Optimization research on operational strategies of energy use in building clusters have generally marginalized the effects of uncertainty in favor of reduced computational expense. This however leads to a significant disconnect between the expected energy cost and the average cost observed under uncertainty. Bridging this divide requires the incorporation of uncertainty analysis which poses both technical and computational challenges. This paper addresses these challenges through the notion of a Pareto band, demonstrating its applicability towards developing resilient operational strategies in a timely and computationally efficient manner. Under the proposed approach, Monte Carlo simulations are leveraged to reveal an envelope of optimality contained within the energy cost solution space. This optimality envelope, formally introduced as a Pareto band, is then used to train generalized linear models (GLMs) enabling robust operational strategy predictions. The results obtained from this approach highlight significant improvements in energy cost performance under uncertainty.

This content is only available via PDF.
You do not currently have access to this content.