The virtualized thermostat reads very accurate input data, which it uses in predictive ways, but the actual response output is tuned and tempered by Bayesian interpretation algorithms and factors particular to the individual consumption instance. Solving mixed integer nonlinear equations using second order variables within a scalable model with input from local and networked sources and passing both complex and “simplified user understood data” requires a two model approach in a cloud based managed solution where one model keeps track of how the system operates in the physical world and the second model communicates with other virtualized devices and computes predicted responses and tracks comparisons.

Benefits of Implementation 1:

By its very nature, a virtualized thermostat can accept input from all types of sources and respond in a complex way, according to specifically tuned and/or predictive (Bayes's) responses.

A virtualized scalable model can contain information that can be presented in any form necessary to interface with simple human interaction or complex machine to machine interface. This allows for simple local user interface with complex control algorithms handled off site in the cloud environment.

By placing highly accurate sensors, it is not only necessary, but possible to enable the use of 2nd order equations to represent complex virtualized interactions and behaviors.

By using a sensor that reads both surface temperature and localized surrounding temperature, more accurate comfort decisions can be made to improve comfort and reduce energy consumption.