Wijaya, Tri KurniawanPapioannou, Thanasis G.Liu, XinAberer, Karl2013-09-032013-09-032013-09-03201310.1109/SustainIT.2013.6685188https://infoscience.epfl.ch/handle/20.500.14299/94441WOS:000330731100002Periods of peak consumer demand in today’s electricity sector are expensive to satisfy and can be the source of power failures. One possible solution is the use of demand-side management (DSM) applying dynamic pricing mechanisms. However, instead of reducing peak loads, these mechanisms can lead to peak-shifting due to the herding effect of consumers’ load-shifting behavior. To overcome this problem, we explore strategies of assigning (non-uniform) participation rates to consumers. We use a generic method to find a near-optimal distribution setting for participation rates. Our method allows DSM designers to tune the system toward consumer convenience. This means less frequent consumption schedule changes, in the price of system performance. In addition, consumers do not need to reveal their detailed consumption schedules (hence, their privacy is preserved). Using experiments, we show the impact of the herding effect and evaluate the effectiveness of the proposed solution. We thereby demonstrate price fairness for consumers. Finally, we apply our solution to a more realistic environment – one where consumers change their consumption behavior every day.demand-side managementherding effectmonte carlo tree searchuctsmart griddynamic pricingparticipation ratedemand responsetext::conference output::conference proceedings::conference paper