基于遗传-蚁群算法优化的电费风险预警模型的研究

Research on the early warning model of electricity tariff risk based on genetic⁃ant colony algorithm optimization

遗传-蚁群算法的“过早收敛”问题易导致电费风险预警反馈出现偏差,为此,利用种群方差和熵共同反映种群的多样性,改进遗传-蚁群算法的选择方式,设计了一种新的电费风险预警模型。根据定性和定量指标设置模型预警阈值,再根据设定的风险预警等级实现对电费风险的预警反馈。以不同阶段的反馈结果为实验测试内容,根据测试结果可知,与传统遗传-蚁群算法构建的预警模型相比,在一个完整的测试周期内,优化后的预警模型在不同阶段均有较准确的风险反馈结果,可见该模型的预警反馈效果较好。

The problem of"premature convergence"of the genetic⁃ant colony algorithm is likely to lead to the deviation of the early warning feedback of electric power risk.Therefore,the population variance and entropy are used to reflect the diversity of the population,and the selection method of the genetic⁃ant colony algorithm is improved to design a new electric power risk early warning model.According to the qualitative and quantitative indicators,the warning threshold value of the model is set,and then the warning feedback of the electric power risk is realized according to the risk warning level.Taking the feedback results of different stages as the experimental test content,it can be seen from the test results that compared with the warning model constructed by traditional genetic⁃ant colony algorithm,the optimized warning model has more accurate risk feedback results in different stages in a complete test cycle,indicating that the model has better warning feedback effect.