Contents I. Introduction
In recent power system, it is becoming an important issue to use efficiently demand side resources because of restrictions for utilizing conventional generation resources. Furthermore, recent advancement of smart grid technology including auto-metering and communication makes easily it possible to develop demand response with a form of program to use demand side resources practically. Demand Response (DR) can be defined as the changes in electric usage by end-use customers from their normal consumption patterns [1]. According to the definition, main agents of DR are not operators but customers, and they will practically exert favorable influences such as improving system reliability and lowering electric price by participating in demand response program (DRP) [1]–[2]. DR extends the customer's participation on the power market, and it assures to take a key role as a paradigm shift in power system because customers would control their load voluntarily as the peak load increases, instead of catching up the increasing load passively with generation resources. A recent research [2] shows how the customers' behaviors are modeled by using an elasticity matrix composed of the price-elasticity of demand. References [3]–[7] show that customer's response has positive influences on the power market performance, nodal price and nodal reliability indices, available transfer capability and spinning reserve, based on the method of reference [2]. The references [3]–[7] are seen as merely modeling customer's response according to the changes of electricity price during a specific hour periods of the day based on the constant elasticity matrix. In practical terms, however, the assumption of the constant elasticity matrix with a certain specific period is not reasonable and moreover, the frequency of DR has not been modeled at all and false and restrictive results may be incurred by the method in using an elasticity matrix. Also the price-elasticity of demand requires a high credibility, but lack of credibility because of the scant information about the response characteristics. For this reason, system operators who operate DRP request information necessary to operate DRP from customers. An example, NYISO requires to provide interruptible load rating, reduction lasting time and response time [11]–[12]. In this paper, aggregated information and their relationship are modeled in a closed form of expression. Also this paper introduces the concept of virtual generation resources converted from demand resources and their marginal cost is calculated according to customer information. Finally, some constraints of DR are manipulated and expressed using the information modeled in this paper with various status flags, and the optimal scheduling combined with generation and DR is proposed by minimizing the system operation cost including generation and DR cost, with the generation and DR constraints developed in this paper.
