Optimal Algorithms for Energy Markets

The deregulation of the energy industry has createdmarketplaces in which producers and consumers canoperate in order to more effectively. This bookdevelops a new model for electricity retail wherecustomers may choose their supplier. It is based onsimultaneous reverse combinatorial auctions, designedas a second-price sealed-bid multi-item auction withsupply function bidding. This mechanism preventsstrategic bidding and allows the auctioneer tomaximise its pay-off. Furthermore, we develop optimalsingle-item and multi-item algorithms for winnerdetermination in such auctions that are significantlyless complex than those currently available in theliterature. Nevertheless, the consumption of theenergy of each singular auctioneer has to be adaptedto the submitted bids to maximise the pay-off. Thus,this work models the constellation of energyconsumers as a distributed constraint optimisationproblem (dCOP). In order to overcome the specificdomain demands, this book presents a novel optimaldCOP algorithm called COBB (Constraint OptimisationBy Broadcasting), and adapts state-of-the-artcounterparts. Empirical comparisons show that COBBclearly outperforms all of them.

Autorentext

Yoseba K. Peña Landaburu (1977, Basque Country) got the PhD. degree under the supervision of Dietmar Dietrich and Nick Jennings. After working at the TU Vienna and the WU Vienna, he is research associate at the University of Deusto. His main research field is the application of AI and Game Theory models to the optimisation of energy consumption.

Klappentext

The deregulation of the energy industry has created marketplaces in which producers and consumers can operate in order to more effectively. This book develops a new model for electricity retail where customers may choose their supplier. It is based on simultaneous reverse combinatorial auctions, designed as a second-price sealed-bid multi-item auction with supply function bidding. This mechanism prevents strategic bidding and allows the auctioneer to maximise its pay-off. Furthermore, we develop optimal single-item and multi-item algorithms for winner determination in such auctions that are significantly less complex than those currently available in the literature. Nevertheless, the consumption of the energy of each singular auctioneer has to be adapted to the submitted bids to maximise the pay-off. Thus, this work models the constellation of energy consumers as a distributed constraint optimisation problem (dCOP). In order to overcome the specific domain demands, this book presents a novel optimal dCOP algorithm called COBB (Constraint Optimisation By Broadcasting), and adapts state-of-the-art counterparts. Empirical comparisons show that COBB clearly outperforms all of them.