JP6697358B2 - Information processing apparatus, information processing system, information processing method, and program - Google Patents

Description

  Embodiments of the present invention relate to an imbalance price prediction device, method, program, and power trading system.

  Under the plan value simultaneous equalization system in the power system reform, the new electric power company submits the demand and supply plan value of the next day to the grid operator, and if there is a discrepancy between the supply and demand plan value and the plan value, the imbalance It is necessary to settle according to the amount. It is possible to reduce imbalance settlement payments by utilizing peak cuts due to demand response and power procurement from the market one hour ago. In order to manage this more rationally, it is necessary to predict in advance the imbalance settlement unit price (imbalance price) per imbalance amount and its fluctuation risk.

  In order to accurately predict the imbalance price, it is necessary to predict not only the expected value of the imbalance price but also the reliability. However, although a technique for predicting only the expected value of the imbalance price has been studied in recent years, it is difficult to predict its reliability.

JP, 2004-145396, A JP, 2007-4646, A Japanese Patent No. 4694828

  An imbalance price prediction device that estimates not only the expected value of the imbalance settlement unit price but also its reliability.

An imbalance price prediction device according to one embodiment,
Based on the past weather forecast value and the past power demand actual value, a model expressing the difference between the power demand forecast value and the actual value by the parameter and the difference between the temperature forecast values at two different timings. Estimating a demand change parameter probability distribution that is a probability distribution of the parameters, a demand change parameter probability distribution estimation unit,
A bid curve parameter probability distribution estimation unit that estimates a bid curve parameter probability distribution, which is a probability distribution of parameters representing the shape of a bid curve, based on contract data relating to past transaction volume in a specific market,
Based on the weather forecast value at the first timing and the second timing after the first timing, and the demand change parameter probability distribution, regarding the power demand at the third timing (the current day) after the second timing, the electric power A demand change probability distribution prediction unit that predicts a demand change probability distribution, which is a probability distribution of the difference between the predicted value of demand and the actual value,
It is an adjustment term of the predicted value of the imbalance price at the third timing and the past contract price based on the past contract market contract data, the bid curve parameter probability distribution, and the demand change probability distribution. A price multiplier probability distribution prediction unit that predicts a price multiplier probability distribution that is a probability distribution of price multipliers,
An imbalance price probability distribution calculation unit that calculates a probability distribution of imbalance prices based on the past specific market contract data and the price multiplier probability distribution.

The sequence diagram which shows the flow of the imbalance price forecasting method. The schematic diagram showing an example of an imbalance price forecasting device. The figure which shows an example of the supply-and-demand curve of an imbalance price. The block diagram which shows an example of the imbalance price prediction apparatus which concerns on 1st Embodiment. The flowchart which shows the process of the imbalance price prediction apparatus which concerns on 1st Embodiment. The figure which shows an example of the output result of the imbalance price reliability which concerns on 1st Embodiment. The figure which shows an example of the output result of the imbalance price reliability which concerns on 1st Embodiment. The figure which shows an example of the output result of the imbalance price reliability which concerns on 1st Embodiment. The block diagram which shows an example of the imbalance price prediction apparatus which concerns on 2nd Embodiment. The flowchart which shows the process of the imbalance price estimation apparatus which concerns on 2nd Embodiment. The figure which shows an example of the output result of the imbalance price reliability which concerns on 2nd Embodiment. The block diagram which shows an example of the electric power trading system which concerns on 4th Embodiment. The figure which shows an example of the hardware constitutions of an imbalance price prediction apparatus.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings.

(Common configuration)
The imbalance price prediction device according to the present embodiment tries to estimate the reliability or confidence interval of the predicted imbalance price by predicting the probability distribution of the imbalance price as well as the imbalance price. Is.

  First, an outline of the imbalance price prediction method will be described with reference to FIG. FIG. 1 is a sequence diagram showing the flow of the imbalance price prediction method.

  In the plan value simultaneous equal amount system, the electric power company submits the planned value of the power supply and demand to the grid operator by the day before the execution of the plan. The time of executing the plan is a predetermined period (for example, 30 minutes) in which the supply and demand plan is executed. In the following, the day including the time of executing the plan will be referred to as the current day. The planning time in FIG. 1 is the timing when the electric power company submits the plan to the grid operator. As shown in FIG. 1, the planning time is included in the previous day. The electric power company who submitted the plan is requested to receive the electric power so that the actual value of the power supply and demand matches the planned value.

  However, in reality, it is conceivable that there will be a gap between the planned value and the actual value. Therefore, when there is a deviation between the planned value and the actual value, the amount of difference power is multiplied by a predetermined unit price to settle the amount. This unit price is called an imbalance price or an imbalance settlement unit price.

  The imbalanced price is announced one month after the actual supply and demand when all supply and demand actual values are fixed, and settled in the following month. That is, the electric power company cannot know the imbalance price at the time of executing the plan. The electric power company acts so that the planned value and the actual value match, but there are various options for realizing this. Options include, for example, additional power procurement in the market one hour ago and peak cut due to demand response.

  The forecast of the imbalance price is preferably performed at least one hour before the execution of the plan. Further, considering the time margin for executing the options according to the predicted value of the imbalance price, the prediction of the imbalance price is preferably performed about 3 hours before the planned execution time.

  Furthermore, in the prediction method according to the present embodiment, the latest change in the weather forecast for the current day (second timing) with respect to the weather forecast for the previous day (first timing) is used. Generally, in the weather forecast, the prediction error increases as the time from the current time to the forecast time increases. It is empirically known that the prediction error is “the time from the current time to the prediction time increases abruptly for 0 to 3 hours, and thereafter increases gradually”. That is, when the time from the current time to the forecast time exceeds 3 hours, the significant difference (change) between the weather forecast at the current time and the weather forecast of the previous day becomes small. Therefore, in predicting the imbalance price, it is preferable to use, as the latest weather forecast, a weather forecast within 3 hours before the plan is executed.

  As described above, in the imbalance price prediction method according to the present embodiment, as shown in FIG. 1, the imbalance price prediction method is preferably executed from 1 hour to 3 hours before the planned execution (third timing). .. The imbalance price is the weighted average value w of the market (spot market) price Py one day ago and the market price Ph one hour ago, and the imbalance price = the weighted average value of the spot market price and the market price one hour ago × α + β Is required as. That is, the imbalance price = (w · Py + (1-w) Ph) α + β. The market price is, for example, a contract price acquired every 30 minutes.

  Here, α is an adjustment term (price multiplier) according to the supply and demand situation of the entire system. Also, β is an adjustment term (area correction value) that reflects the level difference of supply and demand adjustment costs in each region, and is a value obtained by subtracting the national average supply and demand adjustment cost from the annual average supply and demand adjustment cost in that region. is there.

  As shown in FIG. 1, the imbalance price forecast includes the weather forecast of the previous day (first timing), the latest weather forecast of the day (second timing), the spot market price Pd, and the market price Ph one hour ago. It is input to the imbalance price prediction device, and, for example, the prediction process is performed 1 hour to 3 hours before the execution of the plan (third timing). That is, based on the weather forecast of the previous day (first timing) and the latest weather forecast of 1 to 3 hours before the plan execution (third timing) (second timing), the plan execution (third timing) ) Imbalance price forecast.

  Next, a general configuration of the imbalance price prediction device according to the present embodiment will be described. FIG. 2 is a block diagram showing an example of a schematic function of the imbalance price prediction device according to the present embodiment. As shown in FIG. 2, the imbalance price prediction device includes a demand change parameter probability distribution estimation unit 10, a bid curve parameter probability distribution estimation unit 11, an imbalance price reliability prediction unit 20, and an imbalance price reliability output. And a unit 30.

  The demand change parameter probability distribution prediction unit 10 predicts the probability distribution of the demand change parameter based on the past weather forecast value and the past actual power demand value. The demand change parameter is a model parameter for calculating the difference D (imbalance amount) between the demand forecast value and the actual value regarding the demand amount in a certain time zone on a certain day.

  More specifically, the model is, for example, the difference d between the predicted temperature value of the same time zone calculated as of the previous day and the predicted temperature value of the same time zone of the same day calculated for the demand amount of a certain day of the day. Is a model for calculating the imbalance amount D. This model is based on the fact that power demand is greatly affected by weather (temperature, etc.). The model can be expressed by, for example, the formula D = ad + b. The coefficients a and b correspond to demand change parameters. The demand change parameter probability distribution prediction unit 10 obtains the probability distribution of the coefficient a and the probability distribution of the coefficient b by learning. The probability distribution of “a” and the probability distribution of “b” are obtained based on the past, for example, weather forecast values for several years and power demand actual values for several years.

As a learning method, linear regression may be used in the case of the model represented by the formula of D = ad + b. Furthermore, in this case, the probability distribution between a and b can be obtained as a normal distribution. That is, they are represented as _{a to} N (μ _a , σ _a ² ) and _b to N (μ _b , σ _b ² ), and the average μ _a and μ _b and the standard deviations σ _a and σ _b are obtained, Find the probability distribution of a and b. Further, these probability distributions may be expressed as a mathematical expression or may be expressed as a graph like a histogram. Here, N (μ, σ ² ) represents a normal distribution of mean μ and variance σ ² .

  As the model, in addition to the above, the predicted value difference h of the room temperature, which is a parameter other than the temperature, may also be considered, and in this case, a linear model of D = ad + bh + c may be used. Furthermore, the model is not limited to the linear model, and various other models such as a logarithmic model may be used. Further, the probability distribution is not limited to the normal distribution, and for example, a gamma distribution as a continuous probability distribution or a Poisson distribution, a Bernoulli distribution, etc. as a discrete probability distribution may be used. More simply, a triangular distribution or the like may be used. Further, in the above, "past" and "current day" are notations for convenience, and if there are two prediction values with a time difference, even if "two days ago" and "one day ago" are combined, "one week ago" , "One day before" or any other combination.

  The bid curve parameter probability distribution estimation unit 11 estimates the probability distribution of the parameters representing the shape of the bid curve in the spot market or the market one hour ago based on the past contract data. The bid curve is a demand curve showing the relationship between power and price and a supply curve showing the relationship between power and price. The demand curve and supply curve are collectively called the supply and demand curve. The past contract data is available information about market contracts, for example, the past total bid amount, contract amount, contract price, and the like.

As a concrete example of the probability distribution of the parameters, if the total amount of buy bids and the total amount of sell bids, which are publicly available past contract data, are Qd, and the total amount of sell bids is Qs, the shape of the bid curve is defined by the error function erf (x ) Approximates. More specifically, the demand curve is

Supply curve,

As, the probability distribution of Pd, Wd, Ps, and Ws is estimated. Pd, Wd, Ps, and Ws correspond to bid curve parameters. Here, Pd and Ps are the average values of the demand curve and the supply curve, respectively, and Wd and Ws are the values (or standard deviations) that are the variance values of the demand curve and the supply curve, respectively. When (2) is followed, the probability distribution becomes a normal distribution.

Here, the error function erf is represented by the following formula.

  FIG. 3 is a diagram showing a supply and demand curve between electric power and price. A typical demand curve d0 and curves d1 and d2 corresponding to certain confidence intervals for it are shown. A representative supply curve s0 and corresponding curves s1 and s2 corresponding to certain confidence intervals are shown. These confidence intervals correspond to Wd and Ws described above. Estimating the probability distributions of Pd, Wd, Ps, and Ws can be said to be estimating the probability distribution of the demand curve and the probability distribution of the supply curve. Since the price of electricity is determined at the intersection of the demand curve and the supply curve, it is possible to roughly estimate the amount of electricity and the price by using the probability distribution of the demand curve and the probability distribution of the supply curve. Becomes

  Returning to FIG. 2, the imbalance price reliability prediction unit 20 includes a demand change probability distribution prediction unit 21, a price multiplier probability distribution prediction unit 22, an imbalance price probability distribution calculation unit 23, and an imbalance price reliability calculation unit. 24, and predicts the imbalance price and its confidence interval.

The demand change probability distribution prediction unit 21 calculates the difference between the demand predicted value and the actual value from the probability distribution of the demand change parameter, the temperature predicted value t in the weather forecast of the previous day, and the temperature predicted value t ′ in the weather forecast of the day. The probability distribution of D (demand change probability distribution) is calculated. Specifically, first, _a variable a ′ according to the above-described probability distribution N (μ _a , σ _a ² ) and a variable b ′ according to the probability distribution N (μ _b , σ _b ² ) are generated. The variable a ′ is a value sampled from the probability distribution N (μ _a , σ _a ² ) and the variable b ′ is a value sampled from the probability distribution N (μ _b , σ _b ² ). Then, using the difference d (= t−t ′) between the predicted temperature value t in the weather prediction on the previous day and the predicted temperature value t ′ in the weather prediction on the day, the difference D between the demand prediction value and the actual value is calculated as Predict by D = a'd + b '. By performing this process a plurality of times, for example, the probability distribution of the difference D between the demand forecast value and the actual value output in the above example (demand change probability distribution) is calculated. For example, the demand change probability distribution is predicted as _{D to} N (μ _D , σ _D ² ). The calculation example of the demand change probability distribution shown here is an example, and another method may be used as long as it is based on the demand change parameter probability distribution, the predicted temperature value t, and the predicted temperature value t ′. The demand change probability distribution may be a probability distribution other than the normal distribution. For example, a histogram may be used.

  The price multiplier probability distribution predicting unit 22 sets the total of bid bids of the past (for example, the previous day) spot market contract data to Qd and the total amount of sell bids to Qs, and the probability distribution (Pd, Wd, Ps, Pd) of bid curve parameters. Probability distribution) and the demand change probability distribution, the probability distribution of the price multiplier α is predicted. Here, the price multiplier α is a multiplier for adjusting the predicted value of the imbalance price of the day and the past spot contract price.

  The imbalance price probability distribution calculation unit 23 is based on past (for example, the previous day) spot market contract data (price P: weight w is 1), area correction value β, and probability distribution of price multiplier α. Then, the probability distribution of the imbalance price is calculated. The predicted value Pi of the imbalance price is represented by the equation Pi = αP + β. The imbalance price probability distribution calculation unit 23 calculates the probability distribution of the predicted value Pi of the imbalance price.

  The imbalance price reliability calculation unit 24 generates and outputs an index for estimating the imbalance price predicted value and its reliability based on the probability distribution of the imbalance price predicted value Pi. The predicted value is, for example, an average value. The index may be a confidence interval, variance, or standard deviation, or may be another statistic in a histogram. The probability distribution itself may be output.

  The imbalance price reliability output unit 30 outputs the imbalance price predicted value calculated by the imbalance price reliability calculation unit 24 and the reliability index. The output method may be displayed on the screen of the display device, or data may be transmitted to a device held by the user. For example, when a confidence interval is output as an index, the confidence interval may be displayed as additional information on the time series graph of the predicted value of the imbalance price. When the variance or standard deviation is calculated as an index, the value of the graph (predicted value of imbalance price) and the variance or standard deviation value may be output. When a histogram is calculated as an index, the histogram may be shaded based on the predicted value of the imbalance price.

  The details of the embodiment of the operation of each component will be described below.

(First embodiment)
In the imbalance price prediction device 1 according to the present embodiment, in the above configuration, the probability that is input when the probability distribution is calculated in the demand change probability distribution prediction unit 21 and the price multiplier probability distribution prediction unit 22. The imbalance price is predicted by calculating the imbalance price at the time of execution of the plan (third timing) by giving the variable width based on the distribution, and repeating the process of calculating the imbalance price. It aims to increase the reliability of. FIG. 4 is a block diagram showing an example of the configuration of the imbalance price prediction device 1 according to the present embodiment. As shown in FIG. 4, the demand change probability distribution prediction unit 21, the price multiplier probability distribution prediction unit 22, and the imbalance price probability distribution calculation unit 23 of the imbalance price prediction device 1 according to the present embodiment are as follows. It has various configurations.

  The demand change probability distribution prediction unit 21 includes a demand change perturbation unit 210 and a demand change expected value prediction unit 211.

The demand change perturbation unit 210 generates a perturbed demand change parameter based on the demand change parameter probability distribution output by the demand change parameter probability distribution estimation unit 10. Here, adding perturbation to the demand change parameter means randomly raising or lowering the value of the demand change parameter according to a probability distribution. For example, based on the probability distribution parameters μ _a , σ _a , μ _b , and σ _b estimated by the demand change parameter probability distribution estimation unit 10, the probability distribution N (μ _a , σ _a ² ) of a is a ′. , B according to the probability distribution N (μ _b , σ _b ² ) of b. More specifically, as a ′, the probability that a ′ is included in the range of deviation from μ _a of ± 1σ _a is 68.27%, the probability that it is included in ± 2σ _a is 95.45%, and ± 3σ. _The probability of being included in a is 99.73%, ..., And the value of a'is determined. Similarly, the value of b ′ is determined according to the probability distribution. In this way, the demand change perturbation unit 210 generates a ′ and b ′ according to the probability distribution. When the probability distribution is other than the normal distribution, a ′ and b ′ based on the probability distribution are generated.

  The demand change expected value prediction unit 211 includes the demand change parameter generated by the demand change perturbation unit 210, the temperature predicted value t in the weather forecast at the first timing (for example, the previous day), and the weather at the second timing (for example, the day). The expected value of demand change is predicted based on the predicted temperature value t ′ in the prediction. Specifically, the demand change parameters a ′ and b ′ generated by the demand change perturbation unit 210 and the past, for example, the predicted temperature value t at the first timing and the predicted temperature value t at the second timing. The difference D between the demand forecast value and the actual value is calculated according to the model by using the difference d (= t−t ′) from “. For example, it is calculated based on a model represented by D = a'd + b ', and this is output as a prediction of the expected value of demand change.

  The price multiplier probability distribution prediction unit 22 includes a bid curve perturbation unit 220 and a price multiplier expected value prediction unit 221.

  The bid curve perturbation unit 220, based on the probability distribution of the bid curve parameter output by the bid curve parameter probability distribution estimation unit 11 and the contract data of the spot market price at the first timing, determines the bid curve parameter with the perturbation. To generate. For example, from the probability distribution of the bid curve parameters Pd, Wd, Ps, and Ws estimated by the bid curve parameter probability distribution estimation unit 11, similar to the demand change perturbation unit 210 described above, specific bid parameters according to the probability distribution are specified. The values Pd ', Wd', Ps 'and Ws' are generated and output. More specifically, Pd 'and Wd' are generated by the probability distribution according to the equation (1), and Ps 'and Ws' are generated by the probability distribution according to the equation (2). If the probability distribution is not a probability distribution using these error functions erf, it is generated according to the probability distribution.

  The price multiplier expected value prediction unit 221 determines the price multiplier based on the generated bid curve parameters Pd ′, Wd ′, Ps ′, and Ws ′ and the generated difference D between the demand forecast value and the actual value. Predict expected value. For example, the shapes of the demand curve and the supply curve are determined based on the bid curve parameters Pd ', Wd', Ps ', and Ws' to which the perturbation has been generated as described above. After that, the demand curve is shifted in the horizontal direction (power axis direction) using the difference D between the demand forecast value and the actual value, and the expected value of the price multiplier α that is the adjustment term is output as the forecast. That is, the demand curve is shifted in the electric power axis direction by adding the difference D between the generated demand forecast value and the actual value to the estimated Q formula.

  Explaining with reference to FIG. 3, first, each shape of the demand curve and the supply curve is predicted according to the probability distribution in each area. Then, the demand curve is shifted in the power axis direction by using the predicted value of the difference D between the predicted demand value and the measured value. Since the price at the equilibrium point, which is the point where the predicted demand curve and the predicted supply curve intersect, is the predicted value of the imbalanced price that is not area-corrected, the price multiplier α is calculated by dividing this predicted value and Pd. Predict.

  Returning to FIG. 4, the imbalance price probability distribution calculation unit 23 includes an imbalance price predicted value calculation unit 230 and an imbalance price predicted value storage unit 231.

  The imbalance price predicted value calculation unit 230 is based on the first timing, for example, the contract price P of the spot market data on the previous day, the price multiplier α predicted by the price multiplier expected value prediction unit 221, and the area correction value β. , Calculate the forecast value of the imbalance price. Specifically, the predicted value Pi of the imbalance price is calculated as Pi = Pα + β based on the input contract price P, the input area correction value β, and the predicted price multiplier α.

  The imbalance price predicted value storage unit 231 stores and holds the imbalance price predicted value Pi. The imbalance price prediction value storage unit 231 stores the imbalance price prediction values in the next imbalance price reliability calculation unit 24 until the set of imbalance price prediction values stored so far satisfies a predetermined condition. Output. On the other hand, when the predetermined condition is not satisfied, the process is returned to the demand change perturbation unit 210, another perturbation is added to the demand change and the price multiplier, and the calculation of the imbalance price forecast value Pi is repeated. Here, the predetermined condition is a condition for ending the iterative calculation, and for example, the set size of the predicted value Pi of the imbalance price exceeds 100, or the width of the confidence interval is calculated when the confidence interval is calculated. It is a condition such as until it becomes a certain value or less.

  The imbalance price reliability calculation unit 24 and the imbalance price reliability output unit 30 calculate the reliability of the imbalance price based on the probability distribution of the imbalance price Pi, as described in the common configuration described above. Convert it to an appropriate format and output so that it can be recognized by the user.

  The above is the configuration of the imbalance price prediction device 1 according to the present embodiment. Next, the operation of the imbalance price prediction device 1 according to the present embodiment will be described. FIG. 5 is a flowchart showing an example of a processing flow of the imbalance price prediction device 1 according to the present embodiment. In FIG. 5, a confidence interval is calculated as an example of the reliability. In addition, in each process of the following description, the process which is not described in detail shall perform the above-mentioned process.

  First, the imbalance price predicted value storage unit 231 is initialized (step S100). As described above, since the imbalance price predicted value storage unit 231 stores the imbalance price predicted value by the iterative process, the past data stored in the imbalance price predicted value storage unit 231 exists. In this case, initialization is performed by first deleting the stored data (imbalance price prediction value). Further, instead of erasing, the imbalance price prediction value stored so far may not be referred to. The imbalance price predicted value storage unit 231 may be initialized at least before the loop processing, and thus may be initialized at any timing before step S103.

  Next, the demand change parameter probability distribution estimation unit 10 estimates the probability distribution of the demand change parameter from the input past weather forecast value and the past power demand actual value (step S101). Subsequently, the bid curve parameter probability distribution estimation unit 11 estimates the probability distribution of bid curve parameters from past contract data (step S102).

  The order of these processes may be changed. That is, the probability distribution of the demand change parameter may be specified after estimating the probability distribution of the bid curve parameter. Since the loop processing is started from the next step S103, it is preferable that the processing is performed before this, but the processing may be performed in the loop if the data added every moment is used. At least if the probability distribution of the demand change parameter is estimated before the process of perturbing the demand change parameter and the probability distribution of the bid curve parameter is estimated before adding the perturbation to the bid curve parameter, these processes are It may be performed at any timing, or may be processed in parallel.

  Next, the process shifts to a loop process for calculating the imbalance price predicted value. In the loop process, first, the demand change perturbation unit 210 adds a perturbation to the demand change parameter according to the estimated probability distribution of the demand change parameter (step S103). Subsequently, the demand change expected value prediction unit 211 calculates the demand change expected value D based on the demand change parameter perturbed according to the probability distribution and the input temperature predicted value in the weather forecast of the previous day and the current day. Predict (step S104).

  Next, the bid curve perturbation unit 220 adds a perturbation to the bid curve parameters Pd, Wd, Ps, and Ws based on the input spot market price contract data and the estimated probability distribution of the bid curve parameters to calculate the probability. The bid curve parameters Pd ′, Wd ′, Ps ′, and Ws ′ according to the distribution are calculated (step S105). Next, the price multiplier expected value prediction unit 221 uses the bid curve parameters Pd ′, Wd ′, Ps ′, and Ws ′ that are perturbed according to the probability distribution, and the predicted demand change expected value D, based on the predicted demand change expected value D. , The expected value of the price multiplier α is predicted (step S106).

  Next, the imbalance price predicted value calculation unit 230 calculates the imbalance price predicted value based on the predicted expected value of the price multiplier α, the input contract price P of the spot market, and the input area correction value β. Pi (= αP + β) is calculated (step S107). The imbalance price predicted value storage unit 231 stores the calculated imbalance price predicted value Pi (step S108). The imbalance price prediction value storage unit 231 stores the imbalance price prediction value obtained in each loop on the basis of the parameters given different perturbations based on the probability distribution. That is, the imbalance price prediction values Pi calculated in each loop are stored one by one, and the data of the stored imbalance price prediction values Pi are accumulated.

  Next, it is determined whether or not the number of samples of the imbalanced price predicted value Pi stored and stored in the imbalanced price predicted value storage unit 231 is sufficient for calculating the confidence interval (step S109). .. When the number of samples of the imbalance price prediction value Pi has not reached the sufficient number of storages (step S109: No), the process returns to step S103, and the imbalance price prediction value Pi is set by another perturbed parameter. The process of calculating and storing is repeated (steps S103 to S108). In this way, by performing the Monte Carlo simulation, the probability distribution of the imbalance price forecast value Pi is calculated, and the imbalance price forecast accuracy is improved.

  When the number of samples of the imbalance price prediction value Pi has reached the number sufficient to calculate the confidence interval (step S109: Yes), the imbalance price reliability calculation unit 24 determines the imbalance price prediction value storage unit. The imbalance price prediction value Pi, which is an index for the user to estimate the reliability, based on the imbalance price prediction values accumulated in H.231, that is, the plurality of imbalance price prediction values Pi calculated according to the probability distribution. The confidence interval of is calculated (step S110).

  Next, the imbalance price reliability output unit 30 outputs the reliability of the calculated imbalance price predicted value Pi to the user or the like (step S111). For example, when outputting the reliability of the imbalance price prediction value to the user, it may be output using a drawing or a graph so that the user can recognize it via a monitor, or may be output as a numerical value as it is. .. When transferring data to another device, various numerical data may be transferred as numerical values which are raw data, or converted into data that can be easily processed by the imbalance price reliability output unit 30 and transferred. You may do it. Alternatively, the data may be processed and output each time based on the output target.

  FIG. 6 is an example of a graph showing the reliability of the imbalance price prediction value (expected value) and the imbalance price prediction value obtained by the imbalance price prediction device 1 by the above processing. The thick line located around the center of the vertical axis indicates the expected value of the imbalance price obtained by the present embodiment, and the thin lines above and below indicate the confidence interval of the imbalance price. In the example of FIG. 6, the value of the standard deviation is used as the confidence interval, and the upper and lower thin lines show the values of ± σ from the thick line in the center. That is, in this case, the closer the upper and lower lines are to the center line, the higher the reliability of the predicted value is.

  FIG. 7 is an enlarged view of a part of the graph shown in FIG. 6, and does not show ± σ of the expected value of the imbalance price as a graph but shows the standard deviation as an error bar at each expected value. It is a thing. By indicating the error bars as described above, the reliability of the imbalance price at each time with respect to the expected value can be displayed more easily than the graph shown in FIG.

  FIG. 8 is an enlarged view of a part of the graph shown in FIG. 6, and does not show ± σ of the expected value of the imbalance price as a graph. The histogram is shown in shades of color. In this way, by showing the histogram by the shades of colors, it can be seen that the reliability is higher as the spread of the dark areas is narrower. Of course, the histogram at each time may be output as a separate table or graph.

  As shown in FIGS. 6 to 8, there are various kinds of graphs to be output to the user, and it is possible to select and output the graphs according to the use and the preference of the user. In addition to the above-mentioned ones, necessary statistics and other indicators may be output. As described above, the imbalance price reliability output unit 30 outputs the value calculated by the imbalance price reliability calculation unit 24 as a numerical value, a time series graph, an error bar in the time series graph, a histogram, or the like.

  As described above, according to the present embodiment, the imbalance price prediction device 1 outputs the reliability of the imbalance price together with the predicted value of the imbalance price using the Monte Carlo method based on the past and present data. Thus, the user or the like reads the reliability of the prediction result of the imbalance price prediction based on the information output by the imbalance price reliability output unit 30 and uses it as reference information for making a more rational procurement plan. Is possible. Regarding the imbalanced price, the price and transaction volume will be disclosed at a later date, but the shape of the supply and demand curve will not be disclosed. Therefore, by acquiring the reliability of the imbalance price predicted based on the probability distribution of the shape of the supply and demand curve by the imbalance price prediction device 1 according to the present embodiment together with the expected value of the imbalance price, the reliability can be improved. It is possible to predict a high imbalance price.

(Second embodiment)
FIG. 9 is a block diagram showing an example of the imbalance price prediction device 1 according to the second embodiment. As shown in FIG. 3, the imbalance price prediction device 1 according to the present embodiment includes a demand change parameter probability distribution estimation unit 10 and a bid curve parameter probability distribution estimation unit 11 as in the first embodiment described above. An imbalance price reliability estimation unit 20 and an imbalance price reliability output unit 30 are provided. The imbalance price reliability prediction unit 20 includes a demand change probability distribution prediction unit 21, a price multiplier probability distribution prediction unit 22, an imbalance price probability distribution calculation unit 23, an imbalance price reliability calculation unit 24, and Equipped with. Hereinafter, parts different from the above-described first embodiment will be described in detail.

  The demand change probability distribution prediction unit 21 is similar to the first embodiment in that it includes the demand change expected value prediction unit 211, but includes a demand change variance prediction unit 212 instead of the demand change perturbation unit 210. The demand change variance prediction unit 212 predicts the demand change variance based on the past data and the probability distribution of the demand change parameters. The past data is, for example, predicted temperature values of the previous day and the current day, or spot market contract data of the previous day. As an example, in the present embodiment, the variance of demand change is obtained by multiplying the total demand amount Q in the time period by a predetermined value.

  As another example, the variance of demand change is predicted by building a learning model using the above data. In this case, a learning model based on a time series and various conditions is constructed based on the accumulated past data. A generally used learning method can be used to construct the learning model. For example, a Hopfield type neural network, a support vector machine, a random forest, or the like can be adopted. Further, it may be an algorithm including deep learning. A model may be constructed by supervised learning based on the predicted temperature value of past data, the actual temperature data, and the contract data of the spot market price at that time. Also, a simple linear model or the like may be constructed instead of machine learning.

  That is, when the predicted temperature values of the previous day and the current day, the contract data of the spot market price of the previous day, and the probability distribution of the demand change parameter are input, a model for estimating the variance of the demand change is constructed. As the simplest example, the processing in the present embodiment may be one in which a product of the total demand Q in the time zone of interest multiplied by a constant is output as the variance prediction value. As another example, a learning model in which a linear model is prepared and the coefficient is sequentially corrected based on the input data at each time may be used. Further, the model may be based on a higher-order equation, or may be based on an exponent or a logarithm. Further, it may be a model in which a model of the neural network as described above is constructed and the variance is output by changing the point of firing depending on the input condition.

The price multiplier probability distribution prediction unit 22 is similar to the first embodiment in that the price multiplier expected value prediction unit 221 is provided, but the price multiplier variance prediction unit 222 is provided instead of the bid curve perturbation unit 220. The price multiplier variance prediction unit 222 predicts the variance value of the price multiplier based on the past data and the probability distribution of bid curve parameters. The past data is, for example, spot market contract data on the previous day. Specifically, by calculating how the price multiplier α for adjusting the imbalance amount changes from the slope of the tangent line at the intersection of the demand curve and the supply curve approximated by the error function erf, the variance value obtained by the linear approximation To calculate. As an example, the variance σ ² (α) of the price multiplier is predicted and output by the following Expression (4).

Here, α is a price multiplier, ΔD is a demand change amount, Δθ _i is each bid curve parameter, σ ² (ΔD) is a demand change variance, and σ ² (Δθ _i ) is each bid. It is the variance of the curve parameters. In addition, although the method of using the first-order approximation for the prediction of the variance of the price multiplier has been described, the present invention is not limited to this. It may be predicted by a method.

The demand change expected value prediction unit 211 and the price multiplier expected value prediction unit 221 basically perform the same processing as in the first embodiment, but the input value is repeatedly perturbed as in the first embodiment. Rather than calculating, calculate only once. That is, the expected value of the demand change parameter, for example, μ _a and μ _b when the coefficients a and b are represented by the above-described normal distribution are input to the demand change expected value prediction unit 211, and these expected values are input. Is also used to output the expected value E (ΔD) of the demand change. Similarly, the expected value E (Pd), E (Wd), E (Ps), E (Ws) of each bid curve parameter and the expected value E (ΔD) of demand change are input to the price multiplier expected value prediction unit 221. The equilibrium point between the input and demand curves is calculated, and the price at the calculated equilibrium point is divided by E (Pd) to output the expected value E (α) of the price multiplier.

The imbalance price probability distribution calculation unit 23 calculates the expected value E (α) of the price multiplier predicted by the expected price multiplier value prediction unit 221 and the variance σ ² (α) of the price multiplier predicted by the price multiplier variance prediction unit 222. , The probability distribution of the imbalance price Pi is calculated based on the input area correction value β. In this calculation, the imbalance price probability distribution Pi is calculated based on the formula Pi = Pα + β, as in the first embodiment described above. Note that, unlike the above-described first embodiment, the expected value E (α) of the price multiplier and the variance σ ² (α) of the price multiplier are input as the probability distribution of the price multiplier α. That is, unlike the first embodiment, the probability distribution Pi of the imbalance price is calculated without performing the iterative process. For example, when the price multiplier α is represented by a normal distribution, it is represented by a distribution of α to N (E (α), σ ² (α)), so that the imbalance price probability distribution Pi = P × N (E It is expressed as (α), σ ² (α)) + β. Further, the distribution is not limited to this, and may be output according to various probability distributions as in the first embodiment described above, or may be output as a histogram from the obtained expected value and variance. .

  The above is the configuration of the imbalance price prediction device 1 according to the present embodiment. Next, the operation of the imbalance price prediction device 1 according to the present embodiment will be described. FIG. 10 is a flowchart showing an example of the processing flow of the imbalance price prediction device 1 according to the present embodiment. In addition, in the following description, if the details are not described, the processes described in the common configuration and the description of the present embodiment are performed.

  First, the demand change parameter probability distribution estimation unit 10 estimates the probability distribution of the demand variation parameter based on the input past weather forecast value, the actual power demand value, and the like (step S200). Next, the bid curve parameter probability distribution estimation unit 11 estimates the probability distribution of the bid curve parameter based on the input past (previous day) contract market price contract data (step S201).

Next, the demand change expected value prediction unit 211 uses the estimated probability distribution of the demand change parameter and the input past weather forecast values (first timing and second timing), for example, the temperature forecast values of the previous day and the current day. Based on the above, the demand change expected value E (ΔD) is predicted (step S202). Subsequently, the demand change variance prediction unit 212 predicts the demand change variance σ ² (ΔD) based on the estimated probability distribution of the demand change parameters (step S203). The order of predicting these values is not specified, and either value may be predicted.

Next, the expected price multiplier value prediction unit 221 receives the probability distribution of the input bid variable parameters, the past (first timing and second timing) weather forecast values, and the past (first timing) spot market price contract data. Then, based on the predicted demand change expected value E (ΔD), the expected price multiplier value E (α) is predicted (step S204). Subsequently, the price multiplier variance prediction unit 222 predicts the variance σ ² (α) of the price multiplier based on the probability distribution of bid curve parameters and the predicted variance σ ² (ΔD) of demand change (( Step S205). Also, the order of predicting these values is not fixed, and either value may be predicted.

Next, the imbalance price probability distribution calculation unit 23 calculates the predicted price multiplier expected value E (α), the predicted price multiplier variance value σ ² (α), and the input past (first timing). The probability distribution of the imbalance price Pi is calculated based on the contract data of the spot market price and the input area correction value β (step S206). In the present embodiment, since the probability distribution of the imbalance price Pi is calculated based on the expected value and the variance obtained as the random variables, unlike the above-described first embodiment, the imbalance price prediction value is repeatedly processed. Therefore, it is not calculated and stored more than once. That is, instead of performing Monte Carlo simulation, for example, the imbalance price is predicted by a first-order approximation.

  Next, the imbalance price reliability calculation unit 24 calculates the reliability of the imbalance price based on the probability distribution of the imbalance price Pi (step S207).

  Next, the imbalance price reliability output unit 30 converts the reliability of the imbalance price into an appropriate format and outputs it (step S208). As a different output method from the above-described first embodiment, for example, the expected value of the calculated imbalance price Pi satisfies the condition registered in advance, and the reliability of the calculated imbalance price is registered in advance. If the specified conditions are satisfied, a warning may be notified to the user. More specifically, if the imbalance price Pi exceeds the price specified in advance and the probability exceeds a predetermined threshold value and the situation becomes probable, it is possible to reliably procure electric power even if the procurement cost is high. A possible application is to warn the user so that it can be implemented. On the contrary, if the imbalance price Pi becomes lower than a predetermined price specified in advance, a warning that the procurement of electric power may fail, a warning that prompts prediction again, or a caution The user may be alerted to a warning that he / she will trade.

  FIG. 11 is a diagram showing expected values and confidence intervals for imbalance price prediction according to this embodiment. The confidence interval indicates a range of ± σ (standard deviation) centered on the expected value of the imbalance price prediction. The thick line in the center is the expected value of the imbalance price forecast, and the upper and lower thin lines show ± σ. As a comparison, the result according to the first embodiment shown in FIG. As can be seen from this figure, the predicted value of the imbalance price according to the present embodiment has a smoother shape as a whole than the predicted value of the imbalance price according to the first embodiment. As described above, a difference appears in these predicted values, and thus the price may be predicted by selecting a highly reliable one calculated from these two methods.

  Specifically, the prediction result obtained by the Monte Carlo simulation as in the first embodiment does not change the price even if the shape is greatly changed, and thus the high reliability is obtained even if the fluctuation rate of the bid price curve with time is high. It is possible to secure. On the other hand, if the change in the variation rate at each time is small, that is, if the change in the slope of the curve is small, or if the variation rate at each time is small in the first place, the prediction by the linear approximation as in the present embodiment is performed. This makes it possible to ensure high reliability. By predicting the imbalance price in consideration of each of these characteristics, it is possible to predict the imbalance price with higher reliability.

  As described above, according to the present embodiment, the imbalance price is predicted by the first-order approximation having a characteristic different from that of the Monte Carlo simulation of the first embodiment described above, so that the user can perform an event related to the imbalance price, for example, an imbalance price. It will be possible to predict the rise in the balance price in advance, and it will be possible to further rationalize the imbalance cost when procurement fails and the cost of procurement itself.

(Third Embodiment)
The configuration of the imbalance price prediction device 1 according to this embodiment is the same as that of the imbalance price prediction device 1 according to the first embodiment or the second embodiment described above. The difference is the processing content of the imbalance price reliability output unit 30.

  First, up to the prediction of the probability distribution of the imbalance price, calculation is performed using the method of each of the above-described embodiments. Next, the imbalance price reliability calculation unit 24 calculates, as the reliability of the imbalance price, the probability distribution of the total imbalance settlement estimated from the probability distribution of the imbalance price. That is, instead of calculating the imbalance price at each time, the predicted value of the total amount of imbalance risk or the like is calculated. For example, the probability distribution of the imbalance settlement total is calculated by dividing the time every 30 minutes, predicting the imbalance price in each section, and calculating the sum of the predicted values.

  Next, the imbalance price reliability output unit 30 outputs the probability distribution of the imbalance settlement total amount as information regarding prediction of the imbalance settlement total amount when the power procurement fails. As another example, the imbalance price reliability output unit 30 converts the index related to the imbalance price calculated by the imbalance price reliability calculation unit 24 into an index related to the imbalance settlement total amount and outputs the index. Good.

  As described above, according to the present embodiment, in addition to the above-described embodiment, the index related to the total amount of imbalance settlement is output, and therefore, the user pays attention to the cost when the power procurement which the user himself is considering fails. , It becomes possible to understand as a total amount.

(Fourth Embodiment)
FIG. 12 is a block diagram showing an example of the power trading system 2 according to the present embodiment. The configuration of the imbalance price prediction device 1 according to the present embodiment is the same as that of the imbalance price prediction device 1 according to each of the above-described embodiments. 2 includes a transaction device 40.

  The functions and processing of the imbalance price prediction device 1 are the same as those in the above-described embodiments. However, the imbalance price reliability output unit 30 does not output the imbalance price reliability as visual information toward the user or as information that can be perceived by humans, but as imbalance price reliability as numerical data. Is output.

  The imbalance price reliability output from the imbalance price reliability output unit 30 is input to the transaction device 40. Then, the transaction device 40 performs a transaction automatically or semi-automatically based on the input index. For example, when variance is input as the confidence interval, the transaction device 40 automatically calculates the bid price according to the certainty obtained from the variance of the predicted values of the imbalance price, and automatically uses the bid price. Tender. For example, when the reliability is higher than a predetermined value, the transaction device 40 automatically bids at the input imbalance price. On the other hand, when the reliability is equal to or lower than the predetermined value, the imbalance price reliability output unit 30 may output an alert to the user, and the user may input the bid price to the transaction device 40. By doing so, it is possible to save the user the trouble of checking the output result and bidding.

  As another example, the imbalance price reliability output unit 30 transfers the numerical data of the imbalance price probability distribution output by the imbalance price probability distribution calculation unit 24 or the parameter data to the trading device 40 as it is. You may do so. In this case, the transaction device 40 appropriately predicts the imbalance price necessary for the transaction, or calculates the reliability of the imbalance price, and performs the transaction based on the predicted value or the calculated value.

  As described above, according to the present embodiment, the user does not consider the imbalance risk by himself, but the power trading system 2 determines the reliability of the imbalance price predicted by the imbalance price prediction device 1. It is possible to carry out optimal automatic transactions or semi-automatic transactions with consideration.

  In the embodiment described above, the price is estimated using the price of the market one hour ago and the spot market price of the day before, but the imbalance price forecast is performed using the preliminary value of the price multiplier α. May be. Further, in addition to the prediction of the above-described embodiment, the imbalance price at the time of executing the plan may be predicted using the preliminary value of the price multiplier α.

  FIG. 13 is a diagram illustrating an example of the computer 100 that constitutes each of the above-described embodiments. The computer 100 includes a processor 101, an input device 102, a display device 103, a communication device 104, and a storage device 105. The bus 106 interconnects the processor 101, the input device 102, the display device 103, the communication device 104, and the storage device 105 with each other.

  The processor 101 is an electronic circuit including a control device and a computing device of the computer 100. Examples of the processor 101 include a general purpose processor, a central processing unit (CPU), a microprocessor, a digital signal processor (DSP), a controller, a microcontroller, a state machine, an application specific integrated circuit (ASIC), and a field programmable gate array (FPGA). ), Programmable logic circuits (PLDs), and combinations thereof.

  The processor 101 performs arithmetic processing based on data and programs input from each device connected via the bus 106, and outputs an arithmetic result and a control signal to each connected device. Specifically, the processor 101 executes an operating system (OS) of the computer 100, an imbalance price forecasting program (hereinafter referred to as a forecasting program), and controls each device constituting the computer 100.

  The prediction program is a program that causes the computer 100 to realize the configuration of each of the above-described embodiments of the prediction device. The prediction program is stored in a non-transitory, tangible, computer-readable storage medium. The storage medium is, for example, an optical disc, a magneto-optical disc, a magnetic disc, a magnetic tape, a flash memory, or a semiconductor memory, but is not limited to these. The computer 100 functions as the imbalance price prediction device 1 by the processor 101 executing the prediction program.

  The input device 102 is a device for inputting information to the computer 100, and is, for example, a keyboard, a mouse, or a touch panel, but is not limited to these.

  The display device 103 is a device for displaying images and videos, and is, for example, a liquid crystal display (LCD), a cathode ray tube (CRT), a plasma display (PDP), or a display with a touch panel, but is not limited to these. Absent. The display device 103 may display the output data of each functional configuration, may display the result output by the imbalance price reliability output unit 30, or may output the imbalance price reliability output. It may be a part of the unit 30.

  The communication device 104 is a device for the computer 100 to communicate with an external device wirelessly or by wire, and is, for example, a modem, a hub, or a router, but is not limited thereto. Information such as meteorological data and past spot contract data can be input from an external device via the communication device 104.

  The storage device 105 is a storage medium that stores the OS of the computer 100, a prediction program, data necessary for executing the prediction program, data generated by executing the prediction program, and the like. The storage device 105 includes a main storage device and an external storage device. The main storage device is, for example, a random access memory (RAM, DRAM, SRAM), but is not limited to these. The external storage device is, for example, a hard disk (HDD), an optical disk, a flash memory, or a magnetic tape, but is not limited to these. For example, the imbalanced price forecast value storage unit 231 and the like in the above-described first embodiment may be built in the storage device 105 or an external server on the network and can be accessed via the communication device 104. You may do it.

  The computer 100 may include one or a plurality of each of the processor 101, the input device 102, the display device 103, the communication device 104, and the storage device 105. Peripheral devices such as a printer and a scanner are connected to the computer 100. Good.

  Further, the imbalance price prediction device 1 may be configured by a single computer 100, or may be configured as a system including a plurality of computers 100 connected to each other. Furthermore, the prediction program may be stored in advance in the storage device 105 of the computer 100, may be stored in a storage medium external to the computer 100, or may be uploaded to a server on the network. In any case, the function of the imbalance price prediction device 1 is realized by installing and executing the prediction program in the computer 100.

  Although some embodiments of the present invention have been described, these embodiments are presented as examples and are not intended to limit the scope of the invention. These novel embodiments can be implemented in various other forms, and various omissions, replacements, and changes can be made without departing from the spirit of the invention. These embodiments and their modifications are included in the scope and gist of the invention, and are also included in the invention described in the claims and the scope equivalent thereto. Further, it goes without saying that these embodiments can be partially combined as appropriate within the scope of the gist of the present invention.

1: Imbalance price prediction device, 10: Demand change parameter probability distribution estimation unit, 11: Bid curve parameter probability distribution estimation unit, 20: Imbalance price reliability prediction unit, 21: Demand change probability distribution prediction unit, 22: Price Multiplier probability distribution prediction unit, 23: imbalance price probability distribution calculation unit, 24: imbalance price reliability calculation unit, 30: imbalance price reliability output unit, 40: transaction device

Claims (17)

Estimating the demand change parameter probability distribution difference is the probability distribution of the parameters of the models representing the predicted value and the actual value of the power demand on the basis of the difference between the parameter of air temperature predicted values at two timings different demand A change parameter probability distribution estimation unit ,
A bid curve parameter probability distribution estimation unit that estimates a bid curve parameter probability distribution, which is a probability distribution of parameters representing the shape of a bid curve, based on contract data relating to past transaction volume in a specific market,
A weather forecast value of the second timing after the first timing and the first timing, on the basis of the demand change parameter probability distribution, in, for the third timing of power demand after the second timing, the power demand A demand change probability distribution prediction unit that predicts a demand change probability distribution that is a probability distribution of the difference between the predicted value and the actual value,
With a contract data of the past of a specific market, and the bid curve parameter probability distribution, and the demand change probability distribution, based on, calculates the y imbalance value Kaku確 index distribution, the imbalance price probability distribution calculating unit Information processing equipment . The information processing apparatus according to claim 1, wherein the past contract data includes information on a past bid amount, a past contract amount, and a past contract price. On the basis of the past specific market contract data, the bid curve parameter probability distribution, and the demand change probability distribution, an adjustment term for the predicted value of the imbalance price at the third timing and the past contract price A price multiplier probability distribution predicting unit that predicts a price multiplier probability distribution that is a probability distribution of a price multiplier is further provided,
The imbalance price probability distribution calculation unit calculates the imbalance price probability distribution based on the past specific market contract data and the price multiplier probability distribution. Information processing device. The demand change probability distribution predicting unit generates a demand change parameter by adding a perturbation according to the demand change parameter probability distribution, a demand change perturbation unit, and predicts a demand change expected value based on the demand change parameter. A change expected value prediction unit,
The price multiplier probability distribution prediction unit generates a bid curve parameter that is perturbed according to the bid curve parameter probability distribution, a bid curve perturbation unit, and predicts a price multiplier expected value based on the bid curve parameter, a price An expected multiplier value prediction section,
The imbalance price probability distribution calculation unit includes an imbalance price prediction value calculation unit that calculates an imbalance price prediction value,
The information processing apparatus according to claim 3 , further comprising an imbalance price predicted value storage unit that stores the imbalance price predicted value. The demand change probability distribution prediction unit predicts the demand change expected value based on the weather forecast values at the first timing and the second timing and the expected value of the demand change parameter probability distribution. A value prediction unit, and a demand change variance prediction unit that predicts a variance of demand changes based on the demand change parameter probability distribution,
The price multiplier probability distribution predicting unit predicts a price multiplier expected value based on the past spot market contract data and an expected value of the bid curve parameter probability distribution, a price multiplier expected value predicting unit, and The information processing apparatus according to claim 3 , further comprising: a price multiplier variance prediction unit that predicts a variance of the price multiplier based on the bid curve parameter probability distribution. At least one of the demand change parameter probability distribution, the bid curve parameter probability distribution, the demand change probability distribution, the price multiplier probability distribution, and the imbalance price probability distribution has two indices of an expected value and a variance. the information processing apparatus according to any one of claims 3 to 5 represented by the normal distribution represented by. The demand change parameter probability distribution, the bid curve parameter probability distribution, the demand change probability distribution, the price multiplier probability distribution, and at least one of the imbalance price probability distribution 3 to claim represented by the histogram The information processing device according to claim 5 . Based on the imbalance price probability distribution, calculates the reliability of the imbalance price, an imbalance price reliability calculation unit,
An imbalance price reliability output unit that outputs the reliability of the imbalance price,
The information processing apparatus according to any one of claims 1 to 7 , further comprising: The information according to claim 8 , wherein the imbalance price reliability calculation unit calculates the variance of the expected value of the imbalance price, the standard deviation of the expected value of the imbalance price, or the confidence interval of the expected value of the imbalance price. Processing equipment . The imbalance price reliability output unit, the imbalance price reliability value calculation unit has calculated, numeric, time series graph, the error bars in the time-series graph, or claim 8 or claim output as histogram 9 The information processing device described in 1. The imbalance price reliability output unit, the value the imbalance price reliability calculation unit has calculated exceeds a predetermined threshold, or to claim 8 or claim 9 to notify the user as an alert if below The information processing device described. The imbalance price reliability calculation unit calculates the probability distribution of the total imbalance settlement estimated from the imbalance price probability distribution,
The information processing according to claim 8 or 9 , wherein the imbalance price reliability output unit outputs the probability distribution of the imbalance settlement total amount as information regarding prediction of the imbalance settlement total amount when power procurement fails. Equipment . 13. The demand change parameter probability distribution estimation unit estimates the demand change parameter probability distribution based on a past weather forecast value and a past actual power demand value. The information processing device described in 1. The imbalance price reliability output unit, and the information processing apparatus according to any one of the imbalance price reliability calculation unit outputs the reliability of the imbalance price calculated, claims 8 to 12,
The reliability of the imbalance price output by the imbalance price reliability output unit is transferred, and based on the reliability of the transferred imbalance price, automatically performs power trading, and a transaction device,
An information processing system including. The imbalance price reliability output unit, and the information processing apparatus according to any one of the imbalance price probability distribution calculating unit outputs the imbalance price probability distribution calculated, claims 8 to 12,
The imbalance price probability distribution output by the imbalance price reliability output unit is transferred, and based on the transferred imbalance price probability distribution, automatically performs power trading, a transaction device,
An information processing system including. Demand change parameter probability distribution estimation unit, the probability distribution of the parameters of the model representing the difference between the predicted and actual values of the power demand on the basis of the difference between the parameter of air temperature predicted values at two timings different demand Estimating a change parameter probability distribution,
Bid curve parameter probability distribution estimation unit, a step of, based on a trade data regarding the amount of the transaction historical specific market, estimates the bid curve parameters probability distribution is the probability distribution of the parameters representing the shape of the tender curve,
Demand change the probability distribution predicting unit, and a weather forecast value of the second timing after the first timing and the first timing, and the demand change parameter probability distribution, based on the third timing after the second timing For the power demand of, a step of predicting a demand change probability distribution, which is a probability distribution of the difference between the predicted value of the power demand and the actual value,
Price multiplier probability distribution predicting unit includes a contract data of the past of a specific market, and the bid curve parameter probability distribution, and the demand change probability distribution, based on, and calculating the probability distribution of the y imbalance Price Information processing method . Computer ,
Means for estimating the demand change parameter probability distribution is the probability distribution of the parameters of the model representing the difference between the predicted and actual values of the power demand on the basis of the difference between the parameter of air temperature predicted values at two timings different,
Means for estimating a bid curve parameter probability distribution, which is a probability distribution of parameters representing the shape of the bid curve, based on past contract data relating to the trading volume of the specific market,
A weather forecast value of the second timing after the first timing and the first timing, on the basis of the demand change parameter probability distribution, in, for the third timing of power demand after the second timing, the power demand Means for predicting the demand change probability distribution, which is the probability distribution of the difference between the predicted value and the actual value,
And commitments data for the last of a particular market, and the bid curve parameter probability distribution, and the demand change probability distribution, based on, means for calculating a probability distribution of the y imbalance price,
Program to function as.