Importance Scaling for Elastic Appliance for Automated Power Management in Smart Homes

Abstract

This paper examines the effect of importance scaling of individual appliance utility functions in an automated energy management system, considering only appliances which have continuously variable energy consumption settings. These appliances are divided into three categories along the memory property of their utility functions, which are described as full-memory, partial memory, and memoryless appliances. The energy management system is formulated as a convex optimization problem which is solved daily to determine optimal appliance settings for the day. This paper provides a weighting-based mechanism to the overall utility function to scale for the relative importance of individual appliances. Our results indicate greater user satisfaction can be achieved by parameterising the objective function around appliance importance.