In order to accurately determine the auto insurance rate of UBI, this paper proposes to use fuzzy controller to calculate the rate and optimize it by using the simulated annealing particle swarm algorithm with Metropolis criterion. Firstly, a fuzzy controller is constructed by selecting monthly mileage and violation times to calculate the self-underwriting coefficient. In order to eliminate the subjectivity defect of fuzzy controller, the correlation function of independent underwriting coefficient and historical risk data is proposed as the fitness function of evaluating fuzzy rules, using adaptive simulated annealing particle swarm optimization algorithm is intelligent search, according to the fitness value of continual iteration and optimize the optimal fuzzy rules. Finally, the fuzzy controller is reconstructed with the optimal fuzzy rules to estimate the auto insurance rate accurately. The results show that the adaptive simulated annealing particle swarm optimization algorithm can effectively extract the driving behavior information and can calculate the more reasonable and accurate autonomous underwriting coefficient. The results are highly correlated with the number of historical accidents and have the ability and stability of risk quantification.

中文翻译：

---

基于自适应SAPSO优化模糊控制器的UBI汽车保险定价模型研究。

为了准确确定UBI的汽车保险费率，本文提出采用模糊控制器计算费率，并采用基于Metropolis准则的模拟退火粒子群算法对其进行优化。首先，通过选择月度里程和违规时间来构造模糊控制器，以计算自保系数。为了消除模糊控制器的主观性缺陷，提出了独立承保系数与历史风险数据的相关函数作为评价模糊规则的适应度函数，采用自适应模拟退火粒子群算法根据适应度值进行智能搜索。连续迭代，优化最优模糊规则。最后，用最优模糊规则对模糊控制器进行重构，以准确估计汽车保险费率。结果表明，自适应模拟退火粒子群算法能够有效地提取驾驶行为信息，并能计算出更合理，更准确的自主承保系数。结果与历史事故的数量高度相关，并且具有风险量化的能力和稳定性。