algorithm Based on Cauchy Variation and Self-Learning Mechanism(SLCSCA)is proposed and used for this problem.To address the shortcomings such as insufficient search strategy and the tendency to fall into the local optimum,a Cauchy mutation and self-learning mechanism are introduced,which make use of the powerful perturbation ability of the Cauchy mutation to operate on the optimal individual and enhance the global search ability of the algorithm;the self-learning mechanism effectively utilizes the optimal information of the particles,improves the diversity of the population and prevents falling into the local optimum.The effectiveness of the improvement strategy is verified through 12 benchmark function tests.Based on real market data and comparing the experimental results of the four optimization algorithms,the improved algorithm is able to obtain a portfolio with lower risk value more quickly.The experiments show that the improved algorithm is effective in solving the portfolio problem.