International Journal of Service and Knowledge Management, vol.7, no.2, p. IJSKM770, 2023
We created and analyzed various Demand Response (DR) scenarios in the electric power market using multi-agent simulation. We first built a multi-agent simulation for the electric power market using actual data, including prices from the Japanese Electric Power Exchange (JEPX) market, electricity consumption in Japan, and an actual factory. Using this multi-agent simulation, we tested several possible DR scenarios for the factory. We then compared these scenarios using two newly defined indices for assessing the reduction efficiency of cost and CO2 emissions. The results showed that a work time shift in the summer and peak shift in factory demand in the winter were the best in terms of cost and CO2 emission reduction efficiencies. Thus, we demonstrated the usefulness of our multi-agent simulation for examining the effectiveness of DR scenarios by simulating complex interactions that consider the seasonal and time-of-day characteristics of power prices.
Language models; Domain-specific pre-training; Financial market; Natural language processing;
@journal{Hirano2023-ijskm,
title={{Scenario Analysis of Demand Response Using Artificial Electric Power Market Simulations}},
author={Masanori HIRANO and Ryo WAKASUGI and Kiyoshi IZUMI},
journal={International Journal of Service and Knowledge Management},
volume={7},
number={2},
pages={IJSKM770},
doi={10.52731/ijskm.v7.i2.770},
year={2023}
}