Calculation method, program and calculation device

Abstract

To provide a computation method, program and computing device for appropriately supporting transactions in electricity markets.SOLUTION: A computation method disclosed herein comprises: setting bidding conditions for an electricity market; computing a predicted market environment value representing at least one of the amount of power supplied to the electricity market, the amount of power demand in the electricity market, and electricity prices in the electricity market; and computing predicted values of profit and loss with a probability distribution if transactions are made in the electricity market according to the bidding conditions on the basis of the bidding conditions and the predicted market environment value.SELECTED DRAWING: Figure 7

Description

[Technical Field]

[0001] The present disclosure relates to a computing method, a program, and a computing device. [Background technology]

[0002] In an electricity market where electricity is traded, traders decide in advance the selling price and supply amount of electricity and then make bids. For example, Patent Document 1 describes a bidding support system that determines a bidding curve for the market based on a given market electricity supply curve and electricity demand curve so that the amount of electricity, profits, or supply amount that is successfully bid in the market is maximized. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Application Laid-Open No. 2005-339527 Summary of the Invention [Problem to be solved by the invention]

[0004] In Patent Document 1, the predicted value of profit is calculated based on a preset power supply curve and power demand curve. However, since the amount of power supply and the amount of power demand fluctuate depending on weather conditions and the operation status of facilities, the predicted value of profit and loss may deviate from the actual profit and loss, which may make it impossible to appropriately support transactions in the electricity market.

[0005] The present disclosure is intended to solve the above-mentioned problems, and aims to provide a calculation method, a program, and a calculation device that can appropriately support transactions in the electricity market. [Means for solving the problem]

[0006] The calculation method according to the present disclosure includes the steps of: setting bidding conditions for an electricity market; calculating a predicted value of a market environment, which is at least one of the amount of electricity supplied to the electricity market, the amount of electricity demand in the electricity market, and the electricity price in the electricity market; and calculating, based on the bidding conditions and the predicted value of the market environment, a predicted value of profit or loss when trading in the electricity market using the bidding conditions, so as to have a probability distribution.

[0007] The program according to the present disclosure causes a computer to execute the steps of setting bidding conditions for an electricity market, calculating a predicted value of the market environment, which is at least one of the amount of electricity supplied to the electricity market, the amount of electricity demand in the electricity market, and the electricity price in the electricity market, and calculating, based on the bidding conditions and the predicted value of the market environment, a predicted value of profit and loss when trading in the electricity market using the bidding conditions, so as to have a probability distribution.

[0008] The computing device according to the present disclosure includes a bidding condition setting unit that sets bidding conditions for an electricity market; a market environment calculation unit that calculates a predicted value of a market environment, which is at least one of the amount of electricity supplied to the electricity market, the amount of electricity demand in the electricity market, and the electricity price in the electricity market; and a profit and loss calculation unit that calculates a predicted value of profit and loss when trading in the electricity market using the bidding conditions, based on the bidding conditions and the predicted value of the market environment, so as to have a probability distribution. [Effects of the Invention]

[0009] According to the present disclosure, transactions in the electricity market can be appropriately supported. [Brief explanation of the drawings]

[0010] [Figure 1] FIG. 1 is a schematic diagram for explaining the electricity market. [Figure 2] FIG. 2 is a schematic block diagram of the arithmetic unit. [Figure 3] FIG. 3 is a graph showing an example of a forecast value of the market environment. [Figure 4]FIG. 4 is a graph showing an example of predicted values ​​of the market environment. [Figure 5] FIG. 5 is a graph for explaining an example of predicted profit and loss values. [Figure 6] FIG. 6 is a graph for explaining an example of predicted profit and loss values. [Figure 7] FIG. 7 is a flowchart illustrating the calculation flow of the predicted value of profit and loss. [Figure 8] FIG. 8 is a schematic block diagram of a calculation device according to the second embodiment. [Figure 9] FIG. 9 is a flowchart illustrating the flow of the bidding condition optimization process. DETAILED DESCRIPTION OF THE INVENTION

[0011] Preferred embodiments of the present disclosure will be described in detail below with reference to the accompanying drawings. Note that the present disclosure is not limited to these embodiments, and when there are multiple embodiments, the present disclosure also includes configurations in which the respective embodiments are combined.

[0012] (First embodiment) (Electricity Market) FIG. 1 is a schematic diagram illustrating an electricity market. As shown in FIG. 1, the electricity market M is a market where an entity that purchases electricity and an entity that sells electricity trade electricity through bidding. That is, an entity that purchases electricity purchases electricity through the electricity market M, and the purchased amount of electricity is supplied to a demand facility 12 belonging to that entity during a time period specified at the time of purchase. The demand facility 12 is an electricity consuming facility such as a factory. An entity that sells electricity sells electricity through the electricity market M, and the sold amount of electricity is supplied from a supply facility 14 belonging to that entity during a time period specified at the time of sale. The supply facility 14 is an electricity supply facility such as a power plant or a natural energy power generation device. Here, a single electricity market M may include a plurality of demand facilities 12 (entities that purchase electricity) and a plurality of supply facilities 14 (entities that sell electricity). There may also be a plurality of electricity markets M, and each demand facility 12 and supply facility 14 may trade in a plurality of electricity markets M. For example, electricity markets M include a day-ahead market where transactions take place the day before electricity is supplied, an intraday market where transactions take place on the day electricity is supplied, and an adjustment market used to adjust the amount of electricity.

[0013] Here, it is required to appropriately support electricity trading in the electricity market M for at least one of a supply facility 14 (an entity that sells electricity) that sells electricity via the electricity market M and a demand facility 12 (an entity that purchases electricity) that purchases electricity via the electricity market M. The supply facility 14 or the demand facility 12 that is the target of trading support will hereinafter be referred to as a target facility 10. In this embodiment, a calculation device 20 executes calculations regarding electricity trading by the target facility 10, thereby making it possible to appropriately support trading in the electricity market M. The calculation device 20 will be described below.

[0014] (arithmetic device) Fig. 2 is a schematic block diagram of a calculation device. The calculation device 20 is a computer, and as shown in Fig. 2, has an input unit 22, an output unit 24, a communication unit 26, a storage unit 28, and a control unit 30. The calculation device 20 may be configured as a stand-alone device, may be configured integrally with other devices, or may be configured as a system combining various devices such as a calculation circuit and a data server, and is not particularly limited.

[0015] The input unit 22 is a mechanism for accepting user input and may be, for example, a mouse, keyboard, or touch panel. The output unit 24 is a device for outputting information and, in this embodiment, may be a display device for displaying images. The output unit 24 may also be provided with a device other than a display device, such as a speaker. The communication unit 26 is a communication module for communicating with an external device, such as an antenna. The arithmetic device 20 communicates with the external device via wireless communication, but wired communication is also acceptable, and any communication method may be used. The storage unit 28 is a memory for storing various information, such as the contents of calculations and programs performed by the control unit 30, and may include, for example, at least one of a random access memory (RAM), a main storage device such as a read-only memory (ROM), and an external storage device such as an HDD (hard disk drive). The program for the control unit 30 stored in the storage unit 28 may be stored in a recording medium readable by the arithmetic device 20.

[0016] The control unit 30 is a processing device that executes calculations and includes, for example, a calculation circuit such as a CPU (Central Processing Unit). The control unit 30 includes a bidding condition setting unit 32, a market environment calculation unit 34, a profit and loss calculation unit 36, and an output control unit 38. The control unit 30 implements the bidding condition setting unit 32, the market environment calculation unit 34, the profit and loss calculation unit 36, and the output control unit 38 by reading and executing a program (software) from the storage unit 28. The control unit 30 may execute these processes using a single CPU, or may be equipped with multiple CPUs and execute the processes using the multiple CPUs. Furthermore, at least a portion of the processes performed by the bidding condition setting unit 32, the market environment calculation unit 34, the profit and loss calculation unit 36, and the output control unit 38 may be implemented using hardware circuits.

[0017] (Bidding Condition Setting Department) The bidding condition setting unit 32 sets bidding conditions for the target facility 10 to the electricity market M. The bidding conditions refer to the amount of electricity traded (the bid energy amount) via the electricity market M, and in this embodiment, refer to at least one of the amount of electricity planned to be sold (the amount of electricity sold) and the amount of electricity planned to be purchased (the amount of electricity purchased). The bidding conditions refer to conditions related to the supply of electricity from the target facility 10. The type of bidding conditions may be determined appropriately depending on the type of support requested by the target facility 10, and the bidding condition setting unit 32 sets the value of the determined type of bidding conditions. In addition to the bid energy amount, the bidding conditions may also include information indicating a time period during which electricity equivalent to the bid energy amount is to be supplied (bidding time period). The time period here may refer to a length of time in units of month, day, and time (for example, from 1:00 PM on January 1st to 3:00 PM on January 1st), and the same applies hereinafter. This makes it possible to take into account the effects of seasons and day and night.

[0018] The bidding condition setting unit 32 may set the bidding conditions in any manner. For example, the bidding condition setting unit 32 may set the bidding conditions automatically, or may set the bidding conditions input by a user as the bidding conditions. Furthermore, the bidding condition setting unit 32 may set the bidding conditions for each electricity market M.

[0019] (Market Environment Calculation Department) The market environment calculation unit 34 calculates a predicted value of the market environment during the bidding time period. The market environment calculation unit 34 calculates a predicted value of the market environment during the bidding time period based on the bidding time period indicated in the bidding conditions. The market environment is a parameter that affects profit and loss when trading electricity through the electricity market M. In this case, profit and loss refers to revenue from electricity sales when electricity is sold, and refers to loss (payment amount) from electricity purchase when electricity is purchased. The market environment calculation unit 34 may calculate a predicted value of the market environment for each electricity market M.

[0020] The market environment is at least one of the amount of power supplied to the electricity market M, the amount of power demanded in the electricity market M, and the electricity price in the electricity market M. More specifically, in this embodiment, the market environment refers to all of the amount of power supplied, the amount of power demanded, and the electricity price. That is, in this embodiment, the market environment calculation unit 34 calculates the predicted values ​​of the market environment, including the predicted value of the amount of power supplied during the bidding time slot, the predicted value of the amount of power demanded during the bidding time slot, and the predicted value of the electricity price during the bidding time slot. The predicted value of the amount of power supplied during the bidding time slot refers to the amount of power predicted to be supplied via the electricity market M from supply facilities 14 other than the target facility 10 during the bidding time slot. Furthermore, the predicted value of the amount of power demanded during the bidding time slot refers to the amount of power predicted to be demanded via the electricity market M from demand facilities 12 other than the target facility 10 during the bidding time slot. Furthermore, the predicted value of the electricity price during the bidding time slot refers to the price of electricity per unit amount in the electricity market M during the bidding time slot.

[0021] 3 and 4 are graphs showing examples of predicted values ​​of the market environment. In this embodiment, the market environment calculation unit 34 calculates the predicted value of the market environment so as to have a probability distribution. In other words, the market environment calculation unit 34 calculates, as the predicted value of the market environment, information indicating a prediction range, which is a range of values ​​that the market environment can take, and the probability that the market environment will fall within that prediction range. More specifically, the market environment calculation unit 34 calculates the prediction range that the market environment can take for each probability that the market environment will fall within that prediction range. FIGS. 3 and 4 show examples of the calculated prediction ranges. FIG. 3 shows a prediction range AR1 for a certain probability, and FIG. 4 shows a prediction range AR2 for a probability different from the probability in FIG. 3. Note that the number of prediction ranges for each probability calculated by the market environment calculation unit 34 and the probability values ​​used when calculating the prediction ranges may be arbitrary. For example, the market environment calculation unit 34 may calculate a prediction range for each probability by setting the probability that the market environment will fall within the prediction range from 1% to 99% in 1% increments. The higher the probability that the market environment will fall within the forecast range, the wider the forecast range will be.

[0022] In this embodiment, the market environment calculation unit 34 calculates a predicted value of the market environment (prediction range for each probability) using a machine learning technique. Specifically, the market environment calculation unit 34 calculates a reference value of the market environment and the deviation of the market environment from the reference value using a prediction interval estimation technique, thereby calculating a prediction range for each probability. This will be explained in more detail below.

[0023] The market environment calculation unit 34 acquires a first learning model that has been machine-learned to learn the correspondence between the market environment reference value and the time period in which the reference value was obtained. The entity that machine-learns the first learning model may be any entity, such as the computing device 20 or another device. The market environment calculation unit 34 may read the trained first learning model from the storage unit 28 or may receive the trained first learning model from an external device via the communication unit 26. Any method may be used to train the first learning model. For example, the first learning model may be trained by inputting multiple sets of training data into an untrained first learning model, using the market environment reference value and the time period in which the reference value was obtained as one set of training data. As a result, the first learning model learns the correspondence between time periods and the market environment reference value, and becomes a model (program) that can calculate a predicted value of the market environment reference value for a bidding time period when the bidding time period is input. The market environment reference value may be acquired arbitrarily. For example, multiple actual values ​​of the market environment in overlapping time periods may be acquired, and the reference value of the market environment in those overlapping time periods may be calculated based on those multiple values ​​(e.g., by averaging them). Furthermore, when training the first learning model, it is preferable to use training data from different time periods. This allows for highly accurate prediction of the reference value of the market environment in various time periods.

[0024] The market environment calculation unit 34 obtains a reference value of the market environment during the bidding time period by inputting the bidding time period indicated in the bidding conditions into the machine-learned first learning model. Line L1 in Figures 3 and 4 shows an example of the reference value of the market environment during the bidding time period.

[0025] The market environment calculation unit 34 also acquires a second learning model that has been machine-learned to learn the correspondence between the deviation amount of the market environment and the time period in which the deviation amount was obtained. The entity that machine-learns the second learning model may be any entity, such as the computing device 20 or another device. The market environment calculation unit 34 may read the trained second learning model from the storage unit 28 or may receive the trained second learning model from an external device via the communication unit 26. Any method may be used to train the second learning model. For example, the second learning model may be trained by inputting multiple sets of training data into an untrained second learning model, using the deviation amount of the market environment and the time period in which the deviation amount was obtained as one set of training data. As a result, the second learning model learns the correspondence between time periods and deviation amounts, and becomes a model (program) that can calculate a predicted value of the deviation amount for a bidding time period when the bidding time period is input. The deviation amount may be acquired arbitrarily. For example, multiple actual values ​​of the market environment in overlapping time periods may be acquired, and the deviation amount for the overlapping time periods may be calculated based on the deviation of these multiple values. Furthermore, when training the second learning model, it is preferable to use training data from different time periods. This allows for highly accurate prediction of deviation amounts for various time periods.

[0026] The market environment calculation unit 34 obtains the amount of deviation in the bidding time period by inputting the bidding time period indicated in the bidding conditions into the machine-learned second learning model. Any model can be applied as the first learning model and the second learning model, and for example, a CNN (Conventional Neural Network) model may be used.

[0027] The market environment calculation unit 34 calculates a prediction range for each probability for a bidding time period based on the reference value for that bidding time period calculated as described above and the deviation for that bidding time period. For example, the market environment calculation unit 34 extracts a deviation at a probability for which the prediction range is to be calculated from the deviation calculated using the second learning model, and calculates a numerical range that is separated from the reference value by that deviation as the prediction range for that probability. That is, for example, in the example of FIG. 3, the range separated from the line L1, which is the reference value, by the deviation at the probability in FIG. 3 is calculated as the prediction range AR1. Similarly, in the example of FIG. 4, the range separated from the line L1, which is the reference value, by the deviation at the probability in FIG. 4 is calculated as the prediction range AR2. Note that any method may be used to extract a deviation at a certain probability from the deviation calculated using the second learning model. For example, the second learning model may be used to calculate a deviation at each probability, and the deviation at the probability for which the prediction range is to be calculated may be selected from the deviations at each probability calculated using the second learning model. Furthermore, for example, a value indicating the correspondence relationship between the probability and the deviation amount may be calculated as the deviation amount using the second learning model. In this case, the market environment calculation unit 34 may calculate the deviation amount at the probability for which the prediction range is to be calculated, based on the value indicating the correspondence relationship between the probability and the deviation amount and the probability for which the prediction range is to be calculated.

[0028] In this embodiment, the predicted values ​​of the market environment are calculated as predicted values ​​of the power supply amount, the predicted value of the power demand amount, and the predicted value of the power price. In the power market M, power may be supplied from a plurality of supply facilities 14, and power may be demanded from a plurality of demand facilities 12. Therefore, the market environment calculation unit 34 may calculate a prediction range (a prediction value having a probability distribution) for each probability of the total value of the power supply amount from each supply facility 14, or may calculate a prediction range for each probability of the power supply amount for each supply facility 14. Similarly, the market environment calculation unit 34 may calculate a prediction range (a prediction value having a probability distribution) for each probability of the total value of the power demand amount of each demand facility 12, or may calculate a prediction range for each probability of the power demand amount for each demand facility 12.

[0029] The market environment calculation unit 34 calculates the prediction range for each probability (the predicted value of the market environment having a probability distribution) using the method described above. However, the method for calculating the prediction range for each probability is not limited to the above description and may be any method. For example, the prediction range for each probability may be calculated using one learning model. Furthermore, the market environment calculation unit 34 is not limited to calculating the predicted value of the market environment having a probability distribution, and may calculate the predicted value of the market environment as a fixed value. In this case, for example, a set value of the market environment may be set in advance for each time period, and the market environment calculation unit 34 may calculate the set value of the market environment during the bidding time period as the predicted value of the market environment.

[0030] (Profit and Loss Calculation Department) The profit and loss calculation unit 36 ​​calculates the predicted value of profit and loss based on the bidding conditions and the predicted value of the market environment so that the predicted value of profit and loss has a probability distribution. In other words, the profit and loss calculation unit 36 ​​calculates, for each predicted value of profit and loss, a predicted value of profit and loss and information indicating the probability that the actual profit and loss will be the predicted value. Note that the predicted value of profit and loss here refers to the predicted value of profit and loss obtained when trading in the electricity market M using the bidding conditions and supplying electricity during the bidding time period. The profit and loss calculation unit 36 ​​may calculate the total of the predicted values ​​of profit and loss for each electricity market M as the predicted value of profit and loss.

[0031] As mentioned above, profit and loss refers to the revenue from selling electricity or the loss from purchasing electricity. Therefore, when the sales volume of electricity is set as a bidding condition, the profit and loss calculation unit 36 ​​calculates the predicted value of profit based on the bidding conditions and the predicted value of the market environment. On the other hand, when the purchase volume of electricity is set as a bidding condition, the profit and loss calculation unit 36 ​​calculates the predicted value of loss based on the bidding conditions and the predicted value of the market environment.

[0032] 5 and 6 are graphs illustrating an example of a predicted value of profit and loss. In this embodiment, the profit and loss calculation unit 36 ​​calculates a predicted value of profit and loss having a probability distribution by repeating the process of calculating a predicted value of profit and loss using one of the prediction ranges for each probability calculated by the market environment calculation unit 34. The method of selecting a prediction range to be used for calculation from among the prediction ranges for each probability may be arbitrary. For example, in this embodiment, the profit and loss calculation unit 36 ​​uses a predetermined random number to select one prediction range from the prediction ranges for each probability. The profit and loss calculation unit 36 ​​calculates a predicted value of profit and loss based on the selected prediction range and bidding conditions.

[0033] In this embodiment, a bid energy amount is set as a bidding condition, and a prediction range for each probability of the energy supply amount (predicted value of the energy supply amount), a prediction range for each probability of the energy demand amount (predicted value of the energy supply amount), and a prediction range for each probability of the energy price (predicted value of the energy price) are calculated as predicted values ​​of the market environment. In an example where the energy sales amount is set as the bidding condition and the profit and loss is set as revenue, the profit and loss calculation unit 36 ​​may calculate the predicted value of the profit and loss using the following formula (1) based on the selected prediction range and bidding condition.

[0034]

number

[0035] where E(t) is the predicted profit for the bidding time period t, and M pi (t) is the electricity market M i It refers to the amount of electricity sold during the bidding period t. ci (t) is the electricity market M i The predicted value of the electricity price in the bidding time period t is M ci (t) is a value within the predicted range of the electricity price extracted by a predetermined random number, and may be arbitrarily set based on the predicted range of the extracted electricity price. pK (t) indicates the amount of electricity sold in the regulated market during the bidding period t. i (t) is the electricity market M iThe predicted value of the power supply amount in the bidding time period t is i (t) is a value within the predicted range of the amount of power supply extracted by a predetermined random number, and may be arbitrarily set based on the predicted range of the amount of power supply extracted. i (t) is the electricity market M i The predicted value of the power demand during the bidding time period t is i (t) is a value within the forecast range of the electricity demand extracted using a predetermined random number, and may be arbitrarily set based on the forecast range of the extracted electricity demand. That is, the first term on the right side of equation (1) indicates the sum of the forecast values ​​of revenue in each electricity market M excluding the balancing market, and the second term on the right side of equation (1) indicates the forecast value of revenue in the balancing market.

[0036] However, the method for calculating the predicted value of the profit is not limited to using formula (1).

[0037] In the above example, the electricity sales volume was set as the bidding condition. However, even if the electricity purchase volume was set as the bidding condition, the profit and loss calculation unit 36 ​​calculates the predicted value of profit and loss (loss) so as to have a probability distribution based on the bidding conditions and the predicted value of the market environment. For example, the bidding condition setting unit 32 may set at least one of the amount of power generated by the target facility 10 and the amount of power demanded by the target facility 10 in addition to the bidding conditions (sales volume or purchase volume), and the profit and loss calculation unit 36 ​​may calculate the profit and loss. For example, when setting the bidding conditions and the amount of power supply, the profit and loss calculation unit 36 ​​calculates the predicted value of profit and loss so as to have a probability distribution based on the set bidding conditions, the amount of power supply, and the predicted value of the market environment. For example, when setting the bidding conditions, the amount of power supply, and the amount of power demand, the profit and loss calculation unit 36 ​​calculates the predicted value of profit and loss so as to have a probability distribution based on the set bidding conditions, the amount of power supply, and the predicted value of the market environment. The method for calculating the predicted profit and loss values ​​for these patterns can be arbitrarily selected, for example, by setting an appropriate calculation formula for each pattern.

[0038] The profit and loss calculation unit 36 ​​calculates a predicted profit and loss value having a probability distribution by repeating a process of selecting a prediction range and a process of calculating a predicted profit and loss value. In this embodiment, the profit and loss calculation unit 36 ​​determines whether the number of calculations of the predicted profit and loss value satisfies a predetermined condition. If the predetermined condition is not met, the profit and loss calculation unit 36 ​​repeats a process of selecting a prediction range using a predetermined random number and a process of calculating a predicted profit and loss value. On the other hand, if the number of calculations of the predicted profit and loss value satisfies the predetermined condition, the profit and loss calculation unit 36 ​​calculates a predicted profit and loss value having a probability distribution using the predicted profit and loss values ​​calculated so far. Specifically, the profit and loss calculation unit 36 ​​classifies each predicted profit and loss value into one of multiple numerical ranges that the profit or loss can take. Then, the profit and loss calculation unit 36 ​​calculates the frequency of the predicted profit and loss values ​​classified into each numerical range for each numerical range, thereby calculating a predicted profit and loss value having a probability distribution. In other words, the profit and loss calculation unit 36 ​​calculates a predicted profit and loss value having a probability distribution in which the higher the frequency, the higher the probability of the corresponding profit or loss. Fig. 5 shows an example of the frequency of predicted profit and loss values ​​classified into numerical ranges, and line L2 in Fig. 6 shows an example of predicted profit and loss values ​​having a probability distribution. Note that the predetermined condition for the number of calculations may be set as appropriate.

[0039] In this way, the profit and loss calculation unit 36 ​​calculates the predicted value of profit and loss when trading is conducted in the electricity market M using the set bidding conditions, so that the predicted value has a probability distribution. The profit and loss calculation unit 36 ​​may calculate the predicted value of profit and loss with a probability distribution every time bidding conditions are set, i.e., for each bidding condition.

[0040] (Output control section) The output control unit 38 outputs information indicating the predicted value of the profit and loss having the probability distribution calculated by the profit and loss calculation unit 36. For example, the output control unit 38 may cause the output unit 24 to output (display) the information indicating the predicted value of the profit and loss having the probability distribution, or may transmit the information to another device via the communication unit 26.

[0041] As described above, in this embodiment, a predicted value of profit and loss having a probability distribution is calculated from the set bidding conditions and the calculated predicted value of the market environment. According to this embodiment, by predicting profit and loss using a probability distribution, it is possible to predict profit and loss under the set bidding conditions while taking into account cases where the environment, such as weather conditions and the operating status of equipment, changes, and therefore it is possible to appropriately support bidding in the electricity market. Furthermore, in this embodiment, a probability distribution is applied to the market environment, which may change depending on the environment, and it is therefore possible to predict profit and loss while appropriately taking into account cases where the environment changes.

[0042] (Processing flow) Next, the calculation flow of the predicted profit and loss values ​​described above will be explained based on a flowchart. FIG. 7 is a flowchart illustrating the calculation flow of the predicted profit and loss values. As shown in FIG. 7, the calculation device 20 sets bidding conditions using the bidding condition setting unit 32 (step S10), and calculates a prediction range of the market environment for each probability using the market environment calculation unit 34 (step S12). Then, the calculation device 20 selects a prediction range to be used for calculating the predicted profit and loss values ​​for each probability using the profit and loss calculation unit 36 ​​(step S14), and calculates the predicted profit and loss values ​​based on the selected prediction range and the bidding conditions (step S16). Then, if the number of calculations of the predicted profit and loss values ​​does not satisfy the predetermined condition (step S18; No), the process returns to step S14 to repeat the selection of the prediction range and the calculation of the predicted profit and loss values. If the number of calculations of the predicted profit and loss values ​​satisfies the predetermined condition (step S18; Yes), the profit and loss calculation unit 36 ​​calculates the predicted profit and loss values ​​having a probability distribution (step S20).

[0043] (Second embodiment) Next, a second embodiment will be described. The calculation device 20 according to the second embodiment differs from the first embodiment in that it performs optimization calculations for bidding conditions based on predicted values ​​of profit and loss having a probability distribution and an evaluation function. In the second embodiment, explanations of parts of the configuration common to the first embodiment will be omitted.

[0044] 8 is a schematic block diagram of a calculation device according to the second embodiment. As shown in FIG. 8, a control unit 30 of a calculation device 20 according to the second embodiment includes an evaluation function setting unit 37.

[0045] The profit and loss calculation unit 36 ​​in the second embodiment calculates the predicted profit and loss value under the set bidding conditions in the same manner as in the first embodiment. Then, the evaluation function setting unit 37 in the second embodiment sets an evaluation function for evaluating the predicted profit and loss value. The evaluation function may be set appropriately depending on the purpose of transaction support. For example, multiple evaluation functions may be set for each purpose, and the evaluation function setting unit 37 may select an evaluation function to be used this time from among the multiple evaluation functions. For example, when a user inputs an instruction to select an evaluation function, the evaluation function setting unit 37 may select the evaluation function selected by the user as the evaluation function to be used this time. Note that, as the evaluation function for each purpose, for example, one that maximizes the predicted profit value or one that minimizes the probability that the predicted profit value will be low may be mentioned.

[0046] The bidding condition setting unit 32 in the second embodiment determines whether the bidding conditions used to calculate the predicted profit and loss values ​​are appropriate, based on the predicted profit and loss values ​​calculated by the profit and loss calculation unit 36 ​​and the evaluation function set by the evaluation function setting unit 37. Specifically, the bidding condition setting unit 32 performs a calculation to determine whether the predicted profit and loss values, which have a probability distribution, have converged, based on the evaluation function. If the predicted profit and loss values ​​have converged, the bidding condition setting unit 32 determines that the bidding conditions are appropriate (the bidding conditions have been optimized). On the other hand, if the predicted profit and loss values ​​have not converged, the bidding condition setting unit 32 determines that the bidding conditions are inappropriate (the bidding conditions have not been optimized), and resets the bidding conditions. Then, the calculation device 20 calculates the predicted profit and loss values ​​for the reset bidding conditions using the same method, and determines whether the predicted profit and loss values ​​have converged.

[0047] The output control unit 38 according to the second embodiment outputs the bidding conditions that are determined to be optimized. For example, the output control unit 38 may output (display) the bidding conditions that are determined to be optimized on the output unit 24, or may transmit them to another device via the communication unit 26.

[0048] The flow of the optimization process of bidding conditions described above will now be described. FIG. 9 is a flowchart illustrating the flow of the optimization process of bidding conditions. As shown in FIG. 9, the calculation device 20 sets bidding conditions using the bidding condition setting unit 32 (step S30), and calculates the predicted value of profit and loss for the bidding conditions so that it has a probability distribution using the same method as in the first embodiment (step S32). The calculation device 20 sets an evaluation function using the evaluation function setting unit 37 (step S34), and determines whether the predicted value of profit and loss has converged based on the evaluation function and the predicted value of profit and loss (step S36). Note that steps S34 and S36 may be performed in any order. If the predicted value of profit and loss has not converged (step S36; No), the process returns to step S30, the bidding conditions are reset, and the subsequent processes are repeated. On the other hand, if the predicted value of profit and loss has converged (step S36; Yes), the calculation device 20 sets the bidding conditions used to calculate the predicted value of profit and loss as the optimized bidding conditions (step S38).

[0049] As described above, in the second embodiment, the evaluation function is used to perform optimization calculations to determine whether the predicted profit and loss values ​​have converged, thereby optimizing the bidding conditions. This makes it possible to set optimal bidding conditions according to the purpose, thereby more appropriately supporting bidding.

[0050] (effect) A calculation method according to a first aspect of the present disclosure includes the steps of setting bidding conditions for the electricity market M, calculating a predicted value of the market environment, which is at least one of the amount of electricity supplied to the electricity market M, the amount of electricity demand in the electricity market M, and the electricity price in the electricity market M, and calculating, based on the bidding conditions and the predicted value of the market environment, a predicted value of profit and loss when trading in the electricity market M using the bidding conditions, so as to have a probability distribution. According to the present disclosure, predicting profit and loss with a probability distribution makes it possible to predict profit and loss under the set bidding conditions, taking into account cases where the environment fluctuates, and therefore makes it possible to appropriately support bidding in the electricity market.

[0051] A calculation method according to a second aspect of the present disclosure is the calculation method according to the first aspect, wherein in the step of calculating a predicted value of the market environment, the predicted value of the market environment is calculated so as to have a probability distribution. According to the present disclosure, since a probability distribution is given to a market environment that may fluctuate depending on the environment, it becomes possible to predict profits and losses by appropriately taking into account cases in which the environment fluctuates.

[0052] A calculation method according to a third aspect of the present disclosure is the calculation method according to the second aspect, wherein in the step of calculating a forecast value of the market environment, a forecast range, which is a range that the market environment can take, is calculated for each probability that the market environment will be within the forecast range, and in the step of calculating a forecast value of profit and loss, a process of calculating a forecast value of profit and loss using one of the forecast ranges for each probability is repeated to calculate a forecast value of profit and loss so that the forecast value of profit and loss has a probability distribution. According to the present disclosure, a forecast range for each probability is set for a market environment that may fluctuate depending on the environment, and profit and loss is predicted using one of the ranges, so that profit and loss under the set bidding conditions can be appropriately predicted taking fluctuations into account.

[0053] A calculation method according to a fourth aspect of the present disclosure is the calculation method according to the third aspect, wherein the step of calculating a predicted value of the market environment includes calculating the market environment reference value by inputting the bidding time period indicated in the bidding conditions into a first learning model that has been machine-learned to learn the correspondence between a market environment reference value and a time period, calculating the deviation amount by inputting the bidding time period indicated in the bidding conditions into a first learning model that has been machine-learned to learn the correspondence between a deviation amount from the market environment reference value and a time period, and calculating a prediction range for each probability based on the market environment reference value and the deviation amount. According to the present disclosure, the prediction range for each probability is set from the reference value and the deviation amount calculated by machine learning, so that profit and loss under the set bidding conditions can be appropriately predicted taking fluctuations into account.

[0054] A calculation method according to a fifth aspect of the present disclosure is the calculation method according to the third or fourth aspect, wherein in the step of calculating a predicted value of profit and loss, each predicted value of profit and loss calculated for each prediction range is classified into one of a plurality of numerical ranges that the profit and loss can take, and the frequency of the predicted values ​​of profit and loss classified into the numerical range is calculated for each numerical range, thereby calculating the predicted value of profit and loss so that it has a probability distribution. According to the present disclosure, a predicted value of profit and loss having a probability distribution is calculated from the frequency of the numerical range to which the predicted value of profit and loss belongs, so that it is possible to appropriately predict profit and loss under the set bidding conditions, taking fluctuations into account.

[0055] A calculation method according to a sixth aspect of the present disclosure is the calculation method according to any one of the first to fifth aspects, further including a step of setting an evaluation function for profit and loss, and in the step of setting bidding conditions, determining whether the predicted value of profit and loss has converged based on the evaluation function, and if the predicted value of profit and loss has not converged, resetting the bidding conditions. According to the present disclosure, the evaluation function is used to perform an optimization calculation to determine whether the predicted value of profit and loss has converged, thereby optimizing the bidding conditions. This makes it possible to set optimal bidding conditions according to the purpose, thereby more effectively supporting bidding.

[0056] A calculation method according to a seventh aspect of the present disclosure is the calculation method according to any one of the first to sixth aspects, wherein the bidding conditions are the trading volume of electricity in the electricity market M, and the market environment is the amount of electricity supplied to the electricity market M from supply facilities 14 other than the target facility 10, the amount of electricity demanded in the electricity market, and the electricity price in the electricity market. According to the present disclosure, these parameters are used to predict profit and loss for the set bid power amount, thereby enabling appropriate support for bidding.

[0057] A program according to the present disclosure causes a computer to execute the steps of setting bidding conditions for the electricity market M, calculating a predicted value of the market environment, which is at least one of the amount of electricity supplied to the electricity market M, the amount of electricity demand in the electricity market M, and the electricity price in the electricity market M, and calculating, based on the bidding conditions and the predicted value of the market environment, a predicted value of profit / loss when trading in the electricity market M using the bidding conditions, so as to have a probability distribution. According to the present disclosure, it is possible to appropriately support bidding in the electricity market.

[0058] The computing device 20 according to the present disclosure includes a bidding condition setting unit 32 that sets bidding conditions for the electricity market M, a market environment calculation unit 34 that calculates a predicted value of the market environment, which is at least one of the amount of electricity supplied to the electricity market M, the amount of electricity demand in the electricity market M, and the electricity price in the electricity market M, and a profit and loss calculation unit 36 ​​that calculates a predicted value of profit and loss when trading in the electricity market M using the bidding conditions, based on the bidding conditions and the predicted value of the market environment, so as to have a probability distribution. According to the present disclosure, it is possible to appropriately support bidding in the electricity market.

[0059] Although the embodiments of the present disclosure have been described above, the embodiments are not limited to the contents of these embodiments. Furthermore, the above-described components include those that can be easily imagined by a person skilled in the art, those that are substantially the same, and those that are within the so-called equivalent range. Furthermore, the above-described components can be combined as appropriate. Furthermore, various omissions, substitutions, or modifications of the components can be made without departing from the spirit of the above-described embodiments. [Explanation of symbols]

[0060] 20 Arithmetic unit 32 Bidding Condition Setting Department 34 Market Environment Calculation Department 36 Profit and Loss Calculation Department M electricity market

Claims

1. A computing method executed by a computing device, comprising: setting bidding conditions indicating an amount of electricity to be traded through the electricity market and a bidding time period during which the electricity will be traded; calculating a predicted value of a market environment, which is at least one of an amount of electricity supplied to the electricity market, an amount of electricity demanded in the electricity market, and an electricity price in the electricity market; calculating a predicted value of profit or loss when trading in the electricity market using the bidding conditions based on the predicted value of the bidding conditions and the market environment, so as to have a probability distribution; Including, In the step of calculating the predicted value of the market environment, a machine learning technique based on the bidding time period is used to calculate, as the predicted value of the market environment, a predicted range, which is a range of values ​​that the market environment can take during the bidding time period, for each probability that the market environment will fall within the predicted range during the bidding time period; In the step of calculating the predicted value of profit and loss, a process of calculating the predicted value of profit and loss using one of the prediction ranges for each probability is repeated, each predicted value of profit and loss is classified into one of a plurality of numerical ranges into which profit and loss can fall, and the frequency of the predicted value of profit and loss classified into each numerical range is calculated for each numerical range, thereby calculating the predicted value of profit and loss so as to have a probability distribution. Calculation method.

2. In the step of calculating the predicted value of the market environment, calculating the reference value of the market environment by inputting the bidding time period indicated in the bidding conditions into a first learning model that has completed machine learning of the correspondence relationship between the reference value of the market environment and the time period; calculating the deviation amount by inputting the bidding time period indicated in the bidding conditions into a second learning model that has undergone machine learning of the correspondence between the deviation amount from the reference value of the market environment and the time period; The calculation method according to claim 1 , further comprising calculating a prediction range for each probability based on the reference value of the market environment and the deviation amount.

3. further comprising a step of setting an evaluation function for the profit and loss; 3. The calculation method according to claim 1, wherein in the step of setting the bidding conditions, it is determined whether the predicted value of profit and loss has converged based on the evaluation function, and if the predicted value of profit and loss has not converged, the bidding conditions are reset.

4. A calculation method as described in claim 1 or claim 2, wherein the market environment is the amount of electricity supplied to the electricity market from equipment other than the target equipment, which is the equipment for which the bidding conditions are set, the amount of electricity demand in the electricity market, and the electricity price in the electricity market.

5. setting bidding conditions indicating an amount of electricity to be traded through the electricity market and a bidding time period during which the electricity will be traded; calculating a predicted value of a market environment, which is at least one of an amount of electricity supplied to the electricity market, an amount of electricity demanded in the electricity market, and an electricity price in the electricity market; calculating a predicted value of profit or loss when trading in the electricity market using the bidding conditions based on the predicted value of the bidding conditions and the market environment, so as to have a probability distribution; on the computer, In the step of calculating the predicted value of the market environment, a machine learning technique based on the bidding time period is used to calculate, as the predicted value of the market environment, a predicted range, which is a range of values ​​that the market environment can take during the bidding time period, for each probability that the market environment will fall within the predicted range during the bidding time period; In the step of calculating the predicted value of profit and loss, a process of calculating the predicted value of profit and loss using one of the prediction ranges for each probability is repeated, each predicted value of profit and loss is classified into one of a plurality of numerical ranges into which profit and loss can fall, and the frequency of the predicted value of profit and loss classified into each numerical range is calculated for each numerical range, thereby calculating the predicted value of profit and loss so as to have a probability distribution. program.

6. a bidding condition setting unit that sets bidding conditions indicating the amount of electricity to be traded via the electricity market and a bidding time period for trading the electricity; a market environment calculation unit that calculates a predicted value of a market environment, which is at least one of an amount of electricity supplied to the electricity market, an amount of electricity demanded in the electricity market, and an electricity price in the electricity market; a profit and loss calculation unit that calculates a predicted value of profit and loss when trading in the electricity market using the bidding conditions based on the predicted value of the bidding conditions and the market environment, so as to have a probability distribution; Including, the market environment calculation unit calculates, as a predicted value of the market environment, a prediction range, which is a range of values ​​that the market environment can take during the bidding time period, for each probability that the market environment will fall within the prediction range during the bidding time period, using a machine learning technique based on the bidding time period; the profit and loss calculation unit repeats a process of calculating the predicted value of the profit and loss using one of the prediction ranges for each probability, classifies each of the predicted values ​​of the profit and loss into one of a plurality of numerical ranges that the profit and loss can take, and calculates the frequency of the predicted values ​​of the profit and loss classified into each of the numerical ranges for each of the numerical ranges, thereby calculating the predicted value of the profit and loss so as to have a probability distribution; Computing device.

Images