Factored Models for Multiscale Decision-Making in Smart Grid Customers

Active participation of customers in the management demand, and renewable energy supply, is a critical goal Smart Grid vision. However, this complex problem with numerous scenarios that are difficult to test field projects. Rich scalable simulations required develop effective strategies policies elicit desirable behavior from customers. We present versatile agent-based "factored model" enables rich simulation across distinct customer types varying agent granularity. formally characterize decisions be made by as multiscale decision-making show how our factored model representation handles several temporal contextual introducing novel "utility optimizing agent." further contribute innovative algorithms for (i) statistical learning-based hierarchical Bayesian timeseries simulation, (ii) adaptive capacity control using decision-theoretic approximation multiattribute utility functions over multiple agents. Prominent among approaches being studied achieve active one based on offering financial incentives through variable-price tariffs; we also an solution "customer herding" under such tariffs. support contributions experimental results real-world data open platform.