Building users are key participants in demand response (DR) markets, providing significant flexible resources. Due to uncertainty in market clearing prices, various risk-based decision models have been developed to describe their bidding behavior, typically assuming constant risk preferences. However, empirical evidence indicates that users’ risk attitudes vary with the magnitude of load adjustments. To capture this feature, this paper introduces mental accounting theory to model the risk-aware bidding behavior of building users. Total response capacity is divided into three independent mental accounts based on air-conditioning setpoint adjustment magnitude, representing risk-averse, risk-neutral, and risk-seeking behaviors. This framework allows multiple risk preferences to be represented within a unified bidding model. For each account, response quantity and cost models are developed. Bidding strategies under uncertain market clearing prices are formulated by incorporating loss aversion. A multi-agent simulation framework, including building users, a load aggregator, and a grid operator, is established to simulate the market clearing process. A simulation study is conducted using 19 building clusters located in Zhuhai, China. The proposed model is compared with a single-bid model and a step-wise bidding model with constant risk preferences. The results show that it better captures building users’ multiple risk preferences under market clearing price uncertainty. Users tend to secure stable returns through responses with minimal comfort loss, while pursuing excess profits via higher bids for responses involving greater comfort sacrifices.
Deng et al. (2026) studied this question.