With the globalization of the economy, the challenges of financial risk management continue to grow, and the current traditional algorithms are often limited by the lack of search capability and diversity maintenance, which makes it difficult to predict as well as manage financial risks. Therefore, a multi-population multi-objective root system growth algorithm is proposed. The algorithm uses the plant root tip position and growth state as heuristic information to guide the search process. It also introduces adaptive search space to adjust the parameters, a multi-swarm strategies to enhance the exploration ability, and multi-objective optimization to adjust the weight balance among the objectives. The experimental results showed that in the single objective optimization function, the mean value of RSGA model was 5.80E-20, the standard deviation was 1.29E-19, the best position was 2.90E-26, and the worst position was 2.89E-19. In the biobjective optimization function, the average IGD of RSGA model was 2.28E-3. In the three-objective optimization function, the average IGD and HV of RSGA model were 1.05E-1 and 6.53E-1 respectively. In financial risk prediction, the best risk of RSGA model in small-scale investment was 0.1961, the worst risk was 0.2483, and the average risk was 0.2236. The best risk of medium-scale investment was 0.3057, the worst risk was 0.3387, and the average risk was 0.3194. In large-scale investment, the best risk was 0.191, the worst risk was 1.8795, and the standard deviation was 0.3769. Under MV portfolio, the maximum HV value of RSGA model was 1.13E-1, the minimum HV value was 4.20E-1, the average value was 8.74E-1, and the standard deviation was 5.46E-1. Under the RRC portfolio, the maximum HV of RSGA model was 1.49E-0, the minimum was 3.63E-1, the average was 8.17E-1, and the standard deviation was 3.95E-1.
Read full abstract