Application of Incentive Based Scoring Rule Deciding Pricing for Smart Houses Shantanu Chakraborty 1,a) T akayuki Ito 1,b) Abstract: Defining appropriate pricing strategy for smart environment is important and complicated at the same time. In our work, we device an incentive based smart dynamic pricing scheme for consumers facilitating a hierarchical scor- ing mechanism. This mechanism is applied between consumer agents (CA) to electricity provider agent (EP) and EP to Generation Company (GENCO). Based on the Continuous Ranked Probability Score (CRPS), a hierarchical scoring system is formed among these entities, CA-EP-GENCO. As CA receives the dynamic day-ahead pricing signal from EP, it will schedule the household devices to lower price period and report the prediction in a form of a probability distribution function to EP. EP, in similar way reports the aggregated demand prediction to GENCO. Finally, GENCO computes the base discount after running a cost-optimization problem. GENCO will reward EP with a fraction of discount based on their prediction accuracy. EP will do the same to CA based on how truthful they were reporting their intentions on device scheduling. The method is tested on real data provided by Ontario Power Company and we show that this scheme is capable to reduce energy consumption and consumers’ payment. Keywords: Smart Grid, Pricing Scheme, Scoring Rule, Demand Response, Device Scheduling 1. Introduction With the growing needs of environmental sustainability and continuing changes in electric power deregulation, smart grid be- comes an inevitable choice for the society. While such grid in- frastructure in mind, houses started to adopt devices which can be controlled, maintained, monitored and even scheduled as ne- cessity calls. Smart house technology used to make all elec- tronic devices around a house act “smart” or more autonomous. Nearly all major appliances in the future will take advantage of this technology through home networks and the Internet. Such feature of smart grid is a way for ordinary electronics and ap- pliances to communicate with each other, consumers and even energy provider (EP). Recently, smart pricing has attracted much attention as one of the most important demand-side management (DSM) strategies to encourage users to consume electricity more wisely and efficiently [1]. On different note, in order to numerically measure up the actual realization of a probabilistic event which was forecasted ahead, scoring rule was defined [2], [3]. Moreover, it binds the asses- sor to make a careful prediction and hence truthfully elicit his/her private preferences. Which is why, scoring rule has been applied successfully while truthful incentive designing in diverse appli- cations such as voting rules [4] and [5]. Strictly proper scoring rules can be employed by a mechanism designer to ascertain that agents accurately declare their privately calculated distributions, reflecting their confidence in their own forecast. The applicabil- ity of scoring rule is being investigated in field of smart-grid. For 1 Nagoya Institute of Technology, Nagoya, Aichi, Japan a) shantanu.chakraborty@nitech.ac.jp b) ito.takayuki@nitech.ac.jp instance, [6] presented a methodology for predicting aggregated demand in smart-grid. Household devices such as Roomba vacuum cleaners, LG Thinq smart oven [7] are some commercially available smart de- vices that can be controlled and monitored via smart-meter. Us- ing such devices, consumers (actually a consumer agent, refer- eed as CA hereafter, will be responsible to take such decision in conjunction with smart-meter) can respond to day-ahead dy- namic pricing signal by effectively and intelligently managing and scheduling devices, thereby flattening out peak demand and achieving better resource utilization. This paper presents a hierarchical scoring rule based payment mechanism for CA provided by the EP and GENCO in response to the dynamic day-ahead time dependent pricing. The con- sumers will be rewarded a discount on the price to measure up how well they predict the shifting the devices/loads towards the lower demand (lower price as well) periods. These rewards are again a fraction of the discount which were provided by GENCO to the corresponding EP depending on EP’s prediction of required energy demand. The reward mechanism is based on a strictly proper scoring rule. The scoring rule is chosen to reflect to work with continuous variable (the normal distribution, as in the pro- posed method) and measure up how accurate the prediction could be. The Continuous Ranked Probability Score [8] possess such characteristics. EP will formulate an optimization problem to- tal energy demand for its consumers and reports to GENCO. GENCO then run an optimization algorithm that will minimize the cost of providing rewards to EPs while satisfying EPs energy demand. Therefore, the reward is actually dependent on both the consumers prediction and EP’s optimization problem. The rest of the paper is organized as follows. Section 2 intro- 1 ⓒ 2013 Information Processing Society of Japan IPSJ SIG Technical Report Vol.2013-ICS-171 No.14 2013/3/19