JP2007122592A - System, method and program for supporting wholesale merchandise market bidding price setting - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To support the setting of an optimum bidding price for maximizing a profit expected value in a merchandise spot market. <P>SOLUTION: A past time-series data acquiring part 21 acquires spot price past time-series data agreed in a spot market in the past, and a profit calculating data acquiring part 22 acquires production cost data or retail price data as profit calculating data for calculating an expected value of a profit index. A time-series analyzing part 23 generates time-series (predicted price time-series) of a predicted value ahead by one term of a past spot price and time-series of standard deviation or dispersion of predicted errors between the predicted value ahead by one term and a result value ahead by one term by analyzing the spot price past time-series data. A bidding price optimizing part 24 uses a predicted price time-series, the time-series of the standard deviation or dispersion of the predicted errors and the production cost data or the retail price data to thereby calculate an optimum bidding price to maximize an expected value of a preset profit index for a sales transaction and a purchase transaction. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to a technique for supporting price setting for a market participant to bid in order to conduct a sales transaction in a spot market of wholesale merchandise.

  A wholesale spot market has been established for various products such as industrial products, agricultural products, and energy products. In particular, in the spot market of wholesale products where the price fluctuates greatly, in order for market participants to establish their own transactions, that is, to acquire (successfully bid) trading rights, estimate the contract price in the wholesale product market in advance. It is necessary to set an appropriate bid price.

  For example, in the case of a selling transaction, it is necessary to set a bid price that is lower than a contract price that is unknown and that provides an appropriate profit in relation to the production cost of the product. In the case of a buy transaction, it is necessary to set a bid price that is higher than the contract price and that provides an appropriate profit in relation to the product production cost and retail price.

Conventionally, as a method for supporting the setting of an appropriate bid price, for example, there is a method for evaluating a price based on past bid information, product demand information, production cost information, and the like (for example, patents). References 1 and 2).
JP 2004-171180 A JP2005-25377

  By the way, there is an error in the estimation of the contract price of the product in the next day market, so assuming that the actual contract price is normally distributed around the estimated value, when the bidder is bidding on the estimated price, The probability that a company can make a successful bid is about half when there are competitors in the market.

  For example, when there is a competitor in the market in a selling transaction, in order for a market participant to make a successful bid in-house, the bid price for the amount he wants to sell must be less than or equal to the contract price. On the other hand, if the bid price is set too low, an appropriate profit cannot be obtained even if a successful bid can be made.

  In addition, in order for a market participant to make a successful bid in a buying transaction, the bid price must be greater than or equal to the contract price. In this case, if the bid price is set too high, even if successful bids can be made, if the difference from the retail price is small, appropriate profits cannot be obtained. If so, it would be useless to bid at a lower price than your production costs.

  Therefore, it is important to set a bid price that maximizes the expected value of profit due to market transactions. However, in the prior art, since an optimal bid price calculation method that maximizes the expected value of profit has not been established, it has been difficult to set an optimal bid price.

  The present invention has been proposed to solve the above-described problems, and its purpose is to provide wholesale products that can support the setting of the optimal bid price that maximizes the expected value of profit in the spot market of wholesale products. It is to provide a market bidding pricing support system and method, and a program.

  In order to achieve the above-mentioned object, the present invention analyzes time series data of past spot prices of wholesale products, time series of one-term forecast value, one-term forecast value, and one-term forecast value To generate the standard deviation or variance time series of the forecast error and to reflect the next day's execution price and the standard deviation or variance of the forecast error, and to ensure the optimal bid price that maximizes the expected profit It can be calculated well.

  A wholesale product market bidding price setting support system according to the present invention is a wholesale product market bid price setting support system that uses a computer to support the price setting for a market participant to bid in the spot market of wholesale products. , Time series analysis means, bid price optimization means, storage means, and interface means. Here, the time series analysis means analyzes the time series data of the spot price previously executed in the spot market, and thereby the time series of the one-term forecast value of the past spot price, the one-term forecast value, This is a means for generating a standard deviation or a time series of variance of the prediction error from the one-term past actual value. The bid price optimizing means uses the time series of the one-term forecast value and the standard deviation or variance time series to maximize the expected value of the profit index preset for the sell transaction and the buy transaction. This is a means for calculating an optimal bid price. The storage means is means for storing calculation conditions and calculation results, and the interface means is means for inputting data and displaying results.

  The wholesale merchandise market bid price setting support method and the wholesale merchandise market bid price setting support program of the present invention grasp the characteristics of the system from the viewpoints of the method and the computer program.

The definitions of important terms in the present invention are as follows.
The “spot market for wholesale products” is a broad concept that means a spot market for all wholesale products such as industrial products, agricultural products, and energy products including fuels such as electric power, crude oil, heavy oil, and natural gas.
“First period” indicates a unit period of transactions in the spot market of the target wholesale product, and is generally “one day”.

The “expected value of profit index” is appropriately set according to the wholesale product. Generally, in the case of selling transaction, it can be set as “expected value of successful bid amount x (contract price-production cost)”. In addition, in the case of a buying transaction, it can be set as “successful bid amount × (retail price−contract price)” or “expected value of successful bid amount × (production cost−contract price)”.

  According to the present invention, it is possible to provide a wholesale product market bid price setting support system, method, and program capable of supporting the setting of an optimal bid price that maximizes the expected value of profit in the spot market of wholesale products. .

  Hereinafter, a plurality of embodiments of a wholesale product market bid price setting support system according to the present invention will be specifically described with reference to the drawings.

[First Embodiment]
[Constitution]
FIG. 1 is a block diagram showing a configuration of a wholesale product market bid setting support system according to the first embodiment of the present invention. As shown in FIG. 1, a wholesale product market bid price setting support system 1 according to this embodiment includes a calculation unit 2, an interface unit 3, and a storage unit 4 realized on a computer. The interface unit 3 is connected to a market transaction support system 6 that supports transactions in the wholesale merchandise market or an exchange through a communication medium 5 such as a cable or a network. Details of each part 1 to 4 of the system are as follows.

  The calculation unit 2 includes a past time series data acquisition unit 21, a profit calculation data acquisition unit 22, a time series analysis unit 23, a bid price optimization unit 24, and the like. Here, the past time series data acquisition unit 21 is a part that acquires time series data of spot prices (hereinafter referred to as spot price past time series data) of wholesale products promised in the past in the spot market, The profit calculation data acquisition unit 22 is a part that acquires production cost data or retail price data as profit calculation data for calculating the expected value of the profit index.

  The time series analysis unit 23 analyzes the spot price past time series data, thereby making a time series of the past spot price predicted value for one period ahead (hereinafter referred to as a predicted price time series) and a one period ahead forecast. This is a part for generating a time series of standard deviation (or variance defined by the square of the standard deviation) of the prediction error between the value and the one-year past actual value. In addition, the bid price optimization unit 24 is set in advance for a sell transaction and a buy transaction using a predicted price time series, a standard time series of prediction errors, and production cost data or retail price data. This is the part that calculates the optimal bid price that maximizes the expected value of the profit indicator.

  Note that such a calculation unit 2 is specifically configured by a main memory of a computer, a program specialized for setting a bid price for a wholesale product market stored therein, a CPU controlled by the program, and the like. Realized.

  The interface unit 3 is a part that performs communication between the computer and the user and communication between the computer and an external device, an external system, and the like, and includes a data input unit 31, a data output unit 32, a data communication unit 33, and the like. ing. Here, the data input unit 31 is an input device such as a mouse or a keyboard for inputting a signal according to a user operation to the computer, and the data output unit 32 is a data input by the data input unit 31 and an arithmetic unit. 2 is an output device such as a display or a printer that displays or outputs the processing result processed in step 2 to the user. The data communication unit 33 is a communication control device that is generally built in a computer, and transmits and receives data between the computer and the communication medium 5.

  The storage unit 4 is a part that stores in advance various calculation conditions for performing data processing in the calculation unit 2 and stores the calculation results by the calculation unit 2. The storage unit 4 is realized by various types of memory and auxiliary storage devices of a computer.

[Outline of operation]
FIG. 2 is a flowchart showing an outline of the operation of the wholesale product market bid price setting support system according to this embodiment having the above-described configuration.

  As shown in FIG. 2, the calculation unit 2 of the wholesale product market bid price setting support system 1 first uses the past time series data acquisition unit 21 to store spot price past time series data of wholesale products previously executed in the spot market. Is acquired through the data input unit 31 of the interface unit 3 or the storage unit 4 (S210).

  The calculation unit 2 also uses the profit calculation data acquisition unit 22 to calculate production cost data or retail price data as profit calculation data for calculating the expected value of the profit index, and similarly to the data input unit of the interface unit 3. 31 or through the storage unit 4 (S220).

  Next, the calculation unit 2 performs time series analysis on the spot price past time series data obtained by the spot price past data acquisition unit 21 as time series analysis processing by the time series analysis unit 23 (S230). In this time series analysis process, the ARIMA model is usually applied, or the autoregressive model (AR model) or autoregressive-moving average model after extracting and removing the definite trend component and periodic component from the spot price time series. (ARMA model) is applied.

  By this time series analysis process, the spot price at the time of each transaction up to the current day can be estimated from the previous time series data, that is, the expected value of the spot price of the next day, so the spot price of the next day is estimated accordingly. can do. Therefore, a time series of predicted values of the next day spot price, that is, a time series of predicted prices can be obtained.

  In this time series analysis process, it is also possible to estimate the standard deviation of the prediction error, which is the difference between the predicted value for the next term, that is, the predicted value of the next day spot price and the actual value. In this case, since the standard deviation of the prediction error varies every transaction day, a variance variation model defined as the square of the standard deviation of the error is used. For example, an ARCH model, a GARCH model, a probability volatility model, or the like can be used as an analysis model that handles such variance fluctuations. Therefore, in addition to the time series of predicted prices, the time series of standard deviations of prediction errors can be obtained.

  In this case, the time series analysis unit 23 stores the obtained forecast price time series and the standard deviation time series of the prediction error in the storage unit 4 and also to the user by the data output unit 32 of the interface unit 3. To display.

  Next, the calculation unit 2 calculates the time series of the predicted price obtained by the time series analysis unit 23 and the standard series of the prediction error thereof, and the production cost data or retail data obtained by the profit calculation data acquisition unit 22. Based on the price data, the bid price optimization unit 24 calculates the optimal bid price for the next day as the bid price optimization process (S240). Details of the bid price optimization process will be described later.

  The bid price optimizing unit 24 stores the obtained optimal bid price on the next day in the storage unit 4 as a processing result, and displays it to the user by the data output unit 32 of the interface unit 3, and the data of the interface unit 3 The communication unit 33 transmits the optimal bid price data to the market transaction support system 6 through the communication medium 5, and performs a bid on the market from the market transaction support system 6 (S250). In the display of the processing result by the data output unit 32, in order to make it easy for the user to understand the obtained optimum bid price, the standard deviation with respect to the predicted price and the predetermined confidence level is set together with the optimum bid price. The next day price confidence interval obtained from, or if the transaction has been completed, part or all of the actual value of the price is displayed or illustrated for each wholesale product or the past value in time series May be.

[Details of bid price optimization process]
FIGS. 3 and 4 are flowcharts showing an example of the selling bid price optimization process in the case of a selling transaction and the buying bid price optimization process in the case of a buying transaction as the bid price optimization process by the bid price optimizing unit 24. is there.

First, the selling bid price optimization process shown in FIG. 3 will be described. In FIG. 3, the optimal bid price in the selling transaction is expressed by the following equation (1).

  The estimated value of the next day spot price and the standard deviation of the estimation error are values obtained by the time series analysis (S230) by the time series analysis unit 23.

The bid price optimizing unit 24 determines a successful bid for a selling transaction with respect to a given parameter value α (> 0), where S _j is a contract price executed in the market on the past j-th transaction day. The condition and profit index are defined as follows, and the selling bid price optimization process shown in FIG. 3 is performed. In the case of a selling transaction, C is the in-house production cost of the product.

  As shown in FIG. 3, in the selling bid price optimizing process, the bid price optimizing unit 24 performs a loop process (LOOP1) for each parameter value with respect to a given number of parameter values α. After the initial value of the transaction date and accumulated profit (j = 0th transaction date, accumulated profit p = 0) is set for the value α (S310), a loop process (LOOP2) for each transaction date is performed.

  In the loop processing for each transaction day (LOOP2), first, the transaction date is advanced by one (S321), and then the bid price of the next day of the transaction date is calculated by the above equation (1) for the jth transaction date. Calculate (S322). When the calculated bid price on the next day satisfies the successful bid determination condition of the above-mentioned sale transaction (YES in S323), the market participant has been able to make a successful bid at the bid price. Is calculated and added to the accumulated profit p up to that point (S324).

  On the other hand, when the calculated bid price does not satisfy the successful bid judgment condition of the sale transaction (NO in S323), the market participant cannot make a successful bid at the bid price, so there is no profit. Proceed to processing the next trading day.

  By repeating the loop processing (LOOP2) for each transaction date, j = N−1 from the first transaction date with respect to the given parameter value α for the total number (N−1) of past data. Cumulative profit p (α) up to the third transaction date is obtained (S330).

A parameter value that maximizes the accumulated profit is obtained by performing a loop process (LOOP1) for each given parameter value α to obtain the accumulated profit p (α) of the parameter value (S330). α = α _M is selected (S340).

  Here, the optimum parameter value α exists because, according to the above equation (1), if the parameter value α (> 0) is small, the bid price becomes high and the probability of being unable to make a successful bid increases. This is because if the value is increased too much, the expected value of the profit index decreases. Such a parameter value α can generally be determined using a nonlinear optimization method such as Newton's method that maximizes the profit of the objective function.

If the parameter value α = α _M is determined as described above, j = N and α = α _{M in} the above formula (1), so that the optimal bid price in the next day's selling transaction at the present time Can be calculated (S350).

Next, the purchase bid price optimization process shown in FIG. 4 will be described. In FIG. 4, the optimal bid price in the buying transaction is expressed by the following equation (2).

The bid price optimizing unit 24 defines the bid decision condition and profit index of the buying transaction as follows for the given parameter value α (> 0), and sell bid price optimization processing shown in FIG. I do. In the case of a buying transaction, C is the in-house production cost or retail price of the product.

  The contents of the buy bid price optimization process shown in FIG. 4 are the same as those shown in FIG. 3 except that the bid price, the successful bid judgment condition, and the profit index in the sell bid price optimization process shown in FIG. Since it is the same as the content of a process, detailed description about FIG. 4 is abbreviate | omitted.

[effect]
According to the first embodiment as described above, the time series data of the past spot price is analyzed, the time series of the next day forecast price (one-year forecast value), and the next day forecast price (one-term forecast). Value) and the next day's contract price (the actual value for the 1st term) by generating a time series of standard deviation (or variance defined by the square of the standard deviation) By reflecting the standard deviation of the prediction error, it is possible to accurately calculate the optimal bid price that maximizes the expected value of profit.

  Therefore, in particular, in the spot market of wholesale products where the price fluctuates greatly, it is possible to support the setting of the optimal bid price that maximizes the profit from the market transaction of the product based on the market data of the past spot price. . As a result, the burden on market participants in wholesale merchandise transactions can be reduced.

  In addition, by sending the data of the optimal bid price for the next day to the market transaction support system as it is, the market transaction support system can automatically bid on the market. It can also support bidding. As a result, it is possible to further reduce the burden on market participants in wholesale merchandise transactions.

[Second Embodiment]
[Constitution]
FIG. 5 is a block diagram showing a configuration of a wholesale product market bid price setting support system according to the second embodiment of the present invention. As shown in FIG. 5, the wholesale product market bid price setting support system 1 of the present embodiment includes a fluctuation model parameter estimation unit 25 in the time series analysis unit 23 of the calculation unit 2 in the system according to the first embodiment. And a simulation unit 26.

  Here, the fluctuation model parameter estimation unit 25 is a part that estimates the model parameters of a preset spot price fluctuation model by analyzing the spot price past time series data, and the simulation unit 26 also calculates the spot price. This is a part that generates a time series of the predicted value for one period and the standard deviation by simulating the time fluctuation of the spot price by the Monte Carlo method using the fluctuation model and model parameters.

  The other parts have the same configuration as that of the system according to the first embodiment, and thus the description thereof is omitted.

[Outline of operation]
FIG. 6 is a flowchart showing an outline of the operation of the wholesale product market bid price setting support system according to this embodiment having the above-described configuration.

  In this embodiment, the flow of processing is basically the same as the processing of the first embodiment shown in FIG. 2, and the time series analysis processing of spot price past time series data by the time series analysis unit 23 (S230). Only the time series analysis process (S230), which is a feature of the present embodiment, will be described below.

  That is, in the time series analysis process (S230) by the time series analysis unit 23 of the present embodiment, first, the fluctuation model parameter estimation unit 25 performs time series analysis of the spot price past time series data and the preset spot price fluctuation. Model parameters of the model are estimated (S231).

  Next, the simulation unit 26 uses the spot price fluctuation model and model parameter estimation values, and performs a Monte Carlo simulation of spot price temporal fluctuation using a large number of random numbers, thereby simulating a large number of simulated past spot prices. Generate time series data. Then, in the same manner as the time series analysis processing of the first embodiment generated from a large number of spot price past time series data, the predicted price time series and the prediction error are calculated from the simulated many spot price past time series data. A time series of standard deviations is generated (S232).

[effect]
According to the second embodiment as described above, the following effects can be obtained in addition to the effects of the first embodiment described above.

  That is, in the bid price optimization process, in order to improve the accuracy, abundant spot price past data is required. On the other hand, when the past data amount is not sufficient, the spot price fluctuation model is set by the time series analysis processing of the present embodiment based on the past data of the spot price currently obtained. By estimating the parameters and simulating the fluctuation of the spot price, a large number of spot price data necessary and sufficient for determining the parameter value giving the optimum bid price can be obtained.

  Therefore, according to the second embodiment, an optimal bid price that maximizes the expected value of profit can be accurately calculated even if the past data amount is not sufficient.

[Other Embodiments]
It should be noted that the present invention is not limited to the above-described embodiments, and various other variations can be implemented within the scope of the present invention. First, the system configuration and the flowchart shown in the drawings are merely examples, and the specific configuration, operation procedure, and details of each process can be selected as appropriate.

  Further, in the above-described embodiment, the case where the method of the present invention is realized as a system and method by computer hardware and a program has been described. It can also be realized in the form of a computerized computer program only.

  That is, the present invention analyzes time series data of the past spot prices of wholesale products, and calculates the standard deviation of the prediction error between the time series of the one-term forecast value and the one-term forecast value and the one-term forecast value. As long as a time series of (or variance defined by the square of the standard deviation) is generated, the specific embodiment can be freely selected. Further, the present invention can be applied to spot market transactions for all wholesale products such as energy products including fuels such as electric power, crude oil, heavy oil, and natural gas in addition to industrial products and agricultural products.

The block diagram which shows the structure of the wholesale goods market bid price setting assistance system which concerns on the 1st Embodiment of this invention. The flowchart which shows the outline | summary of operation | movement by the wholesale goods market bid price setting assistance system which concerns on 1st Embodiment. The flowchart which shows an example of the selling bid price optimization process in the case of selling transaction as a bid price optimization process which concerns on 1st Embodiment. The flowchart which shows an example of the buying bid price optimization process in the case of buying transaction as a bid price optimization process which concerns on 1st Embodiment. The block diagram which shows the structure of the wholesale goods market bid price setting assistance system which concerns on the 2nd Embodiment of this invention. The flowchart which shows the outline | summary of operation | movement by the wholesale goods market bid price setting assistance system which concerns on 2nd Embodiment.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 ... Wholesale product market bid price setting support system 2 ... Operation part 21 ... Past time series data acquisition part 22 ... Profit calculation data acquisition part 23 ... Time series analysis part 24 ... Bid price optimization part 25 ... Fluctuation model parameter estimation part 26 ... Simulation unit 3 ... Interface unit 31 ... Data input unit 32 ... Data output unit 33 ... Data communication unit 4 ... Storage unit 5 ... Communication medium 6 ... Market transaction support system

Claims (7)

In a wholesale product market bidding pricing support system that uses a computer to support the bid setting for market participants to bid in the spot market for wholesale products,
By analyzing the time series data of the spot price previously executed in the spot market, the time series of the one-term future forecast value of the past spot price, the one-term future forecast value, and the one-year ahead actual value A time series analysis means for generating a standard series of prediction errors or a time series of variances;
Using the time series of the one-term forecast value and the time series of the standard deviation or variance, an optimal bid price that maximizes the expected value of the profit index set in advance for the sell transaction and the buy transaction is calculated. Bid price optimization means,
Storage means for storing calculation conditions and calculation results;
A wholesale product market bidding price setting support system comprising interface means for inputting data and displaying results. The time series analysis means includes
Fluctuation model parameter estimation means for estimating model parameters of a preset spot price fluctuation model by analyzing time series data of spot prices previously executed in the spot market;
Simulating means for generating a time series of the one-term forecast value and a time series of the standard deviation or variance by simulating the spot price temporal fluctuation by the Monte Carlo method using the spot price fluctuation model and the model parameter The wholesale product market bid setting support system according to claim 1, further comprising: In the case of a sell transaction, the expected value of the profit index used in the bid price optimization means is set based on the bid price and the production cost of the market participant of the spot market transaction unit. The wholesale product market according to claim 1 or 2, wherein the wholesale product market is set based on a bid price and a production cost of the market participant of the spot market transaction unit or a retail price of the wholesale product. Bid pricing support system. 4. The wholesale commodity market bid price setting according to claim 1, wherein the time series analysis unit is configured to calculate the standard deviation or variance for each trading day. 5. Support system. In a wholesale product market bid price setting support method for supporting price setting for a market participant to bid in order to conduct sales transactions in a spot market of wholesale products using a computer having a calculation unit, a storage unit, and an interface unit,
By the calculation unit,
By analyzing the time series data of the spot price previously executed in the spot market, the time series of the one-term future forecast value of the past spot price, the one-term future forecast value, and the one-year ahead actual value A time series analysis step to generate a time series of standard deviation or variance of the prediction error;
Using the time series of the one-term forecast value and the time series of the standard deviation or variance, an optimal bid price that maximizes the expected value of the profit index set in advance for the sell transaction and the buy transaction is calculated. Perform a bid optimization step,
The storage unit performs a result storage step of storing the processing result of the arithmetic unit,
A wholesale product market bid price setting support method, wherein the interface unit performs a result display step of displaying a processing result of the arithmetic unit. 6. The wholesale product according to claim 5, wherein the interface unit performs a result transmission step of transmitting the processing result of the arithmetic unit to a means for supporting a transaction at a wholesale product transaction market or an exchange through a communication medium. Market bid setting support method. In a wholesale product market bidding pricing support program that uses a computer to support the price setting for market participants to bid in the spot market for wholesale products,
By analyzing the time series data of the spot price previously executed in the spot market, the time series of the one-term future forecast value of the past spot price, the one-term future forecast value, and the one-year ahead actual value A time series analysis function that generates a time series of standard deviation or variance of the prediction error;
Using the time series of the one-term forecast value and the time series of the standard deviation or variance, an optimal bid price that maximizes the expected value of the profit index set in advance for the sell transaction and the buy transaction is calculated. Bid price optimization,
A result storage function for storing the processing result of the arithmetic unit;
A wholesale product market bidding price setting support program for causing a computer to realize a result output function for outputting a processing result of the arithmetic unit.

Images