基于仿射跳跃-扩散过程的电力市场电价随机模型

基于金融工程理论的电价随机模型对于竞争性电力市场中的电力衍生产品定价、风险管理、资产定价及不确定条件下的投资决策等具有基础意义。文中提出了基于仿射跳跃-扩散过程的、具有2个和3个跳跃分量的电价随机模型以及一种新的参数标定方法。所提出的近似参数标定方法可利用历史电价数据快速求解模型参数，且计算量与电价样本点数量无关。由于直接根据历史电价的数字特征求解，不存在误差累积的问题。根据法国Powernext、德国EEX和荷兰APX等3个主要欧洲电力市场的历史日前小时电价数据标定了模型参数并采用Monte Carlo方法分析了模型的模拟效果。计算表明，文中提出的电价随机模型能够较为准确地描述电价的整体概率分布，满足进一步研究的需要。

Stochastic Electricity Price Models Based on the Affine Jump-diffusion Process

In the competitive electricity market, the stochastic electricity price model originated form financial engineering theory is crucial to the pricing of the electricity derivatives, the electricity market risk management, the appraisal of asset and the investment decision under uncertainty. This paper proposes the stochastic electricity price models based on the affine jump- diffusion process with two or three jump components and a novel approximate parameter calibration method, with which the model parameters can be solved quickly using the historical electricity price data and the calculation needed will not increase with the number of the historical electricity prices. This calibration method will not cause the accumulation of computational error because it solves the model parameters directly from some numerical characteristic values of the historical electricity price. The calibration results for the Powernext, EEX and APX market using hourly spot prices are presented and Monte Carlo simulations are performed to validate the effectiveness of the model. The results illustrate that the calibrated models are capable of depicting the probability distribution of the hourly electricity price accurately and provide a useful tool for further research.