数学建模方法对碳排放、碳达峰问题的研究

以河北省碳排放为研究目标,本文基于传统KAYA恒等式和STIRPAT模型,提出了一个改进的STIRPAT模型,并用于预测河北省碳排放以及碳达峰情况。首先使用ARIMA模型、灰色预测方法、长短期记忆(LSTM)网络预测模型分别对改进STIRPAT模型中的驱动指标进行了预测和比较,得到了各指标的最优预测方法。之后,结合时序型预测方法,分别采用BP神经网络回归、Lasso回归和岭回归方法分别对STIRPAT模型种的碳排放进行了预测。结果表明,随着第三产业占比的持续扩大和清洁能源占比的增加,河北省的碳排放总量将呈现下降趋势。其中Lasso回归方法具有最小均方误差和最优拟合优度,这种回归方法预测出河北省的碳排放总量在2035年将达到456.8百万吨,并在2023年左右达到碳排放峰值。河北省的碳排放预测对同类省份的碳排放情况给出了参考,这对于中国全国实现2035年碳达峰目标具有重要意义。

A Research on Demands Forecasting and Personnel Training of Tourism Talents A Case Study of Zhejiang Province

By analyzing the recent 15 years＇ statistical data of Zhejiang tourism human resources, this paper analyzes the status of Zhejiang tourism talents. ARIMA （Autoregressive Integrated Moving Average） model is a method of time series prediction. This paper predicts the trends of the next three years＇ demands of Zhejiang tourism talents based on ARIMA model in order to promote the tourism in Zhejiang Province. According to the demands forecasting, the number of the employees required by the hotels is 10 times of travel agencies in 2015. At last, some solutions and suggestions are provided such as strengthening the talents training establishing tourism talents mobility mechanism and improving tourism talents excitation mechanism