DLC Portfolio Optimization under Weather Risk

Abstract

This paper describes a kind of portfolio selection problem under weather-risk that arises in development of DLC (Direct Load Control) execution plan, mainly for the coordinated power curtailment of office buildings' HVAC systems. Inspired by the Markowitz model of investment theory, we invented two-stage approximate means for scalable solution of the optimization problem. A functional relation between expected value of demand reduction and corresponding minimized risk, which is an analog of efficient frontier in investment theory, is plotted and analyzed its behavior along with selected portfolio's constraints. A numerical experiment is done to show the proposed method can improve reliability of HVAC demand reduction compared to non-optimized portfolios.