基于模拟退火优化的海马齿生长参数确定

为确定海马齿生长的相关参数，对海马齿进行了室内水培实验，连续监测海马齿生长及生态因子状况。考虑营养盐循环，基于一级衰减模式构建海马齿生长模型。使用基于单纯形实验设计的Morris法进行全局定性灵敏度分析发现，对所有状态变量均敏感的参数为氨氮硝化率和有机氮矿化常数，表明所构建模型主要由氮的循环系统所支配；最适光照是对海马齿生长最为灵敏的参数，光照是影响海马齿生长的最主要生态因子。以各状态变量的最大均方误差为代价函数，使用模拟退火算法对模型参数进行优化及确定，选取多指标对模型进行评估。结果表明，模拟值可以很好地拟合实测值，最大平均绝对百分误差为5.023%；对参数分析发现，海马齿对硝酸盐氮的吸收具有一定的偏向性，对磷具有较高的耐受性；海马齿生长速率较快，具有一定的开发前景。

Parameter determination of Sesuvium portulacastrum based on simulated annealing

In order to determine the parameters related to the growth of Sesuvium portulacastrum, hydroponic experiments are performed on the S. portulacastrum to continuously monitor the growth of S. portulacastrum and the status of ecological factors. Considering nutrient cycling, the model of S. portulacastrum is constructed based on first order attenuation pattern. The global qualitative sensitivity analysis using Morris method based on the simplex experimental design shows that the parameters sensitive to all state variables are ammonia nitrogen nitrification rate and mineralization constant of organic nitrogen, indicating that the model is mainly dominated by the nitrogen circulation system. The optimum illumination is the most sensitive parameter for the growth of S. portulacastrum, and light is the most important ecological factor that affects the growth of the S. portulacastrum. Based on the minimum mean squared error, the simulated annealing algorithm is used to optimize and determine the parameters of the model, and multiple indicators are selected to evaluate the model. The results show that the simulated values can be well fitted with the measured values, and the maximum mean absolute percentage error is 5.023%. It is found that the absorption of the S. portulacastrum to NO₃-N has a certain bias to the parameter analysis, and it has high tolerance to phosphorus. The growth rate of S. portulacastrum is fast, and it has a certain development prospect.