GPT Prompt Engineering for Scheduling Appliances Usage for Energy Cost Optimization

Marco Siino, Ilenia Tinnirello

Abstract: In this paper, we propose a novel approach that makes use of a GPT model and of prompt engineering to build a proper input to GPT, given a domestic energy dataset. Specifically, given a residential energy consumption dataset, we ask a GPT model - in order to reduce the cost of energy - for planning the timing usage of house appliances, while preserving the same utilization of each appliance on a daily basis. To the best of our knowledge, this is the first attempt to schedule appliances usage taking advantage of the planning ability of a GPT model. Thanks to this preliminary study, we highlight interesting results to be further investigated and enabling certain room for improvements in this domain.