Automated tariff design for energy supply–demand matching based on Bayesian optimization: Technical framework and policy implications

With the emergence of renewable energy sources, designing tariffs becomes crucial to match unstable energy supply with varying energy demand. However, traditional tariff design frameworks depend on a certain type of tariff and the demand response of customers, which makes designing tariffs rigid and hard to automate. To overcome the challenges, in this paper, we propose an automated tariff design framework for energy supply–demand matching based on Bayesian optimization. The proposed framework allows us to design an optimized tariff for any given objective of energy matching automatically. This automation can provide flexibility in designing tariffs for energy management. Moreover, we propose a novel matching evaluation metric focusing on the mismatch of actual energy supply–demand patterns related to the stability and reliability of energy management systems. Adopting it along with traditional economic utility, actual energy pattern matching and economic values can be deliberately balanced in the automated tariff design. Through simulations with real datasets, we demonstrate that the proposed framework can effectively design tariffs in the real world such as time-of-use and demand charge tariffs, considering energy patterns and economic costs separately. Finally, we provide potential policy implications resulting from the automated tariff design.