Yoshio Izui

Application of data mining to customer profile analysis in the power electric industry

In Japan, the Revised Electric Utility Industry Law went into effect on March 21, 2000. The partial deregulation into the retail electric power supply sector means that electric power for extra-high voltage customers can now be supplied from companies other than the 10 major electric power companies (EPCos) and electricity rates charged to customers can be determined freely according to negotiations for the liberalized sector between these companies and their customers. Due to the introduction of this partial deregulation, the 10 EPCos have found it necessary to form customer strategies against new entrants, power producers and suppliers (PPSs). In other industries (such as retail trade) in which competition with smaller companies is fierce, it is indispensable to assess preferred customers yielding more profits and form marketing strategies (i.e., business intelligence) to strengthen relationships with these preferred customers by providing unique services. Even in the electric power industry, it is expected to become necessary to form the same marketing strategies as these unrelated industries in consideration of future fierce competition. In order to draw up marketing strategies, it is first indispensable to understand customers, that is, the analysis of customer data. Since the nature of electric power differs from the products that the retail trade business targets, it is necessary to carry out a type of customer data analysis that differs from the retail trade business. The authors explain an example of marketing method to establish customer strategies, using data mining technique based on customer profile data.

Data mining for customer load profile analysis

In the progressing liberalization of the electric power industry worldwide, Japan too revised its Electric Utility Industry Law in May 1999 and began partial liberalization in March 2000. The points of the liberalization were partial retail liberalization targeting extra-high-voltage customers whose power supply is 20 kV or above and whose contracted demand is 2000 kW or above, and to make it possible for these extra-high-voltage customers accounting for about 30% of the total power demand to freely select their electric power suppliers. In March 2002, there were nine Power Producers and Suppliers (PPSs) that applied to the Ministry of Economy, Trade and Industry as new companies in the electric power industry. At present, there are about 50 examples showing that customers changed to new companies entering the electric power business from existing electric power utilities, but this number is very few if we compare the number with the number of customers of existing electric power utilities in terms of contracted electric power quantity. However, there are also examples showing a drop of about 20% in electric power charges by customers changing to new companies, and it should not be forgotten that more customers will change to other electric power sources in the future. In addition, there are many views that targets of liberalization will increase from the reconsideration of the partially liberalized system in 2003. If we suppose the range of liberalization will increase to include high-voltage customers, we can predict fierce competition between existing utilities and new entrants or other major utilities, since customers in liberalized sector should increase significantly compared with the current number. In a liberalized market, Customer Relationship Management (CRM) is very important. CRM involves assessing customers yielding profits and constructing relationships with these customers through the implementation of ideal measures directed at the customers. In other words, although there are market-driven approaches looking at what market segments the traditional marketing serves, CRM is a customer-based market-driven approach that continuously offers products aiming at improving customer satisfaction for each and every customer as well as the idea of one-to-one marketing in the marketing field. The electric power industry is an industry that had allowed local monopolies to date in the background of economies of scale, so it was unnecessary for electric power utilities to deeply understand customers. However, in order to effectively oppose new companies in liberalized sector, it is necessary for these utilities to recognize preferred customers based on customer profiles (e.g., load profile history).

A study of resource constraint project scheduling problem for energy saving

In terms of energy consumption, the environment surrounding the manufacturing industries has become considerably strict. As a framework containing many production scheduling problems in manufacturing industries, a great deal of researches on the resource-constrained project scheduling problem (RCPSP) has been performed. We have proposed a RCPSP formulation called RCPSP/πRC, which can deal with realistic energy constraints such as power restriction during peak hours, contract demand, and energy consumption during the setup operations. In this paper, we evaluate RCPSP/πRC through computational experiments. The proposed method is compared with two heuristic rules. Results of computational experiments show the effectiveness of the proposed method.