Information processing device, information processing method, and computer program

Abstract

To provide an information processing device, information processing method, and information processing program, which provide assistance in bidding for products on a trading market.SOLUTION: An information processing device disclosed herein comprises a processing unit configured to use a price fluctuation model representing market price fluctuations according to time elapsed from a target time and a bid volume in a tradable period of a product trading market in order to determine one or more bidding times for bidding during a period after the target time in the tradable period and a bid volume to be bid in the trading market.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] This embodiment relates to an information processing device, an information processing method, and a computer program. [Background technology]

[0002] JEPX (Japan Electric Power Exchange) operates the spot market (one-day-ahead market) as its primary trading market. The spot market handles trading of electricity to be delivered the following day. JEPX also operates an hour-ahead market as an adjustment market to address any surplus or shortage of electricity after trading in the spot market.

[0003] The difference between the amount of electricity traded in over-the-counter transactions and the actual amount of electricity generated is considered an imbalance, and the power generator is penalized for it. Generally, the difference between the planned amount of electricity generated and the amount of electricity traded in over-the-counter transactions is put up for bidding in the spot market. Since imbalances caused by changes in electricity generated after the spot market closes directly affect the power generator's profits, it is desirable for power generators to use the pre-market to avoid imbalances and secure profits at the same time.

[0004] However, due to the low liquidity of the current pre-market, the buying and selling activities of power generators themselves significantly impact market prices. Furthermore, the closer it gets to the end of the trading period, the higher the risk of unsuccessful bids. In addition, the longer the period from the current time to the bidding time, the higher the price fluctuation risk tends to be. If power generators trade electricity to resolve imbalances in the pre-market without considering the impact of their own buying and selling activities, the risk of unsuccessful bids, and the price fluctuation risk, there is a concern that their profits will decrease. [Prior art documents] [Patent Documents]

[0005] [Patent Document 1] Japanese Patent Publication No. 2019-215693 [Overview of the project] [Problems that the invention aims to solve]

[0006] This embodiment provides an information processing device, an information processing method, and a computer program that support bidding on goods in a trading market. [Means for solving the problem]

[0007] The information processing device disclosed herein includes a processing unit that determines one or more bidding timings to be used for bidding during the period after the target time within the trading period for a product in the trading market, and the amount of bids to be placed in the trading market, based on a price fluctuation model that represents the fluctuation of the market price in accordance with the elapsed time from a target time and the amount of bids placed during the trading period for a product in the trading market. [Brief explanation of the drawing]

[0008] [Figure 1] A block diagram of an example of a trading strategy calculation device, which is an information processing device according to this embodiment. [Figure 2] Here are 48 examples of products. [Figure 3] Diagram explaining the trading period. [Figure 4A] A diagram illustrating an example of a predicted imbalance. [Figure 4B] A diagram illustrating another example of predicted imbalance. [Figure 5] A diagram showing an example of a tree structure representing all bidding patterns for predicted imbalances. [Figure 6] A diagram illustrating a specific example of the processing performed by the optimal trading strategy calculation unit. [Figure 7] This figure shows an example of recommended result data based on the bid content recommended by the optimal trading strategy calculation unit. [Figure 8] A flowchart illustrating an example of the processing of the trading strategy calculation device according to this embodiment. [Figure 9] An explanatory diagram for branch pruning exploration. [Figure 10] A diagram showing an example of the hardware configuration of the device in Figure 1 or Figure 13. [Modes for carrying out the invention]

[0009] Hereinafter, this embodiment will be described while referring to the drawings.

[0010] FIG. 1 is a block diagram of an example of a trading strategy calculation device 1 which is an information processing device according to this embodiment. The trading strategy calculation device 1 includes a market data acquisition unit 11, a price impact table generation unit 12, a price impact table input unit 13, a power generation company database 14, a power generation amount prediction unit 15, a predicted imbalance calculation unit 16, an input data acquisition unit 17, an optimal trading strategy calculation unit 18, an output data acquisition unit 19, an output unit 20, and an imbalance charge predicted value acquisition unit 10. At least a part of each of the units 10 to 13 and 15 to 20 corresponds to a processing unit that performs the processing according to this embodiment.

[0011] The trading strategy calculation device 1 is connected to a pre-time market trading device 21, a power generation device 31, a weather server 41, and a user device 51 via a communication network. The communication network is, for example, the Internet, and includes a wired network, a wireless network, or a hybrid network thereof.

[0012] The power generation device 31 includes, for example, a renewable energy power generation device such as a solar power generation device or a wind power generation device. The power generation device 31 is used by a power generation company. The power generation company supplies (sells, etc.) the electric power generated by the power generation device 31 to consumers.

[0013] The weather server 41 is a server that stores and manages past weather information and performs future weather prediction by numerical weather calculation.

[0014] The pre-time market trading device 21 is a device that performs trading processing in the pre-time market installed in JEPX (Japan Electric Power Exchange) as an example of the electric power trading market. In this embodiment, it is assumed that electric power is traded as a commodity, but the commodity is not limited to electric power as long as it is something traded in the market such as gas, hydrogen, coffee beans, corn, etc.

[0015] The user device 51 is used by a user to input various instructions or data to the trading strategy calculation device 1. The user device 51 may also transmit bidding instructions to the pre-market trading device 21. The user belongs to, for example, a power generation company or a company commissioned by such a company.

[0016] (Market data acquisition unit 11) The market data acquisition unit 11 is connected to the advance market trading device 21 via a communication network and acquires data related to the advance market (electricity trading market) from the advance market trading device 21.

[0017] Figures 2 and 3 will be used to explain the time-ahead market. A day is divided into 48 time slots (commodity slots) based on the unit of electricity measurement (30 minutes), and electricity (commodity) is traded separately for each time slot. Each time slot is independent of the others as a commodity. A one-hour-ahead market exists for each time slot.

[0018] Figure 2 shows examples of 48 products. For example, the top product is electricity, which is traded during the time frame of "00:00-00:30" on a given day. For each time frame, trading (bids and agreements) is repeated in real time during the bidding period (trading period). This trading method is known as the continuous trading method.

[0019] Figure 3 is an explanatory diagram of the trading period. The trading period for each commodity is from the market opening time until the market closing time one hour before the commodity is to be delivered. For example, the trading period for a commodity between 2:00 and 2:30 on a given day is from 5:00 PM the previous day to 1:00 AM on the current day, and the trading period for a commodity between 4:00 and 4:30 on a given day is from 5:00 PM the previous day to 3:00 AM on the current day.

[0020] In the example in Figure 3, an example of the trading period for a product from "04:00 to 04:30" on a given day is shown. The start time (opening time) of the trading period is 17:00. The start time may be the same for all products. The end time (closing time) is 03:00, which is one hour before the product delivery time of 04:00. In the example in Figure 3, the current time is 19:30, and a user can place a bid at one or more arbitrary times during the remaining time until closing. The electricity trading market according to this embodiment is not limited to a pre-trading market; it may be any other market in which trading (bids and agreements) is repeated in real time during the trading period.

[0021] The pre-hour market trading device 21 processes transactions between market participants in the pre-hour market according to the continuous trading method, and also stores and manages trading rules and market trading performance data. The pre-hour market data acquired by the market data acquisition unit 11 includes trading rules in the pre-hour market and market trading performance data between market participants. Trading rules include information on the trading period for each commodity unit, the minimum bid unit [kWh] (in JEPX, the minimum bid unit is 50 [kWh]), etc. Examples of market trading performance data include the contract price, contract quantity, and contract time of a commodity (electricity). Market trading performance data may also include statistical quantities such as the average value of past trading results.

[0022] In this embodiment, when a power generator sells generated electricity in the pre-market, the system determines and recommends to the user the timing (bidding time frame) and content (bid quantity and bid price) of the bid to be placed during the trading period (bidding period) in order to optimize the power generator's profits. However, power generators can purchase electricity in the pre-market as well as sell it. Furthermore, the entity whose profits are to be optimized in this embodiment is not limited to power generators, but may also be an electricity consumer (e.g., an aggregator). In this case, the system determines the timing (bidding time frame) and content (bid quantity and bid price) of the bid to be placed by the electricity consumer in order to optimize the electricity consumer's profits, and recommends these decisions to the user. Note that the recommendation of a bid price to the user may be omitted. In this case, the transaction may be conducted at the market price (market price) that is actually tradable in real time at the time of bidding in the pre-market, in the form of a so-called market order (real-time order).

[0023] (Power Generation Company Database 14) The power generation company database 14 stores power generation plan data, including the planned power generation values ​​for each future day of the power generation company. The planned power generation values ​​are stored for each time period into which the day is divided. For example, the time unit for the planned power generation values ​​may be the same as the time unit for the pre-market commodity slots (30 minutes in this example), or it may be a different time unit from the commodity slots.

[0024] The planned power generation value indicates the amount of electricity that a power generator has promised to supply to other businesses (aggregators, etc.). The planned power generation value may also be the amount of electricity per commodity unit agreed upon in the spot market (day-ahead market), which opens before the hourly market. A brief explanation of the spot market follows.

[0025] In the spot market, similar to the day-ahead market, the day is divided into 48 time slots (commodity slots) based on electricity measurement units (30 minutes), and electricity is traded for each time slot. In the spot market, bidding for the next day's time slots closes every morning. For each time slot, the contract price (transaction price) and contract quantity are determined using a blind single-price auction system. Regarding the electricity for the next day, the day-ahead market opens after the spot market closes, and any surplus or shortage of electricity for the next day can be traded in the day-ahead market.

[0026] The power generation plan value is not limited to the amount of electricity (generation amount) that the power generator has contracted for in the spot market, but may also be the amount of electricity that the power generator has determined through bidding with other companies. The power generation plan may be updated, for example, for each bidding time slot (every 30 minutes in this example) during the trading period of the commodity unit.

[0027] (Power generation prediction unit 15) The power generation forecasting unit 15 forecasts the amount of power generated each day in the future and calculates power generation forecast data including the forecast values ​​of power generation. The forecast values ​​of power generation may be calculated for each time period into which the day is divided. In this case, the time unit for the forecast may be the same as the time unit for the pre-market commodity slots, or it may be a different time unit from that of the commodity slots. The power generation forecast for a commodity slot may be performed, for example, for each bidding time frame (every 30 minutes in this example) during the trading period of the commodity slot, and the forecast values ​​may be updated.

[0028] There are no particular limitations on the method for predicting power generation. As an example of a simple prediction method, one could identify the amount of power generated for the same time period in the past for multiple days from the power generation history data of the power generation equipment, and then average the identified amount of power generated. The days to be averaged may be limited to the same day of the week in the past. Alternatively, the average may be calculated over several consecutive days in the past (for example, the past week).

[0029] Alternatively, the amount of power generated may be predicted using weather information in addition to power generation history data. For example, a model is generated that predicts the amount of power generated for each time period on a given day using past weather information, and the amount of power generated for each time period on a given day is predicted from the generated model and the weather forecast values ​​for each time period on that day. During the trading period of the product slot, weather forecasts may be performed for each bidding time slot (every 30 minutes in this example), and the predicted amount of power generated may be updated. Past weather information and weather forecast values ​​are obtained from the weather server 41. The power generation prediction unit 15 may also be equipped with the weather forecasting function of the weather server 41.

[0030] (Predicted Imbalance Calculation Unit 16) The predicted imbalance calculation unit 16 obtains power generation plan data from the power generation business operator database 14 and power generation forecast data from the power generation amount forecast unit 15. The predicted imbalance calculation unit 16 calculates the predicted imbalance for each product time slot (time slot) by subtracting the power generation amount forecast value from the power generation amount plan value.

[0031] Figures 4A and 4B are explanatory diagrams of predicted imbalances. Figure 4A shows an example of a predicted imbalance when the planned generation amount is greater than the predicted generation amount. Figure 4B shows an example of a predicted imbalance when the predicted generation amount is greater than the planned generation amount. Thus, the predicted imbalance is the difference between the planned generation amount and the predicted generation amount. The predicted imbalance may be updated by calculating it during the trading period, for example, for each bidding time slot (every 30 minutes in this example). To resolve the predicted imbalance, in the case of Figure 4A, it is necessary to purchase electricity equal to the predicted imbalance in the pre-market. In the case of Figure 4B, it is necessary to sell electricity equal to the predicted imbalance in the pre-market.

[0032] (Price Impact Table Generation Unit 12) The price impact table generation unit 12 generates a table that shows the fluctuation of the market price in relation to the remaining time (i.e., elapsed time from the target time) from the target time (calculation start time) to the closing time in the trading period of the pre-market, and the amount of bids made. This table is called the price impact table. The price impact table shows the impact of remaining time and bid amount on the market price. The price impact table is an example of a price fluctuation model that models the fluctuation of the market price in relation to remaining time and bid amount. The price fluctuation model is not limited to a table format and may be represented by functions or other means.

[0033] The price fluctuation model simultaneously represents how a bidder's own bid (sell bid or buy bid) can lead to a decrease in the selling price or an increase in the buying price, and how the risk of price fluctuations increases as closing time approaches (as time decreases). In the following explanation, we assume that the bidder is a power generation operator and that the bid is a sell bid.

[0034] When the trading period is divided into multiple bidding time slots of a fixed duration, in sell bids, the more bidding time slots there are from the start of calculation to the closing time, and the larger the volume of bids in each time slot, the greater the risk of price decline. Conversely, in buy bids, the more time slots there are from the start of calculation to each bidding time slot, and the larger the volume of bids in each time slot, the greater the risk of price increase. The price impact table stores this price behavior as multiple coefficients (impact coefficients) corresponding to the multiple bid volumes for each of the remaining bidding time slots.

[0035] In this embodiment, for the sake of simplicity, the calculation start time is assumed to be the start time of each bidding time slot. However, the calculation start time can also be at other times, for example, in the middle of each bidding time slot, or before the start of the trading period. Actual bids can be made at any time within each bidding time slot. An example of a price impact table is shown below as Table 1. Details of Table 1 will be described later. [Table 1]

[0036] The following is an example of generating a price impact table for the target product. A price impact table is generated for each product. The generated price impact table may be updated or regenerated periodically (e.g., daily) or irregularly.

[0037] (Step 1) The price impact table generation unit 12 acquires market data via the market data acquisition unit 11. It divides the trading period of the target product into multiple bidding time slots (1st to Nth bidding time slots). The start time of each of the multiple bidding time slots can become the calculation start time (target time).

[0038] The price impact table generation unit 12 defines the maximum number of bidding time slots existing from the calculation start time to the closing time as the maximum number of time slots (Tmax). In other words, the number of bidding time slots from the start time of the very first bidding time slot to the closing time corresponds to the maximum number of time slots.

[0039] Furthermore, the price impact table generation unit 12 determines the minimum bid unit (let's call it u[kWh]) and the upper limit of the bid amount (let's call it Qmax[kWh]). In this example, we assume that the minimum bid unit is 50[kWh], similar to JEPX. The upper limit of the bid amount means, for example, the upper limit of the total amount of bids that can be placed within the trading period of a product slot. The upper limit of the bid amount may be the maximum amount of power that the power generation equipment can generate during the time of the product slot (rated power generation), or it may be the maximum amount of bids that power generation companies have actually placed for the same product slot on past days. Alternatively, the user may arbitrarily specify the maximum bid amount via the user device 51.

[0040] The maximum number of bids (let's call it N) is calculated by dividing the upper limit of the bid amount Qmax by the minimum bid unit u. If the result of dividing Qmax by u is a decimal, the decimal part is truncated.

[0041] The price impact table generation unit 12 generates a table with (N+1) rows and a Tmax column size, called Price Impact Table I imThis generates a table that will serve as the basis for generating the following. The value to enter in the cell at row i, column j. The value JPEG0007864625000002.jpg6161 is calculated using the method described below. This value represents the coefficient (price impact coefficient) when bidding on (i-1)u [kWh] of electricity, where j is the number of bidding time slots from the start of the calculation to the bidding timing (bidding time slot).

[0042] (Step 2) The price impact table generation unit 12 calculates the maximum price fluctuation rate [%] for the target product using past daily market transaction data as parameter α. For example, it identifies the maximum and minimum price values ​​from past market transaction data and calculates α by (maximum value - minimum value) / maximum value or its average. That is, α is calculated by subtracting the minimum value from the maximum value and dividing the result by the maximum value, or by taking the average of the values ​​obtained by the division.

[0043] Furthermore, the price impact table generation unit 12 uses the calculated α to: Calculate TIFF0007864625000003.tif9161. That is, calculate β by dividing α by the sum of N and the square of Tmax.

[0044] (Step 3) JPEG0007864625000004.jpg66170

[0045] (Step 4) JPEG0007864625000005.jpg36170

[0046] In this way, the price impact table shown in Table 1 above is generated. The price impact table contains multiple coefficients corresponding to multiple bid quantities for each of the multiple (1st to Nth) bidding time slots after the calculation start time.

[0047] (Step 5) The price impact table generation unit 12 updates or regenerates the price impact table at fixed or arbitrary time intervals following the procedures of steps 2 to 4. For example, the price impact table may be updated by acquiring market trading performance data after the pre-market has closed. Alternatively, the price impact table may be updated by receiving market trading performance data in real time during the trading period (for example, by receiving market trading performance data at the end of each bidding time frame), updating parameters α and β, and updating the price impact table.

[0048] The price impact table input unit 13 retrieves the price impact table generated by the price impact table generation unit 12 and sends the retrieved price impact table to the input data acquisition unit 17. The price impact table input unit 13 may also send the price impact table when it receives a request for retrieval from the input data acquisition unit 17. The price impact table input unit 13 sends the latest price impact table to the input data acquisition unit 17 before any calculations related to the target product are performed.

[0049] (Imbalance charge prediction value acquisition unit 10) The imbalance fee forecast acquisition unit 10 acquires imbalance fee forecasts (surplus imbalance forecast (lower limit price) in the case of a sale, and deficit imbalance fee forecast (upper limit price) in the case of a purchase). The acquisition source may be an external server or a storage unit such as internal memory. The imbalance fee forecast functions as the upper or lower limit price of the bid price. The imbalance fee forecast may differ for each product item or may be common to all product items.

[0050] (Input data acquisition unit 17) The input data acquisition unit 17 acquires information on the predicted imbalance of the target product slot (the difference between the power generation plan value and the power generation forecast value for the target product slot) from the predicted imbalance calculation unit 16. If at least one of the power generation plan value and the power generation forecast value is updated at regular or arbitrary time intervals, the input data acquisition unit 17 may acquire the updated predicted imbalance at each time interval of the bidding time frame during the trading period of the product slot.

[0051] The input data acquisition unit 17 acquires the latest market transaction performance data from the market data acquisition unit 11 at fixed or arbitrary time intervals (for example, every hour in a bidding time frame), and identifies the latest execution price, etc., from the acquired market transaction performance data. The identified latest execution price is treated as the latest execution price immediately preceding the calculation start time in the optimal trading strategy calculation unit 18. If the calculation start time is the start time of the first bidding time frame, the last execution price during the trading period of the same product on the previous day may be used. Alternatively, it is also possible to use the price at which the first transaction occurred after bidding actually started.

[0052] The input data acquisition unit 17 calculates the remaining time from the calculation start time (in this example, the start time of each bidding time slot) until the market closing time, each time a bidding time slot arrives. For example, it obtains the time from a clock and calculates the length of time from the obtained time until the market closing time to obtain the remaining time.

[0053] The input data acquisition unit 17 acquires imbalance charge forecast values ​​from the imbalance charge forecast value acquisition unit 10.

[0054] The input data acquisition unit 17 also acquires the price impact table for the target product from the price impact table input unit 13.

[0055] (Optimal trading strategy calculation unit 18) The optimal trading strategy calculation unit 18 uses a price impact table for the target product to calculate the bid quantity and bid price for each bidding time frame (30-minute time frames in this example) from the calculation start time to the market closing time through optimization calculations. Then, the bid quantity and bid price for the bidding time frame that includes the calculation start time (the target time frame) are presented to the user as the recommended bid quantity and bid price for that target time frame. The optimization calculation is performed based on an optimization criterion that optimizes the total transaction amount of the power generator (sum of contract price * contract quantity; equivalent to revenue in the case of selling) with the constraint that the sum of bid quantities in all bidding time frames should match or approach the predicted imbalance. This determines the bid quantity and bid price for each bidding time frame after the target time frame. The optimization criterion corresponds to maximizing the total transaction amount (revenue) in the case of selling bids, and to minimizing the total transaction amount in the case of buying bids. Once the time frame specified above has elapsed during trading hours, the target time frame is changed to the next bidding time frame (by changing the calculation start time to 30 minutes later), and the same process is repeated. In this way, the optimization calculation is repeated for each bidding time frame. Once the processing for the last bidding time frame is complete, the process ends.

[0056] Specifically, the optimal trading strategy calculation unit 18 generates the following mathematical programming model as an optimization problem. This optimization problem maximizes (or near-maximizes) the objective function in equation (1) under the constraints shown in equations (2) and (3), u t (t=1...T) and p t Calculate (t=1...T). TIFF0007864625000006.tif10161TIFF0007864625000007.tif13161TIFF0007864625000008.tif6161

[0057] The following symbols are defined. JPEG0007864625000009.jpg74170

[0058] The amount of electricity bid q in each bidding time frame t (t=1...T) t[kWh] is obtained by multiplying the minimum bidding unit u by the number of bids u t (q t =u t ·u). The bid power quantity q t is discretized by the minimum bidding unit u. Therefore, the possible number of bids (denoted as ht) within the bidding time frame t is within the range shown in (4). TIFF0007864625000010.tif14161

[0059] The term Σ included in the objective function of Equation (1) represents the revenue (total contract amount or total transaction amount) obtained when bidding only the number of bids u t at the market predicted price p t for each bidding time frame t (t = 1…T) after the start time of calculation, assuming that all bids are accepted. The start time of the bidding time frame of t = 1 corresponds to the start time of calculation. However, any time within the time of the bidding time frame of t = 1 may be used as the start time of calculation, and it is not excluded to use a time before the bidding time frame of t = 1 as the start time of calculation. The market predicted price is the price at which it is expected that a transaction will be concluded (a bid will be accepted) in the market, and it may also be the predicted value of the average price (for example, the weighted average price of trading volume) in the corresponding bidding time frame.

[0060] The constraint condition of (2) corresponds to the constraint that the total amount of bids is equal to the predicted imbalance (the difference between the predicted power generation amount of the power generation equipment of the power generation company and the planned power supply amount of the power generation company), or the difference between the total bid price and the predicted imbalance is less than or equal to the minimum bidding unit of the bid amount.

[0061] By maximizing the objective function under the constraint conditions, the number of bids (recommended bid number) u t and the bid price (recommended bid price) p t can be calculated as the solution of the optimization problem. In the case of a buy bid, it is only necessary to minimize the Σ formula included in the objective function. The target time frame is advanced one by one every 30 minutes, which is the unit time of the bidding time frame, and the above optimization problem is calculated. Therefore, each time the calculation is performed, the value of T in Equation (1) and Equation (2) will decrease by one.

[0062] In other words, the calculation start time (target time) is set to the start time of the first bidding time slot (the input time slot at t=1) among multiple bidding time slots, or a time before the start time, and each of the multiple bidding time slots is designated as a bidding timing. The number of bids (bid quantity) and bid price for each bidding time slot are calculated as the solution to the optimization problem. When the bidding time slot (target time slot) containing the target time ends, the calculation start time (target time) is updated to the start time of the next bidding time slot, and based on the updated calculation start time, the number of bids (bid quantity) and bid price for each subsequent bidding time slot are calculated as the solution to the optimization problem.

[0063] Regarding equation (3), the market price p1 corresponding to the bid amount in the first bidding time frame (t=1) is predicted based on the market price p0 shown by the actual data before the start of the first bidding time frame and the corresponding coefficient. The market price corresponding to the bid amount in the Xth bidding time frame (where X is an integer between 2 and N) is the market price p predicted for the X-1st bidding time frame. x-1 The prediction is made based on the corresponding coefficients.

[0064] Optimization problems can be solved using various methods for solving mathematical programming models. One example is an algorithm based on pruning search. The following describes an algorithm based on pruning search.

[0065] Search is a method of enumerating all possible combinations of solutions and finding a pattern that matches the constraints and objective function.

[0066] All bid patterns for a given bidding time frame t (t=1...T) can be represented as a tree structure. Figure 5 shows an example of a tree structure representing all bid patterns for a predicted imbalance X.

[0067] Every path from the root of the tree to each leaf corresponds to all possible bidding patterns, and each pattern represents a single feasible solution. Each path contains the number of bids for each bidding time frame from the start of the calculation to the closing time of the market. In other words, multiple nodes are generated that relate to multiple bid amounts in the Y-th (Y is an integer greater than or equal to 1) bidding time frame, and as child nodes of these nodes, multiple nodes representing multiple bid amounts in the Y+1-th bidding time frame are generated. By recursively repeating this process, a tree structure in which multiple nodes are hierarchically connected is generated.

[0068] The revenue corresponding to the path can be calculated based on equations (1) to (3). Market forecast price p t The number of bids is u t and the bidding time frame t and the number of bids u t This can be calculated using the price impact coefficient corresponding to the pair. Since there is always a path (bid pattern) that maximizes total revenue among all paths, the optimal bid pattern can be found by exploring all paths.

[0069] In other words, from among multiple paths leading from the root node to the leaf nodes, the path that yields the maximum total revenue is selected, and the bid amounts included in the selected path are determined as the bid amounts for each of the first to N bidding time slots. Here, the market price associated with each node is predicted based on equation (3) as follows: For example, for the first bidding time slot, the market price corresponding to multiple bid amounts is predicted based on the market price of the actual data and a coefficient corresponding to the bid amount. The predicted market price is associated with the node associated with the bid amount. For the Xth bidding time slot (where X is an integer between 2 and N), the market price corresponding to multiple bid amounts is predicted based on the market price associated with the parent node of multiple nodes and a coefficient corresponding to the bid amount. The predicted market price is associated with the node associated with the bid amount.

[0070] (Output data acquisition unit 19) The output data acquisition unit 19 acquires output data that includes the number of bids and bid prices (predicted market prices) for each bidding time frame calculated by the optimal trading strategy calculation unit 18, specifically for the target time frame (the bidding time frame with t=1, and the bidding time frame with the calculation start time as the start time), as the number of bids and bid prices recommended to the user. The output data may also include the number of bids and input prices calculated for each bidding time frame that is later in time than the target time frame (bidding time frames from t=2 onwards). However, if the market prediction price calculated by the optimal trading strategy calculation unit 18 is below the lower limit price (in the case of selling) or above the upper limit price (in the case of buying), the output data acquisition unit 19 will recommend the lower limit price (in the case of selling) or the upper limit price (in the case of buying) to the user as the recommended bid price instead of the calculated price (market prediction price). The recommended bid quantity to the user may be the calculated bid quantity. Setting a lower or upper limit on the bid price may reduce the possibility of execution, but it can reduce the losses of the power generation operator. Therefore, this leads to stable revenue for power generation operators. In addition, the output data acquisition unit 19 may include the amount of electricity bid, which is the number of bids multiplied by the minimum bid unit, in the output data. The number of bids or the amount of electricity bid is just one example of the amount of electricity bid.

[0071] (Output section 20) The output unit 20 outputs the output data obtained by the output data acquisition unit 19 in a way that is visible to the user. The output unit 20 may be a display that shows the data, a communication device that transmits the data to the user terminal, or it may include both.

[0072] The user determines the number of bids and the price to bid within the target time frame (bid time frame t=1) based on the output data presented by the output unit 20. For example, the user decides to bid with the number of bids and input price shown in the output data. The user inputs bid data, including the number of bids and bid price, from the user device 51 within the target time frame, and the output unit 20 transmits the bid data to the pre-market trading device 21. Alternatively, the bid data may be transmitted from the user device 51 to the pre-market trading device 21. The pre-market trading device 21 inputs the bid data into the pre-market for the target product based on the bid data. The pre-market trading device 21 executes bids (sell bids and buy bids) that match in price from the user and various other users. The user's bid may or may not be executed. The user may also place a bid by specifying only the number of bids without specifying a bid price (market order). This allows the bid to be executed immediately at the market price, reducing the risk of the bid not being executed.

[0073] Once the target time slot ends, the next 30-minute bidding time slot is set as the target time slot, and the optimization calculation described above is performed with the start time of the target time slot as the calculation start time. That is, the bidding time slot where t=2 in the previous optimization calculation is set as the bidding time slot (target time slot) where t=1 in the current calculation. Before performing the calculation, the optimal trading strategy calculation unit 18 receives the latest predicted imbalance and the latest market trading data, updates the actual market price p0, and reflects the amount of electricity corresponding to the number of executed bids in the predicted imbalance. The optimal trading strategy calculation unit 18 performs the optimization calculation and determines the number of bids and bid price (market predicted price) for each bidding time slot from the target time slot onward. The output data acquisition unit 19 acquires output data in which at least the number of bids and bid price for the target time slot are set as the recommended number of bids and recommended bid price, respectively, from the number of bids and bid price for each bidding time slot calculated by the optimal trading strategy calculation unit 18. Once the current target time slot ends, the next bidding time slot is set as the target time slot, and the same process is repeated thereafter.

[0074] Figure 6 illustrates a specific example of the processing performed by the optimal trading strategy calculation unit 18. In the same figure as Figure 3, each bidding time slot from the current time onward is assigned the codes F1 to F15. It is assumed that the user has not placed any bids from the opening time until 19:30, and will place bids from bidding time slot F1 onward. Initially, bidding time slot F1 is set as the target time slot, and since there are 15 bidding time slots from the target time slot onward, T=15. Bidding time slot F1 corresponds to t=1, bidding time slot F2 to t=2, ..., and bidding time slot F15 to t=15. The optimization calculation is performed with the current time as the calculation start time, and output data including the recommended number of bids and bid price for bidding time slot F1 is presented, and the user places a bid in bidding time slot F1. Note that in addition to placing bids with the number of bids and prices shown in the output data, the user may also place bids with different numbers of bids and different bid prices at their own discretion.

[0075] When bidding time slot F1 ends, bidding time slot F2 is set as the next target time slot. Since there are 14 bidding time slots after the target time slot, T=14. Bidding time slot F2 corresponds to t=1, bidding time slot F3 to t=2, ..., and bidding time slot F15 to t=14. The start time of bidding time slot F2 is used as the calculation start time for the optimization calculation. At this time, as described above, the latest predicted imbalance and the latest market transaction data are received to update the market price p0, and the amount of electricity for the number of executed bids is reflected in the predicted imbalance X. As a result of the calculation, output data including the recommended number of bids and bid price for bidding time slot F2 is presented, and the user places a bid in bidding time slot F2.

[0076] The process is repeated in the same manner until the calculation for bidding time slot F15 is completed. The process may be terminated when the predicted imbalance falls below the minimum bid unit. Alternatively, even if the predicted imbalance falls below the minimum bid unit, it may be updated to exceed the minimum bid unit later, so the process may be continued until the calculation for bidding time slot F15 is completed. While the predicted imbalance is below the minimum bid unit, the number of bids in the optimization result may be zero.

[0077] Figure 7 shows an example of recommended result data, which displays the recommended bid price and recommended bid amount (recommended number of bids × minimum bid unit) for each bidding time frame F1 to F15 recommended by the optimal trading strategy calculation unit 18. The optimal trading strategy calculation unit 18 may generate the recommended result data after the last bidding time frame F15 has ended and present the recommended result data to the user via the output data acquisition unit 19 and the output unit 20. Alternatively, the recommended result data may be updated as the bidding time frame progresses and presented to the user sequentially. The user may verify whether they actually placed bids correctly according to the recommended bids by comparing the recommended result data with their actual trading result history.

[0078] Figure 8 is a flowchart of an example of the processing of the trading strategy calculation device 1 according to this embodiment. In this process, a trading strategy (number of bids and bid price for each bidding time frame) is calculated for a product slot during a certain time period on the target day in the pre-market.

[0079] The price impact table generation unit 12 generates a price impact table based on past market transaction data and sends the generated price impact table to the optimal trading strategy calculation unit 18 via the price impact table input unit 13 and the input data acquisition unit 17 (S101). The price impact table generation unit 12 may also read a pre-created price impact table and send it to the optimal trading strategy calculation unit 18.

[0080] The predicted imbalance calculation unit 16 calculates the predicted imbalance based on the predicted power generation data and the planned power generation data for the time period of the product slot (S102).

[0081] The input data acquisition unit 17 or the optimal trading strategy calculation unit 18 calculates the number of bidding time slots until the market closes based on the remaining time from the calculation start time to the closing time (S103). For simplicity of explanation, the calculation start time is assumed to be the start time of the bidding time slot in which the first trade begins. The input data acquisition unit 17 or the optimal trading strategy calculation unit 18 calculates the number of bidding time slots until the market closes based on the remaining time from the determined calculation start time to the closing time.

[0082] The optimal trading strategy calculation unit 18 obtains the latest market price immediately preceding the calculation start time based on market trading performance data. Based on the latest market price, price impact table, predicted imbalance, predicted imbalance fee value, and number of bidding time slots, the optimal trading strategy calculation unit 18 calculates the optimal number of inputs and bid price for each bidding time slot after the calculation start time, subject to constraints and optimization criteria (S104). The bidding time slot including the calculation start time corresponds to the target time slot.

[0083] The optimal trading strategy calculation unit 18 generates output data that includes the number of bids and bid prices for at least the target time slot among the bidding time slots, as the number of bids and bid prices recommended to the user. The optimal trading strategy calculation unit 18 presents the output data to the user via the output data acquisition unit 19 and the output unit 20 (S105).

[0084] The optimal trading strategy calculation unit 18 determines whether the calculation for the last bidding time slot has been completed (S106). If it has been completed, it terminates this process. If the calculation for the last bidding time slot has not been completed, the optimal trading strategy calculation unit 18 waits for the calculation start time for the next bidding time slot to begin (S107). When the calculation start time for the next bidding time slot arrives, it performs the processes in steps S102 to S105.

[0085] In this embodiment, a price impact table is defined that reflects the impact of the power generator's own trading on market prices and the risk of future price fluctuations. Based on the price impact table, the number of bids and bid prices to be placed in the target time slot are determined according to the number of remaining bidding time slots (remaining time) until the end of the trading period, under optimization criteria and constraints. This makes it possible to reduce or eliminate the predicted imbalance, which is the difference between the planned power generation value and the predicted power generation value, by the time the market closes. In this way, an optimal trading plan (recommended number of bids and bid prices for each bidding time slot) can be provided that reduces the power generator's imbalance penalty and maximizes profits.

[0086] (Variation 1) In the embodiment described above, the profit for all paths was calculated in a tree structure, and the path that yielded the maximum profit was determined by search. While this exhaustive search can find the optimal solution, the number of bidding patterns becomes enormous when the predicted imbalance is large or when the number of bidding time slots is large. In other words, the size of the search space becomes large, and it takes time to obtain the optimal solution.

[0087] In this modified version, pruning search is used as the algorithm for the optimal trading strategy calculation unit 18 in order to reduce the size of the search space. In pruning search, nodes and their descendants that are unlikely to be the optimal solution in the tree structure are ignored. This allows the optimal solution to be calculated from a smaller search space than a brute-force search, thus reducing the computational complexity.

[0088] Figure 9 is an explanatory diagram of pruning search. At each node in the tree structure, the number of bids u for each node included in the path from the root to that node. t From there, the number of remaining bids is l t This is calculated using the following formula (5). TIFF0007864625000011.tif13161

[0089] JPEG0007864625000012.jpg28170

[0090] The optimal trading strategy calculation unit 18 calculates the partial profit (partial transaction amount) from bids up to bidding time frame t using the following formula (6) based on the bids for each bidding time frame from the first bidding time frame (initial calculation start time) to bidding time frame t. t Calculate [yen]. TIFF0007864625000013.tif10161

[0091] Equation (6) allows us to calculate partial revenue at any node in the tree structure. Based on the relative magnitudes of the partial revenues at each node, we determine whether there is a prospect of reaching the optimal solution and decide which nodes to continue the search at and which to terminate the search at.

[0092] JPEG0007864625000014.jpg30170

[0093] Thus, if multiple nodes with the same number of remaining bids exist within the same bidding time frame t, nodes with both low partial revenue and low market forecast price are expected to have lower total revenue until market close compared to nodes with high partial revenue and low market forecast price. Therefore, we decide not to explore the descendants of nodes with both low partial revenue and low market forecast price.

[0094] For example, suppose circle A is larger than circle B, and circle a is larger than circle b. In this case, nodes from node N2 onward are excluded from the search. In the diagram, excluding nodes from the search, i.e., pruning, is represented by "×". This reduces the search space and makes the search for solutions more efficient.

[0095] (Modification 2) In the embodiment described above, the number of bids and the bid price were optimized, but the amount of electricity bid (number of bids × minimum bid unit) and the bid price may also be optimized. In this case, equations (1) to (3) in the optimization problem may be changed to equations (1A) to (3A) below. In this case, the rows of the price impact table are changed to u t From q t Change it. TIFF0007864625000015.tif10161TIFF0007864625000016.tif13161TIFF0007864625000017.tif6161If the bid amount obtained as a solution to the optimization problem is not a multiple of the minimum bid unit, the fractional part of the bid amount may be truncated so that it becomes a multiple of the minimum bid unit.The output data presented to the user shall include at least the bid amount and bid price that are recommended to be bid on within the target time frame.The user may convert the bid amount into the number of bids and decide the actual number of bids to be bid.Alternatively, the device may convert the bid amount into the number of bids and include the converted number of bids in the output data.By optimizing the bid amount instead of the number of bids in this way, the number of parameters in the objective function and constraints can be reduced, and thus the computational load can be reduced.

[0096] (Variation 3) In the embodiment described above, if, after a user has placed a bid within the target time frame, new market transaction data is acquired, such as when another user's bid is executed, the price impact table may be updated and the number of bids and input price may be re-determined. The device may then present output data to the user, including the re-determined number of bids and bid price. The user may then cancel their bid and place a new bid with the re-determined number of bids and input price. This allows for bidding that reflects the latest market transaction data, which is expected to increase profits.

[0097] (Modification 4) If there are any unexecuted bids after a predetermined time has elapsed from the start time of the last bidding time slot (bidding time slot F15 in the example in Figure 6), the bid prices may be changed to immediately execute all bids. For example, if a user's bid is a sell bid, the sell bids will be matched starting with the highest-priced buy bid among all buy bids in the pre-market. While this may reduce revenue, it reduces the predicted imbalance compared to when bids are not executed. Therefore, it can reduce the actual penalty incurred by the power generator.

[0098] (Hardware configuration) Figure 10 shows the hardware configuration of the information processing device (trading strategy calculation device) shown in Figure 1. The device in Figure 1 consists of a computer device 600. The computer device 600 includes a CPU 601, an input interface 602, a display device 603, a communication device 604, a main memory 605, and an external memory device 606, which are interconnected by a bus 607.

[0099] The CPU (Central Processing Unit) 601 executes an information processing program, which is a computer program, on the main memory 605. The information processing program is a program that implements each of the above-described functional configurations of this device. The information processing program may not be a single program, but rather a combination of multiple programs or scripts. Each functional configuration is realized when the CPU 601 executes the information processing program.

[0100] The input interface 602 is a circuit for inputting operation signals from input devices such as keyboards, mice, and touch panels to this device. The input interface 602 corresponds to the user device 51.

[0101] The display device 603 displays data output from this device. The display device 603 is, for example, an LCD (liquid crystal display), an organic electroluminescent display, a CRT (cathode ray tube), or a PDP (plasma display), but is not limited to these. Data output from the computer device 600 can be displayed on this display device 603. The display device 603 corresponds to the output unit 20.

[0102] The communication device 604 is a circuit for this device to communicate with an external device wirelessly or via a wired connection. Data can be input from an external device via the communication device 604. The data input from the external device can be stored in the main memory 605 or the external memory 606.

[0103] The main memory 605 stores information processing programs, data necessary for the execution of the information processing programs, and data generated by the execution of the information processing programs. The information processing programs are deployed and executed on the main memory 605. The main memory 605 is, for example, RAM, DRAM, or SRAM, but is not limited to these. Each storage unit or database in Figure 1 may be built on the main memory 605.

[0104] The external storage device 606 stores information processing programs, data necessary for executing the information processing programs, and data generated by the execution of the information processing programs. These information processing programs and data are read into the main memory 605 when the information processing programs are executed. The external storage device 606 is, for example, a hard disk, optical disk, flash memory, and magnetic tape, but is not limited to these. Each storage unit or database in Figure 1 may be built on the external storage device 606.

[0105] The information processing program may be pre-installed on the computer device 600, or it may be stored on a storage medium such as a CD-ROM. Furthermore, the information processing program may be uploaded to the internet.

[0106] Furthermore, this device may consist of a single computer device 600, or it may be configured as a system consisting of multiple interconnected computer devices 600.

[0107] It should be noted that the present invention is not limited to the embodiments described above, and the components can be modified and implemented in practice without departing from the spirit of the invention. Furthermore, various inventions can be formed by appropriately combining the multiple components disclosed in the embodiments described above. For example, a configuration in which some components are removed from all the components shown in each embodiment is also conceivable. Moreover, components described in different embodiments may be appropriately combined.

[0108] This embodiment can also be configured as follows. [Item 1] A processing unit determines one or more bidding timings to be used for bidding during the trading period after the target time, and the amount of bids to be placed in the trading market, based on a price fluctuation model that represents the fluctuation of the market price in accordance with the elapsed time from a target time and the amount of bids to be placed in the trading market during the trading period of the product. Equipped with an information processing device. [Item 2] The processing unit predicts the market price at the bidding timing based on the price fluctuation model, The processing unit determines the one or more bidding timings and the bid amounts based on constraints relating to the total bid amounts to be submitted at the one or more bidding timings and optimization criteria for the total transaction amount based on the bid amounts and the market price. The information processing device described in item 1. [Item 3] The processing unit further determines the bid price to be placed in the trading market based on the predicted market price. The information processing device described in item 2. [Item 4] The aforementioned product is electricity, The constraint is that the total amount of bids matches the difference between the predicted power generation amount of the power generation equipment of the power generation company and the planned power supply amount of the power generation company, or the difference between the total and the difference is less than or equal to the minimum bid unit of the bid amount. An information processing device as described in item 2 or 3. [Item 5] The optimization criterion is to maximize the total transaction amount when the bid is a sell bid, or to minimize the total transaction amount when the bid is a buy bid. An information processing device as described in any one of items 2 to 4. [Item 6] The processing unit generates output data that recommends to the user the bidding timing and bid amount to be placed in the trading market. An information processing device described in any one of items 2 to 5. [Item 7] The processing unit predicts the market price at the bidding timing based on the price fluctuation model, and includes the predicted market price in the output data as a bid price that recommends bidding in the trading market. The information processing device described in item 6. [Item 8] If the predicted market price exceeds the upper limit price, the processing unit includes the upper limit price in the output data as the recommended bid price for bidding in the trading market; if the predicted market price falls below the lower limit price, it includes the lower limit price in the output data as the recommended bid price for bidding in the trading market. The information processing device described in item 7. [Item 9] The system includes a market data acquisition unit that acquires transaction performance data from a trading market device that processes transactions in the aforementioned trading market. The processing unit calculates the amount of bids that were executed from among the bids made based on the output data to the trading market based on the actual data, subtracts the calculated amount of bids from the total amount of bids, The processing unit updates the target time and, based on the updated target time, determines one or more bidding timings and bid amounts for submitting bids with the reduced bid amount. An information processing device as described in any one of items 6 to 8. [Item 10] The aforementioned trading period includes multiple bidding time slots, The processing unit sets the start time of the first of the multiple bidding time slots, or a time prior to the start time, as the target time, and determines the bid amount for each of the multiple bidding time slots, with each of the multiple bidding time slots being the bidding timing. An information processing device as described in any one of items 2 through 9. [Item 11] The price fluctuation model includes multiple coefficients corresponding to multiple bid quantities for each of the first to N bidding time slots, The system includes a market data acquisition unit that acquires transaction performance data from a trading market device that processes transactions in the aforementioned trading market. The processing unit generates multiple nodes related to multiple bid amounts in the Y-th (Y is an integer greater than or equal to 1) bidding time frame, and generates multiple nodes related to multiple bid amounts in the Y+1-th bidding time frame as child nodes of the said nodes, thereby generating a tree structure in which multiple nodes are hierarchically connected. For the first bidding time frame, the market price corresponding to the multiple bid quantities is predicted based on the market price of the actual data and the coefficient corresponding to the bid quantity, and the predicted market price is associated with the multiple nodes related to the multiple bid quantities. For a bidding time frame of number X (where X is an integer between 2 and N), market prices corresponding to multiple bid quantities are predicted based on the market price associated with the parent node of the multiple nodes and the coefficient corresponding to the bid quantity, and the predicted market prices are associated with the multiple nodes associated with the multiple bid quantities. For each of the multiple paths from the root node to the leaf nodes of the tree structure, the total transaction amount is calculated based on the bid amount and market price associated with the nodes included in the path. Based on the calculated total transaction amount, a path is selected from the multiple paths, and the bid amount associated with each node included in the selected path is determined as the bid amount for each of the first to Nth bidding time slots. The information processing device described in item 10. [Item 12] For the bidding time frame of Y+1, identify the partial paths from the root node to each of the multiple nodes generated, calculate the partial transaction amount at each of the multiple nodes based on the partial paths, compare the partial transaction amount at each of the multiple nodes with the market price, select a node from the multiple nodes, and do not generate child nodes for the selected node. The information processing device described in item 11. [Item 13] The processing unit predicts the market price by multiplying the market price by the coefficient and adding the multiplication result to the market price. The information processing device described in item 11 or 12. [Item 14] The processing unit calculates a coefficient that is larger or smaller the later the bidding time frame and the larger the bid amount. The information processing apparatus according to claim 11. [Item 15] The processing unit calculates market price fluctuation information based on past transaction data in the trading market, generates random numbers based on the fluctuation information, and calculates the coefficient using the random numbers. The information processing device described in item 14. [Item 16] The aforementioned fluctuation information represents the rate of change in the market price, The processing unit determines a range of values ​​for the random number based on the rate of change, the number of the first to Nth bidding time slots, and the number of the multiple bid amounts, and generates the random number from that range. The information processing device described in item 15. [Item 17] The range of values ​​for the aforementioned random numbers is the range in which the predicted rate of change in the market price is less than or equal to the rate of change. The information processing device described in item 16. [Item 18] The market data acquisition unit acquires updated transaction performance data from the trading market device. The processing unit updates the volatility rate based on the updated transaction performance data, and updates the price fluctuation model based on the updated volatility rate. The information processing device described in item 16 or 17. [Item 19] The aforementioned trading market is the pre-hours market for electricity, The aforementioned planned power supply amount is the amount of electricity sold by the power generator in the market one day prior. The information processing device described in item 4. [Item 20] The aforementioned product is electricity. Information processing devices as described in items 1-19. [Item 21] The processing unit further determines the bid price to be submitted to the trading market. An information processing device described in any one of items 1 to 20. [Item 22] Based on a price fluctuation model that represents the price fluctuation of the commodity in relation to the elapsed time from a target time and the bid volume during the trading period of the commodity in the trading market, one or more bidding timings to be used for bidding during the period after the target time and the bid volume to be used for bidding in the trading market are determined. Information processing methods. [Item 23] A step of determining one or more bidding timings to place bids during the period after the target time, and the amount of bids to place in the trading market, based on a price fluctuation model that represents the price fluctuation of the said product in accordance with the elapsed time from the target time and the amount of bids placed in the trading market during the period in which the product is available for trading in the trading market. A computer program that causes a computer to execute something. [Explanation of symbols]

[0109] 11 Market Data Acquisition Department 12. Price Impact Table Generation Unit (Processing Unit) 13. Price Impact Table Input Section 14. Power Generation Company Database 15. Power generation prediction unit (processing unit) 16 Predicted Imbalance Calculation Unit (Processing Unit) 17 Input Data Acquisition Unit 18. Optimal Trading Strategy Calculation Unit (Processing Unit) 19 Output data acquisition unit 20 Output section 21 hours ago Market trading device 31 Power generation equipment 41 Weather Server 51 User device 600 Computer devices 602 Input Interface 603 Display device 604 Communication equipment 605 Main storage 606 External storage device 607 Bus

Claims

1. A processing unit determines one or more bidding timings to be used for bidding during the trading period after the target time, and the amount of bids to be placed in the trading market, based on a price fluctuation model that represents the fluctuation of the market price in accordance with the elapsed time from a target time and the amount of bids to be placed in the trading market during the trading period of the product. Equipped with an information processing device.

2. The processing unit predicts the market price at the bidding timing based on the price fluctuation model, The processing unit determines the one or more bidding timings and the bid amounts based on constraints relating to the total bid amounts to be submitted at the one or more bidding timings and optimization criteria for the total transaction amount based on the bid amounts and the market price. The information processing apparatus according to claim 1.

3. The processing unit further determines the bid price to be placed in the trading market based on the predicted market price. The information processing apparatus according to claim 2.

4. The aforementioned product is electricity, The constraint is that the total amount of bids matches the difference between the predicted power generation amount of the power generation equipment of the power generation company and the planned power supply amount of the power generation company, or the difference between the total and the difference is less than or equal to the minimum bid unit of the bid amount. The information processing apparatus according to claim 2.

5. The optimization criterion is to maximize the total transaction amount when the bid is a sell bid, or to minimize the total transaction amount when the bid is a buy bid. The information processing apparatus according to claim 2.

6. The processing unit generates output data that recommends to the user the bidding timing and bid amount to be placed in the trading market. The information processing apparatus according to claim 2.

7. The processing unit predicts the market price at the bidding timing based on the price fluctuation model, and includes the predicted market price in the output data as a bid price that recommends bidding in the trading market. The information processing apparatus according to claim 6.

8. If the predicted market price exceeds the upper limit price, the processing unit includes the upper limit price in the output data as the recommended bid price for bidding in the trading market; if the predicted market price falls below the lower limit price, it includes the lower limit price in the output data as the recommended bid price for bidding in the trading market. The information processing apparatus according to claim 7.

9. The system includes a market data acquisition unit that acquires transaction performance data from a trading market device that processes transactions in the aforementioned trading market. The processing unit calculates the amount of bids that were executed from among the bids made based on the output data to the trading market based on the actual data, subtracts the calculated amount of bids from the total amount of bids, The processing unit updates the target time and, based on the updated target time, determines one or more bidding timings and bid amounts for submitting bids with the reduced bid amount. The information processing apparatus according to claim 6.

10. The aforementioned trading period includes multiple bidding time slots, The processing unit sets the start time of the first of the multiple bidding time slots, or a time prior to the start time, as the target time, and determines the bid amount for each of the multiple bidding time slots, with each of the multiple bidding time slots being the bidding timing. The information processing apparatus according to any one of claims 2 to 9.

11. The price fluctuation model includes multiple coefficients corresponding to multiple bid quantities for each of the first to N bidding time slots, The system includes a market data acquisition unit that acquires transaction performance data from a trading market device that processes transactions in the aforementioned trading market. The processing unit generates multiple nodes related to multiple bid amounts in the Y-th (Y is an integer of 1 or more) bidding time frame, and generates multiple nodes related to multiple bid amounts in the Y+1-th bidding time frame as child nodes of the said nodes, thereby generating a tree structure in which multiple nodes are hierarchically connected. For the first bidding time frame, the market price corresponding to the multiple bid quantities is predicted based on the market price of the actual data and the coefficient corresponding to the bid quantity, and the predicted market price is associated with the multiple nodes related to the multiple bid quantities. For a bidding time frame of number X (where X is an integer between 2 and N), market prices corresponding to multiple bid quantities are predicted based on the market price associated with the parent node of the multiple nodes and the coefficient corresponding to the bid quantity, and the predicted market prices are associated with the multiple nodes associated with the multiple bid quantities. For each of the multiple paths from the root node to the leaf nodes of the tree structure, the total transaction amount is calculated based on the bid amount and market price associated with the nodes included in the path. Based on the calculated total transaction amount, a path is selected from the multiple paths, and the bid amount associated with each node included in the selected path is determined as the bid amount for each of the first to Nth bidding time slots. The information processing apparatus according to claim 10.

12. For the bidding time frame of Y+1, identify the partial paths from the root node to each of the multiple nodes generated, calculate the partial transaction amount at each of the multiple nodes based on the partial paths, compare the partial transaction amount at each of the multiple nodes with the market price, select a node from the multiple nodes, and do not generate child nodes for the selected node. The information processing apparatus according to claim 11.

13. The processing unit predicts the market price by multiplying the market price by the coefficient and adding the multiplication result to the market price. The information processing apparatus according to claim 11.

14. The processing unit calculates a coefficient that is larger or smaller the later the bidding time frame and the larger the bid amount. The information processing apparatus according to claim 11.

15. The processing unit calculates market price fluctuation information based on past transaction data in the trading market, generates random numbers based on the fluctuation information, and calculates the coefficient using the random numbers. The information processing apparatus according to claim 14.

16. The aforementioned fluctuation information represents the rate of change in the market price, The processing unit determines a range of values ​​for the random number based on the rate of change, the number of the first to Nth bidding time slots, and the number of the multiple bid amounts, and generates the random number from the range. The information processing apparatus according to claim 15.

17. The range of values ​​for the aforementioned random numbers is the range in which the predicted rate of change in the market price is less than or equal to the rate of change. The information processing apparatus according to claim 16.

18. The market data acquisition unit acquires updated transaction performance data from the trading market device. The processing unit updates the volatility rate based on the updated transaction performance data, and updates the price fluctuation model based on the updated volatility rate. The information processing apparatus according to claim 16.

19. The aforementioned trading market is the pre-hours market for electricity, The aforementioned planned power supply amount is the amount of electricity sold by the power generator in the market one day prior. The information processing apparatus according to claim 4.

20. The aforementioned product is electricity. The information processing apparatus according to claim 1.

21. The processing unit further determines the bid price to be submitted to the trading market. The information processing apparatus according to claim 1.

22. Based on a price fluctuation model that represents the price fluctuation of the commodity in relation to the elapsed time from a target time and the bid volume during the trading period of the commodity in the trading market, one or more bidding timings to be used for bidding during the period after the target time and the bid volume to be used for bidding in the trading market are determined. Information processing methods.

23. A step of determining one or more bidding timings to place bids during the period after the target time, and the amount of bids to place in the trading market, based on a price fluctuation model that represents the price fluctuation of the said product in accordance with the elapsed time from the target time and the amount of bids placed in the trading market during the period in which the product is available for trading in the trading market. A computer program that causes a computer to execute something.

Images