Decision support challenge: The main challenge is to collect, store, and analyze massive amounts of Smart Grid usage data, which takes time — and yet be able to make quick decisions that trade off various parameters while not relying on stale data. Another challenge is to make decisions when only a fraction of the data is available, because one should not wait months to collect enough data and make such decisions. Similarly, we have to recognize and purge wrong, bad, stale, or downright malicious data. Yet another challenge is to ensure privacy in this massive data collection without hurting the public good.
Decision makers or end users: Federal, State, and local governments can make informed decisions about their energy use (for example, is it more efficient to use cloud computing or local computing). Internet companies operating large data centers can better optimize their use (e.g., migrate expensive computation to a data center whose energy costs are lowest at the moment): this could reduce carbon footprints considerably as well as save millions of dollars in energy costs. Even organizations with substantial computing infrastructure, such as universities and hospitals, will benefit from this work. Eventually, these technologies can become commoditized and available to individual users (e.g., smart homes).
Research challenges relating to decision-making process: Energy efficiency is one of the most important challenges facing our planet, affecting the climate and much more. Future Smart Grids will have to ingest data and energy from potentially billions of sources of varying capabilities: the sheer scale of this problem requires massive data processing. Smart Grids will have to take into account user preferences, fluctuating energy costs, supply and demand projected for the next few minutes, hours, or days — and make informed economic decisions that optimize for individual as well as global needs. Collecting, storing, and processing all this data quickly will require new research and techniques.
Data science skills necessary to improve decision making: Students studying Smart Grids need a good understanding of math, electronics, and computer science — i.e., spanning what is normally three separate degree programs without tripling their educational burden. Students would need to be proficient in machine learning, data mining, big data sciences, visualization, and modeling. Students would also need a good understanding of basic economics so as to integrate proper business models.
Assessing improved decision making: One key metric would be how much energy is being saved overall (e.g., how much wasted energy could have been sent elsewhere, and whether a nation’s overall consumption is reduced). Another key metric is how quickly do we adapt to rapidly changing conditions and needs (e.g., energy cost fluctuations, natural disasters). A third metric would be how closely do we address users’ trade-off preferences.