Dynamic Pricing Model Of Power Sales Companies Based On Bi-level Optimization

The issuance of “Electrical Reform No.9” marked the beginning of a new round of power reform.The focus of the reform is liberalization for the power sale side.The power sales company emerged as the emerging market entity for this reform,and the pricing of electricity power was the key factor for winning in the competitive electricity market,reasonable pricing can reduce electricity market risks and maximize own revenue,which is also the most important part of the sales cycle of the power sales company.This topic carries out the following research on the technical problems existing in the dynamic pricing of power sales companies under the new electricity market.Aiming at the risks brought about by the purchase-sales decision-making and the spot market volatility involved in the pricing model,a pricing risk assessment method for power sales companies based on CVaR is proposed.First,the power sales business and the power sales business of the power sales company are modeled separately;secondly,the pricing risk assessment function of the power sales company is established by introducing the conditional risk value;the final experiment simulates the risks caused by several sets of business decisions,and sells the electricity.The company's comprehensive utility optimal operation method validates the effectiveness of the pricing risk assessment method.Aiming at the problem of uncertainty in the parameters of sales power involved in the pricing model,a method of forecasting the sales volume of electricity sales companies considering residuals was proposed.Firstly,the growth trend of sales electricity is modeled by nonlinear regression analysis,and the mapping relationship between the influencing factors and the sales volume is constructed.The least squares method is used to estimate the nonlinear regression equations,and the result set of the prediction sequence is obtained;Minimize the method of calculating the variance of the fitted variance,and calculate the residual sequence weights of the RBF neural network and the least squares vector machine to obtain the residual sequence result set.Finally,based on the cuckoo algorithm,the inertia weight is reduced nonlinearly.Strategies,optimization of model parameters for extended search space and model solution.The reliability of the prediction model was verified by experiments.Based on the above research,a double-layer optimized pricing model for power sales companies based on demand response was designed.Firstly,the two-level model is constructed,and the comprehensive utility function of conditional risk value is introduced as the upper optimization model.The target function is the maximization of revenue from sales companies,and the demand curve between electricity price and demand response users is used as the lower optimization model.The objective function is electricity.The user's power utility is maximized;then the lower model is transformed into the upload model constraint by the KKT condition;the CPLEX solver is used to solve the model.The double-layer optimization model is validated by the data from the California market.The experimental results show that considering the dynamic interaction between the power sales company and the power customer can provide a better real-time power price for the power sales company.The work of the paper focuses on the new power.New problems in the market provide guidance for the pricing of power sales companies.