Agent-based simulation of renewable electricity consumption under dual policy mechanisms: A complex network and social learning perspective

Thermal power still dominates overall electricity consumption, whereas renewable electricity, though essential for decarbonization, remains limited in market acceptance. Therefore, identifying the drivers of renewable electricity adoption is crucial for facilitating the sustainable transition. This study investigates electricity market participants’ decision-making and the dynamics of renewable electricity consumption, considering climate policies, generator competition, and consumer social learning. Specifically, we develop a two-layer dynamic decision-making framework that incorporates carbon emission trading (CET) and Renewable Portfolio Standard (RPS). The upper layer captures price competition between heterogeneous power generators, while the lower layer models consumer interactions within networks through social learning. Simulations are conducted based on agent-based modeling and multi-agent reinforcement learning. The results indicate that social learning raises thermal power price, accompanied by reduced thermal power consumption and expanded renewable electricity consumption. The profits of renewable energy generators rise significantly. Moreover, social learning not only raises the share of renewable electricity and enhances social welfare but also reduces carbon emissions and amplifies the policy effects of CET and RPS. This study reveals the evolutionary dynamics of renewable electricity consumption in complex environments, offering valuable theoretical insights and practical implications for advancing energy transition.