Residential Demand Response under Uncertainty
Paul Scott
ARTIFICIAL INTELLIGENCE SEMINAR PhD monitoringDATE: 2013-05-29
TIME: 12:00:00 - 12:30:00
LOCATION: NICTA CRL Boardroom (level 2)
CONTACT: JavaScript must be enabled to display this email address.
ABSTRACT:
Real-time pricing provides an incentive for electricity consumers to actively participate in balancing the electricity network. This can help delay the need for expensive network upgrades and enable more renewables and electric vehicles to be connected. Such a market raises an optimisation problem for home automation systems where they need to schedule consumption activities to reduce costs, whilst maintaining a base level of comfort and convenience for occupants. This optimisation problem faces uncertainty in real-time prices, weather conditions, and occupant behaviour. This talk will present an online stochastic combinatorial optimisation algorithm that produces fast, high-quality solutions to this problem. This algorithm is compared with reactive control strategies and an approach using an expected scenario. Our results demonstrate the value of stochastic information and online stochastic optimisation in residential demand response.
