中国电机工程学报2026，Vol.46Issue(14)：5794-5807,中插8,15.DOI:10.13334/j.0258-8013.pcsee.250570
考虑用户禀赋效应的需求响应增量激励策略
Incremental Incentive Strategy for Demand Response Considering User Endowment Effect
摘要
Abstract
Developing demand response strategies for users is one of the effective ways for electricity retailers to cope with price risks in the spot market.However,existing demand response strategies have not fully considered the differences in user response elasticity and psychological behavioral characteristics,resulting in an imbalance in consumer surplus allocation and limited incentive effects.Therefore,this article proposes a demand response incremental incentive strategy that considers the differences in user endowment effects and response elasticity.Firstly,by introducing endowment effect factors to quantify the subjective psychological cost of users' existing electricity rights,a time-varying incremental incentive function that dynamically matches the marginal cost and psychological loss of users is constructed to avoid excessive allocation of consumer surplus in the initial response stage.Secondly,to resolve the asynchronous coupling issue between the retailer's day-ahead bidding and intraday incentive strategies,this paper designs a novel asynchronous coupling solution method based on curriculum learning-enhanced twin delayed deep deterministic policy gradient(TD3)algorithm.This method progressively learns complex high-uncertainty environments—such as user behavior and market electricity prices—from simple to challenging stages,achieving coordinated optimization of asynchronous coupling problems.Theoretical analysis shows that the proposed incremental incentive strategy can reduce the unit incentive cost of electricity retailers and enhance the depth of high elasticity user response.Finally,the simulation results indicate that the proposed strategy dynamically redistributes consumer surplus by following changes in user marginal costs,reducing excessive consumer surplus in the initial response phase and achieving fair allocation of incentive resources.In addition,the proposed algorithm has significantly improved convergence compared to the deep deterministic policy gradient(DDPG)algorithm,and its convergence speed has increased by 21.88％compared to the TD3 method without curriculum learning.关键词
禀赋效应/深度强化学习/响应弹性/异步耦合优化/增量激励策略Key words
endowment effect/deep reinforcement learning/response elasticity/asynchronous coupling optimization/incremental incentive strategy分类
信息技术与安全科学引用本文复制引用
华昊辰,贾韵思,陈星莺,宫家凯,刘迪,余昆,甘磊..考虑用户禀赋效应的需求响应增量激励策略[J].中国电机工程学报,2026,46(14):5794-5807,中插8,15.基金项目
国家重点研发计划项目(2022YFE0140600) （2022YFE0140600）
国家自然科学基金项目(52377093).National Key R&D Program of China(2022YFE0140600) （52377093）
Project Supported by National Natural Science Foundation of China(52377093). （52377093）
