JP2007004646A - Power transaction support system and power transaction support program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To support the preparation of a transaction strategy obtained by taking into consideration a generator set of one's company, and the profit and risk of a transaction position in a power market. <P>SOLUTION: This power transaction support system is provided with: a power price distribution estimating part 301 for storing or receiving power market price data and weather forecast data 221, transaction data 271, power demand data 261, customer data and power cost calculation data and estimating a relation model of the power market price data and the weather forecast data and an error distribution; a fixed probability calculating part 501 for calculating a fixed probability of a power price; a trading profit and loss calculating part 401 for calculating a difference between profit brought about by certain bidding and expected profit when bidding is not performed; a bidding pattern calculating part 701 for calculating a bidding price at which the increment of the expected profit is maximum and a bidding amount, a weather amount distribution estimating part for calculating a weather forecast amount distribution at a certain future point of time; and a business risk calculating part 601 for controlling each estimating part or calculating part so as to repeat processing that seeks profit with an input pattern in which the increment of the profit is maximum and calculating a profit distribution. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to a power trading support system and a power trading support program for supporting trading in a power market.

  The power wholesale transaction market was established in April 2005, and power liberalization has been expanded in Japan. One purpose of electricity liberalization is to lower the price of electricity using the principle of competition, but in other countries, as seen in California in the United States in 2000, It may fluctuate greatly and is not always stable at a low price. Electricity market participants want to trade at a lower price when buying and at a higher price when selling, and the price depends on the relationship between supply and demand. In Japan's wholesale power market as of April 2005, there is a trading market in which electricity is delivered at a fixed time every 30 minutes the next day. It is determined by matching electronic bills and bids for electric energy. That is, the transaction is established by a market participant who puts a bid with a price lower than the matching price and a bid with a price higher than the matching price, and the established transaction price is a uniform matching price.

  In this way, the electricity price in the electricity market is determined based on the relationship between demand and supply. However, since electricity cannot be stored basically in an inexpensive manner, the power generated on the supply side is instantaneously on the demand side. If it is consumed and its supply and consumption do not match, an accident will occur. As a result, power prices fluctuate greatly. For example, the electricity price during the day is nearly twice the electricity price at night, but this is a buyer's market because the demand for electricity at night is generally low, and only market participants with low priced bids can buy electricity. On the other hand, since it is a relatively buyer market in the daytime, only participants who have placed a bid with a certain high price can sell power.

  The electricity wholesale exchange market participants are mainly electric power companies, generators, and electric power wholesalers, and fluctuations in the electricity price of the electricity wholesale exchanges are affected by the fluctuations in the electricity price for each of the companies. The technology to manage the risks that the company has on business profits has become important.

  Among the prior arts, a risk management technique that takes into account the power price includes a power generation facility operation system that takes into account the power price in Patent Document 1. This system is a system that calculates the risk value that impairs the optimality of the optimal operating conditions of the power generation equipment owned in-house from the viewpoint of profit, using the probability distribution of the power transaction price prediction value and the power demand prediction value as inputs.

  Document 2 is a conventional risk management system that takes into account the relationship between power demand and price. This management system is a system that determines the price of financial derivatives for risk hedging by reorganizing the portfolio to reduce the risk. The power demand, price data, power demand / price relationship and future demand for a certain period in the past. Means for estimating future fluctuations in power prices from the fluctuation prediction data.

  As a conventional technique related to a bid strategy, there is a power trading plan creation method and apparatus disclosed in Document 3. This device is designed so that the user can visually determine the best plan from multiple plans, based on the information on the power that has already been supplied, the power generation facilities of the company, and the estimated price of the market. As a device for calculating and displaying the relationship between revenue and transaction volume.

JP 2005-4435 A JP 2004-252967 A JP 2005-51866 A Takamitsu Sato regression analysis Asakura Shoten (1979)

  Participants in the power wholesale exchange want to buy more at a cheaper price or sell more at a higher price. The higher the price, the higher the probability of buying and the lower the probability of selling. In addition, if the amount of bids is large, the profit varies greatly depending on whether or not the transaction is established. As of April 2005, in Japan's power wholesale exchange, trading participants can bid on multiple bills, so the power you want to sell is cheap, and if you want to make a high profit, but not sure, a slightly higher price Various bid patterns can be considered when considering the combination of price and quantity for multiple tags, such as bidding. Among these patterns, it is necessary to determine the bidding pattern in consideration of risks and returns. In addition, if you own a generator, trading participants with a contract to sell power to customers whose demand has already fluctuated due to their own power generation capacity and unit price, relative, etc. must consider the fluctuating demand. I must.

  In the system of Patent Document 1, the influence on the market price or the company's profit is stochastically obtained based on the probability distribution of the power price. However, although the profits of trading on exchanges vary greatly depending on the bid strategy and whether or not a successful bid is made, the optimal bidding pattern on the exchange is not considered.

  The system of Patent Document 2 models the random fluctuation of the future power price (probability distribution) from the relational expression between the past power demand and the power price and the prediction of the future power demand, and further, one day, week, year The amount of risk of the portfolio is calculated taking into consideration the regularity of the market, but arbitrarily determined amounts and prices such as bidding on exchanges are not considered. In addition, fluctuations in profits that depend on whether or not to make a successful bid at an exchange are not taken into account.

  The device of Patent Document 3 basically supports the user to select a plurality of proposals visually by visually displaying the bid plan proposal, and the selection criterion is the user's subjectivity. It is also quite a skill to visually understand the fluctuation range of profits, the probability distribution of transaction volume and the probability distribution of market price at a time, and to determine the relationship between risk and expected profit and to formulate a bid plan. Seems to need. The object of the present invention is to determine a bidding pattern in which profits are maximized below a certain allowable risk, taking into account the difference in profits between successful and unsuccessful bids and the probability of successful bids depending on the bid price. In addition, it is intended to support transaction bidding and risk management by calculating profit fluctuation risk for a specific period in consideration of annual price fluctuations and bidding patterns.

  Electricity price distribution that uses power market price data, weather forecast data, transaction data, power demand data, customer data, and power cost calculation data as input or data, and estimates the relationship model and error distribution between power market price data and weather forecast data The estimation unit, the contract probability calculation unit that calculates the contract probability for the power price, the trading profit / loss calculation unit that calculates the difference between the profit that a certain bid brings and the expected profit when not bidding, and the bid price and bid that maximize the expected profit increase Each estimation unit or the bidding pattern calculation unit that calculates the amount, the weather amount distribution estimation unit that calculates the weather forecast amount distribution at a certain time in the future, and the process of calculating the profit and loss in the input pattern that maximizes the revenue A power transaction support system characterized by comprising a business risk calculation unit that controls the calculation unit and obtains the distribution of revenue .

  By the above means, the bidding pattern that maximizes the profit below a certain permissible risk is calculated in consideration of the difference between the profits when the bid is made at the exchange and the probability of the successful bid that depends on the bid price. In addition, it is intended to support transaction bidding and risk management by calculating the earnings fluctuation risk for a specific period in consideration of annual price fluctuations and bidding patterns.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings. FIG. 1 is a diagram showing an example of the configuration of a power trading support system according to the present invention. The power price distribution estimation unit 301 receives the power market price data 201 and the weather forecast data 221 as input, and estimates a relationship model between the power price and the weather forecast information data. The power demand distribution estimation unit estimates a relational model of weather forecast data, customer data, and power demand data from the weather forecast data 221, customer data 241, and power demand data 261. Further, the trading profit / loss calculation unit 401 calculates the expected profit for the power demand from the power demand data 261 or the power demand distribution estimated by the power demand distribution estimation unit 361, the transaction data 271 and the power cost calculation data 281. The contract probability calculation unit 501 calculates the contract probability of a price under the latest weather forecast data from the relationship model and error distribution estimated by the power price distribution estimation unit 301. The bid pattern calculation unit 701 estimates the bid pattern from the contract probability at each price estimated by the contract probability calculation unit 501 and the expected profit for the power demand calculated by the trading profit / loss calculation unit 401. The weather quantity distribution estimation unit 331 estimates the weather forecast distribution from the weather forecast data. The business risk calculation unit randomly generates from the weather amount distribution by controlling the meteorological amount distribution estimation unit 331, the power demand distribution estimation unit 361, the power sale profit / loss calculation unit 401, the contract probability calculation unit 501, and the bid pattern calculation unit 701. Based on the meteorological data, the tender pattern is calculated based on the distribution of power demand and the contract probability, the profit in the tender pattern is repeatedly calculated, and the frequency distribution of the business profit is calculated. In the input / output interface, the results of the contract probability calculation unit 501, the bid pattern calculation unit 701, and the business risk calculation unit 601 are displayed. FIG. 2 is a diagram showing an example of power price data. The electricity price data is a contract price in the market stored for each date, delivery time, and area. An area refers to any trading area such as a 50 Hz area and a 60 Hz area, and a Hokkaido area and a Tohoku area. FIG. 3 shows an example of weather forecast data. The meteorological data is meteorological forecast data corresponding to each date, time, place, and forecast date and time, and is data such as temperature, weather, and humidity. However, it is not always necessary to divide by the forecast date and time. Further, this data is weather forecast data effective for explaining the electricity price, and all of temperature, weather and humidity are not necessarily required. Here, the weather is composed of data corresponding to sunny, rainy, cloudy, and the like. Moreover, the maximum temperature, the minimum temperature, the average temperature, etc. may be used when the weather forecast is in units of one day or half a day.

  FIG. 4 is a diagram illustrating an example of customer data. Customer data is the data used when estimating the customer's demand and estimating the demand. The data corresponding to each date and contract, whether the day is a holiday or a weekday, opening / opening, closing / The closing time is recorded. Here, since the customer may have a plurality of contracts, it is described for each contract, not for the customer.

  FIG. 5 is a diagram showing an example of power demand data. The power demand data is data in which the power demand corresponding to each date, contract, and time is described.

  FIG. 6 is a diagram showing an example of transaction data. The transaction data is data in which a contract power amount and a contract unit price corresponding to each contract are described. Since the contract unit price is different for each time zone and season, the unit price is described according to the charge system.

  FIG. 7 is a diagram showing an example of data relating to generator characteristics in the power calculation data. The data on the generator characteristics includes the maximum output corresponding to each generator and date, and the coefficient of the approximate expression of power generation cost. When stopped at the time of inspection or the like, the maximum output 0 is described. This data may have separate data on the date when the generator stops, such as the coefficient of the approximate expression of the generator and periodic inspection. In addition, since the performance of the generator may deteriorate due to aging deterioration, etc., even if the coefficients of the approximate expression are stored separately from the schedule of the periodic inspection, predictions corresponding to each year or each year The coefficient data of the approximate expression to be used may be recorded.

  FIG. 8 is a diagram illustrating an example of data relating to an imbalance fee among the power calculation data. The unit price corresponding to each type is described in the data on the imbalance fee. Here, imbalance is a charge when excess or deficiency occurs in the power with respect to the reserved amount of the transmission line reserved in advance or the transaction amount, and when the supply side transmits extra power to the demand amount. The surplus power take-up price and backup fee to be purchased when it is known that the power is insufficient in advance are also included.

  Next, a relation model of the power price distribution estimation unit 301 and a calculation method of the error distribution will be described. FIG. 9 is a diagram for explaining an example of a relationship model between the power price and weather forecast data. This figure shows temperature data corresponding to each power price data plotted on a graph with temperature on the horizontal axis and power price on the vertical axis. The first order is such that the error is minimized by the least squares method. The equation is a straight line described in the figure. Here, the variable for explaining the power price is the temperature data, but it is not necessarily limited to the temperature, and may be weather, humidity, or a combination thereof. The day of the week may also be an explanatory variable. If a statistical method such as multiple regression analysis is used, an equation for estimating the power price composed of a plurality of explanatory variables, that is, a relational model can be obtained. The power price distribution estimation unit 301 employs a model with the smallest error among models obtained by combining combinations of explanatory variables. In addition, the electricity price model is the same throughout the year due to seasonal effects such as the consumption of electricity during the daytime due to cooling during the summer, while the consumption from the evening to the board due to heating during the winter. It is more likely that the error will be smaller than using the estimation formula depending on the season and so on. Therefore, instead of using long-term data such as one year, when obtaining each model, it is more error to obtain each model using the weeks just before the current day and the weeks in the same season for several years. Is likely to be small. Here, the model with the smallest error is obtained for each product listed for each delivery time zone listed on the exchange. Statistical methods such as multiple regression analysis are described in Non-Patent Document 1, for example.

  FIG. 10 is a diagram illustrating an example of an error distribution in a model for estimating a power price. The error distribution in this case is a distribution of errors between the model estimated by the power price distribution estimation unit 301 and the actual power price when a certain expected temperature is given. The distribution of the difference between the combination and the electricity price estimated by the relational model from the weather forecast data. This distribution is not limited to one, but may be classified according to weather forecast data and may be an error in the category. For example, the error distribution is created by dividing the temperature into less than 15 ° C. and 15 ° C. or more.

  The power price distribution estimation unit 301 uses a statistical technique such as multiple regression analysis as described above to estimate a relational model of two data that minimizes the error between the power price data and the weather forecast data. An error distribution is calculated, and the value is stored in a storage medium as necessary.

  Here, the error distribution shows the deviation of the actual value with respect to the expected value by the probability, and the area of the distribution on the right side with respect to the bid at a certain price is the probability that it can make a successful bid. The area is the probability of winning a bid. Therefore, FIG. 10 is an example showing the relationship between price and execution probability. As an example showing the relationship between price and execution probability, cumulative probability density is also effective. When considering the bidding pattern, the expected value and distribution obtained by inputting the latest weather forecast into the relational model are displayed.

  FIG. 11 shows an example in which the relationship between the contract probability and the time is displayed. Each of these curves finds a price with a probability corresponding to each time from the expected value obtained by inputting the latest weather forecast into the relational model estimated by the power price distribution estimation unit 301 and the cumulative density distribution, FIG. 11 may be created.

  FIG. 12 is a distribution diagram of demand and its appearance frequency. This distribution map replaces electricity prices with electricity demand, and uses statistical techniques such as multiple regression analysis in the same way as the relationship data between weather forecast data and electricity price data. A relational model can be obtained, and it can be obtained as a distribution of the difference between the combination of power demand data and weather forecast data and the power price estimated by the relational model from the weather forecast data. Electricity demand is not very realistic when it comes to predicting the demand of each customer when there are many customers, because the calculation becomes enormous, but when there are few customers, it is useful to predict demand for each customer. Become. In that case, in addition to the weather forecast data, the customer data is also added to the explanatory variables to obtain the distribution, and the total demand distribution is created by adding the distributions. The power demand distribution estimation unit 361 performs these series of processes.

  FIG. 13 is a diagram showing an example of the relationship between revenue and demand. In the case of electric power wholesalers, they purchase cheap electricity and sell profits to consumers, so profits increase as demand increases up to a certain demand. However, since the amount of cheap power that can be purchased is limited, if the demand exceeds that amount, high power must be purchased. For this reason, when a cheap purchase unit price has to purchase electric power that exceeds the sales unit price, the profit decreases as the demand increases. Even in the case of power producers, there are many ways to adjust the power generators to meet demand by moving them in the order of efficiency, which is economically efficient. In addition, if the sales volume exceeds the power generation capacity, it will be necessary to procure a shortage of electricity, but if the power to be procured is higher than the sales unit price, it will rise upward as in the case of power wholesalers. It becomes the curve of. Expected profit can be obtained by making this curve correspond to the appearance frequency of the demand in FIG. 12 and adding the product of the appearance probability and the profit for each demand. In addition, for the price and quantity to be traded on the exchange, revenue can be calculated from the product of price and quantity, and the cost can be calculated from the quantity and power cost calculation data. You can get the expected profit corresponding to. These series of operations are performed by the power trading profit / loss calculation unit 401. That is, for each demand in FIG. 12, revenue is obtained from transaction data and expense is calculated from power cost calculation data to obtain income, and the product of the difference revenue and appearance probability is added to obtain expected income. This series of processing is performed for a combination of a predetermined bid amount and price. Revenue is the difference between revenue and cost, but if the components that fluctuate due to fluctuations in demand are assumed to be variable costs and variable revenues, it is only necessary to pay attention to these quantities. It is sufficient that at least data capable of calculating the component is included. Therefore, it is sufficient that the power cost data includes at least generator characteristics and imbalance fees, and the transaction data includes at least metered fee data.

  From the viewpoint of risk management, the maximum loss or minimum profit that can occur at X% or less is calculated. However, in the combination of appearance probability and profit, the probability of appearance is added to X% from the largest profit order. If this is the case, the power trading profit / loss calculation unit 401 also performs a series of processing for obtaining the maximum loss or the minimum profit, which is obtained by the last combination of profits.

  FIG. 14 shows an example of the relationship between the trading volume, the bid price, and the appearance frequency. Once the bid amount and the bid price are determined, the expected profit can be calculated from the trading profit / loss calculation unit 401. Therefore, if the expected profit is calculated for each bid price and the trading volume, The right graph in FIG. 14 can be drawn. On the other hand, the diagram on the left represents the contract probability distribution calculated by the contract probability calculation unit 501 with the vertical axis representing the bid price and the horizontal axis representing the appearance frequency. By using the graphs on the left and right, it is possible to obtain the profit when making a successful bid and the contract probability for a certain bid price and quantity. Similarly, since the price at the time of a successful bid and the profit at that time can be obtained, the expected profit at this bid price and quantity can be calculated.

  When bidding multiple tags, for example, if the A tag sells 1 MW with a price of 10 yen / kWh, and the B tag sells 11 MW / kWh with an amount of 1 megawatt, the A tag Since the price is cheaper, the probability that both tags are not executed and the profit at that time, the probability that only the A tag is executed and the expected profit, and the probability that both tags are executed and the expected profit are calculated. In addition, if only the A bill is executed, the transaction amount is increased by 1 mega to the expected demand when the transaction is not executed. If both bills are executed, the transaction amount is the expected demand when the transaction is not executed. On the other hand, it is calculated by the amount increased by 2 mega. In the case of three or more, similarly, after obtaining the transaction amount by the execution probability of each tag and adding the product of the execution probability and the profit at the bid price, the expected return by a plurality of tags can be obtained. If the same consumer cannot place a bid and a bid with the same price, there are restrictions on the generator's capacity, the consumer's contracted energy, the minimum unit of transaction, etc. Limited number. The bid pattern calculation unit performs a series of processes for searching for a combination having the highest expected value for the finite number. In other words, for each bid volume and price pattern, the expected value is calculated based on the probability and profit when each tag is executed and not based on the execution probability distribution and trading revenue, and the bid pattern with the highest expected value is calculated. Explore. This search can be performed by full search, but can be solved efficiently by using a method such as linear programming.

  A method for obtaining the maximum loss or minimum profit at X% or more for a certain bidding pattern will be described. In the example of the A tag and the B tag, there are three cases in which both are not executed, only the A tag is executed, and both are executed, and the appearance probability of the three cases can be calculated from the execution probability distribution. In addition, since the transaction volume is known in all three cases, the occurrence probability of demand and the profit can be calculated for each case. Therefore, by multiplying the occurrence probability of each demand by the execution probability and obtaining the conditional probability, and adding up to X% in order of the corresponding revenue, it corresponds to the appearance probability added last. Revenue is the desired value. This series of processing is performed by the trading revenue calculation unit. In each bid pattern, when finding a bid pattern that maximizes profit under the condition that X% or more and the maximum loss is Y or more, the maximum loss is calculated according to the above procedure, and the loss is Y or more. If the condition is not satisfied, it is possible to obtain a bid pattern that maximizes the profit in consideration of the risk by removing it from the options.

  FIG. 15 is a diagram illustrating an example of a weather forecast distribution. This distribution is the appearance probability distribution obtained by counting the values of the weather forecast data, and is calculated by the meteorological amount distribution estimation unit 331. Since the weather distribution is seasonal, it is calculated separately for each season. When obtaining the simultaneous distribution of temperature, weather and humidity, the variables are divided and the combinations that meet the conditions are counted for each combination of the categories. For example, there are three distributions: a temperature distribution when it is clear, a temperature distribution when it is cloudy, and a temperature distribution when it is raining. In addition, the probability of clearing, the probability of raining, and the probability of cloudy are also obtained.

  FIG. 16 is a diagram illustrating an example of the business revenue appearance probability distribution. The appearance probability distribution of business revenue is determined from the contract probability distribution estimation unit and the contracted probability distribution estimation unit to the power demand distribution estimation unit according to the weather conditions randomly generated from the weather forecast distribution. Search for a bidding pattern. At this time, the profit distribution when each tag in the selected pattern is executed is obtained, and each profit and its conditional appearance probability are held. By repeating this series of steps and dividing the number of repeated conditional occurrence probabilities, adding the conditional probabilities that result in the same revenue, and arranging them in ascending order of revenue, a revenue distribution in a certain time zone is obtained. It is done. In this way, the profit distribution by time, season, and day of the week is obtained. From these distributions, an appropriate distribution is selected according to the season, time zone, and day of the week in the evaluation target period, and revenue is randomly generated from the distribution, and the total profit is obtained as the profit in the evaluation target period. By repeating this series of processes many times, the distribution shown in FIG. 16 can be obtained.

The example of a structure of an electric power transaction support system. Example of electricity price data. An example of weather forecast data. Example customer data. An example of power demand data. An example of transaction data. An example of generator characteristic data in the power cost calculation data. An example of penalty charges in the power cost calculation data. An example of the relationship between electricity price and temperature. An example of the appearance frequency of electricity prices. An example of the appearance frequency of electricity prices. An example of the frequency of demand. An example of the relationship between revenue and demand. Example of relationship between trading volume and revenue, market price and probability distribution. An example of the frequency of air temperature. An example of revenue frequency.

Explanation of symbols

DESCRIPTION OF SYMBOLS 101 ... Electric power transaction support system 201 ... Electric power market price data, 221 ... Weather forecast data, 241 ... Customer data, 261 ... Electric power demand data, 271 ... Transaction data, 281 ... Electric power cost calculation data, 301 ... Electric power price distribution estimation part 331 ... Meteorological quantity distribution estimation unit, 361 ... Electric power demand distribution estimation unit, 401 ... Trading profit / loss calculation unit, 501 ... Contract probability calculation unit, 601 ... Business risk calculation unit, 701 ... Bid pattern calculation unit, 901 ... Input / output interface face.

Claims (10)

  A system that supports the examination of transactions in the electric power market, and estimates the relation model of the two data and the error distribution of the model from data including at least the market price of electric power and weather forecast data, and the relation model and the error distribution. A power trading support system comprising a power price distribution estimation unit for estimating a power price distribution after a short time from the latest weather forecast data and an input / output interface.   The power trading support system according to claim 1, wherein a relationship model between a market price and weather forecast data for each power product for which a delivery time zone is determined and an error distribution of the model are estimated, and the relationship model and the error are estimated. The power price distribution estimation unit that estimates the power price distribution of each product from the latest weather forecast data using the distribution, and the lowest price that can be awarded with a certain probability or more from the estimated power price distribution of each product. A power trading support system comprising: a contract probability calculation unit for calculating; and an input / output interface for displaying a relationship between price and product or price and probability on a graph.   The power trading support system according to claim 1 or 2, wherein a combination of meteorological data and electric power market price for the last few weeks and a few weeks ago is used to estimate a relationship model between market price and weather forecast data. An electric power transaction support system comprising the electric power price distribution estimation unit to be used.   4. The power trading support system according to claim 1, wherein a profit and a new trading are obtained by adding a new trading volume and a contract price at a specific time from the power demand data and power cost calculation data of a contract already executed. A power trading support system comprising a trading profit / loss calculation unit for calculating a difference in profits when there is not.   5. The power trading support system according to claim 4, wherein a relational model of these data and an error of the relational model are estimated from at least weather measurement data or weather forecast data, power demand data, and customer data, and the latest weather. Power demand distribution model that predicts power demand distribution in a short time from forecast data, customer data, and transaction data, estimated power demand distribution, power cost calculation data, and transaction data; A power trading support system comprising a trading profit / loss calculation unit for obtaining a difference between the two.   The power trading support system according to claim 4 or 5, wherein the relationship between the trading volume calculated by the trading profit / loss calculation unit, the difference between the contract price and the expected profit, and the power price distribution estimated by the power distribution estimation unit. A power trading support system comprising a bid pattern calculation unit for calculating a combination of a bid price and a bid amount that maximizes an increase in expected profit.   7. The electric power transaction support system according to claim 4, wherein an increase in expected profit is a minimum or a decrease in power demand distribution, transaction data, and electric power cost calculation data under a probability specified in each trading volume. A buy / sell profit / loss calculation unit for obtaining a risk value, and a bid pattern calculation unit for calculating a bid pattern combination in which the risk value is equal to or less than the specified risk tolerance and the expected profit increase is maximum. A featured power trading support system.   8. The power trading support system according to claim 4, wherein each of the relationship between the trading volume calculated by the trading profit / loss calculation unit, the contract price, and the profit difference, and the power price distribution estimated by the power distribution estimation unit, A bidding pattern calculation unit that calculates the increase in expected revenue for a combination of trading volume and contract price, and an input / output interface that displays a graph of the combination of trading volume and contract price that is the increment value of a specific expected profit. A featured power trading support system.   9. The power trading support system according to claim 1, wherein a meteorological amount distribution estimating unit that calculates a meteorological forecast amount distribution at a certain time in the future from meteorological forecast data and a meteorological amount distribution calculated by the meteorological amount distribution estimating unit. Therefore, the electricity price distribution estimator for estimating the electricity price distribution under the randomly generated meteorological amount from the past electricity price data and the meteorological data of the same time period at a certain time in the future for the randomly generated weather amount; Input pattern that calculates the bid pattern that maximizes the increase in revenue from the relationship between the contract price, trading volume, and increase in revenue obtained by the trading profit and loss calculation section using the calculated power demand distribution as input. Repeat the series of processes for calculating profits and losses in the input pattern that maximizes the calculated increment in the electricity price generated from the electricity price distribution. Power trading support system, characterized in that control each estimator or calculating unit, and a business risk calculation unit for obtaining a distribution of revenue as.   The computer estimates a relational model of the two data from the data including at least the market price of electricity and weather forecast data and the error distribution of the model, and uses these relational models and error distribution to shorten the latest weather forecast data. Electricity price distribution estimation means for estimating the electricity price distribution after time, and an electricity transaction support program that functions as an input / output interface.

Images