Autonomous HVAC Control, A Reinforcement Learning Approach
Recent high profile developments of autonomous learning thermostats by companies such as Nest Labs and Honeywell have brought to the fore the possibility of ever greater numbers of intelligent devices permeating our homes and working environments into the
- PDF / 351,184 Bytes
- 17 Pages / 439.37 x 666.142 pts Page_size
- 30 Downloads / 249 Views
2
Schneider Electric, Cityeast Business Park, Galway, Ireland Schneider Electric, 800 Federal Street, Andover, MA 01810-1067, USA {Enda.Barrett,Stephen.Linder}@schneider-electric.com http://www.schneider-electric.com/
Abstract. Recent high profile developments of autonomous learning thermostats by companies such as Nest Labs and Honeywell have brought to the fore the possibility of ever greater numbers of intelligent devices permeating our homes and working environments into the future. However, the specific learning approaches and methodologies utilised by these devices have never been made public. In fact little information is known as to the specifics of how these devices operate and learn about their environments or the users who use them. This paper proposes a suitable learning architecture for such an intelligent thermostat in the hope that it will benefit further investigation by the research community. Our architecture comprises a number of different learning methods each of which contributes to create a complete autonomous thermostat capable of controlling a HVAC system. A novel state action space formalism is proposed to enable a Reinforcement Learning agent to successfully control the HVAC system by optimising both occupant comfort and energy costs. Our results show that the learning thermostat can achieve cost savings of 10% over a programmable thermostat, whilst maintaining high occupant comfort standards. Keywords: HVAC control · Reinforcement learning · Bayesian learning
1
Introduction
Thermostats for controlling Heating, Ventilation and Air Conditioning (HVAC) systems in the home and office can largely be broken into two main categories: programmable and manual. Programmable thermostats allow the user to schedule heating and cooling to achieve patterns that work best for one’s schedule. A thermal set-point is specified by a user and it governs the temperature and humidity levels that must be reached when the controller is active. Manual thermostats are non-programmable and require an external operator (human) to turn on and off the functions of heating and cooling as required. Manual thermostats are usually cheaper than their programmable counterparts. Recently there has been a surge in the development of intelligent thermostats which boost the ability to autonomously control HVAC systems. These include c Springer International Publishing Switzerland 2015 A. Bifet et al. (Eds.): ECML PKDD 2015, Part III, LNAI 9286, pp. 3–19, 2015. DOI: 10.1007/978-3-319-23461-8 1
4
E. Barrett and S. Linder
offers from companies such as Nest Labs[3] and Honeywell[2]. They often only require the user to enter temperature set-points, while the schedule is learned automatically, with the objective of minimizing energy consumption while still allowing for occupant comfort. The unit attempts to learn a user’s preference over time based on their manual adjustments and produce a schedule which is deemed optimal for the observed patterns of occupancy. The principal characteristics of these units is that they promote some no
Data Loading...
