基于改进果蝇算法优化GRNN的碳排放权价格研究

Study on the Carbon Emission Price of GRNN Based on Improved Fruit Fly Algorithm

随着国内外碳排放交易机制的不断发展，研究国际碳市场排放权的价格对我国碳市场价格的研究具有重要意义。为了提高国际碳价的预测精度，首先用自适应Lasso算法进行变量选择和参数估计，接着用改进的果蝇优化算法（IF-OA）优化广义回归神经网络（GRNN）的平滑参数来预测国际碳价。仿真结果表明，该模型的收敛速度和预测精度优于传统的GRNN和FOA优化的GRNN模型，说明基于IFOA优化的GRNN算法可以有效降低平滑参数设置的盲目低效性，改善原模型易陷入局部极值的问题，从而提高国际碳价的预测精度。

With the development of carbon emission trading mechanism, it is of great significance to study the price of carbon emission in the international carbon market for the study of the carbon market in China. In order to improve the prediction accuracy of international carbon price,variable seleetion and parameter estimation are firstly carried out by using adaptive Lasso algorithm. Then, the international carbon price is predicted by generalized regression neural network （GRNN） optimized by the improved fruit fly optimization algorithm （IFOA）. The simulation results show that the convergence rate and prediction accuracy of this model are better than the traditional GRNN and FOA - GRNN model. It shows that the GRNN algorithm based on IFOA optimization can effectively reduce the blind inefficiency of smoothing parameter setting, and improve the problem that the original model is easy to fall into the local extreme, which can raise the forecast accuracy of the international carbon price.