SYSTEM AND METHOD FOR SUPPORTING PURCHASE OR PRODUCTION OF PRODUCTS BY POTENTIAL DEMAND PREDICTION

Abstract

To support a purchase or a production of a product by accurately predicting a sold amount of the product. A system that supports a purchase or a production of a product, the system including an input section for accepting an input of a history of a supplied amount and a sold amount of the products, a function generating section for representing a conditional probability function showing probability distribution of a sold amount when the sold amount is restricted by the supplied amount by means of a potential demand probability function including a parameter showing probability distribution of the sold amount when it is supposed that the sold amount is not restricted by the supplied amount and computing a value of the parameter maximizing a value of a likelihood function of the conditional probability function using the input history as a sample to generate the potential demand probability function, and a supplied amount computing section for computing a supplied amount of the product maximizing a profit by a sale of the product, based on the generated potential demand probability function and a predetermined selling price and supplying price of the product, and outputting the amount as a quantity of the product to be purchased or produced.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a view showing a functional construction of a support system 10 according to an embodiment of the present invention.

FIG. 2 is a view showing an example of a data structure of a History of Sale DB 20 according to the embodiment of the present invention.

FIG. 3 is a flowchart of processing for computing a potential demand probability function by the support system 10 according to the embodiment of the present invention to obtain a supplied amount.

FIG. 4 is a view showing an example of a potential demand probability distribution and a conditional probability distribution.

FIG. 5 is a view showing a first result of an experiment for computing a potential demand probability distribution.

FIG. 6 is a view showing a second result of an experiment for computing a potential demand probability distribution.

FIG. 7 is a view showing a functional construction of a supplied amount computing section 330 according to a first modification example.

FIG. 8 is a flowchart of processing for computing a supplied amount and a selling price by the supplied amount computing section 330 according to the first modification example.

FIG. 9 is a view showing an example of a distribution form in which a second modification example should be applied.

FIG. 10 is a view showing a functional construction of a supplied amount computing section 330 according to the second modification example.

FIG. 11 is a flowchart of processing for computing a supplied amount by the supplied amount computing section 330 according to the second embodiment example.

FIG. 12 is a conceptual view of processing for computing a supplied amount for each seller in a third modification example.

FIG. 13 is a view showing an example of a hardware construction of an information processing apparatus 400 functioning as a main body 30 of the support system 10 in the embodiment or any of the modification examples stated above.

Claims

1. A system that supports at least one of a purchase and a production of a product, comprising: an input means for accepting an input of a history of a supplied amount of the product and a history of a sold amount that is a quantity of the product sold from the supplied amount of the product to form an input history;a function generating means (i) for representing a conditional probability function showing probability distribution of the sold amount when the sold amount is restricted by the supplied amount by means of a potential demand probability function including a parameter showing probability distribution of the sold amount when the sold amount is not restricted by the supplied amount and (ii) for computing a value of the parameter maximizing a value of a likelihood function of the conditional probability function using the input history as a sample, in order to generate the potential demand probability function; anda supplied amount computing means for computing a supplied amount of the product maximizing a profit by a sale of the product, and outputting the computed supplied amount as a quantity of the product to either be purchased or produced, based on the generated potential demand probability function and a predetermined selling price and supplying price of the product.

2. The system according to claim 1, wherein the supplied amount computing means computes the supplied amount of the product that maximizes a profit by a sale of the product, including a cost for discarding some of the product in dead stock.

3. The system according to claim 1, wherein the function generating means represents the conditional probability function as an equation using the potential demand probability function, substitutes each probability of generating sold amounts in the history under restriction of a supplied amount in the history for each of the conditional probability functions included in the likelihood function, and replaces each of the probabilities of generating the sold amount in the history under restriction of the supplied amount in the history, which has been substituted for the conditional probability function, by a probability equation using the potential demand probability function in order to numerically compute a value of the parameter maximizing a value of the likelihood function, relative to the substituted supplied amount and sold amount in the history by a computer.

4. The system according to claim 1, wherein the function generating means: represents the conditional probability function as a function for computing a same probability value as that of the potential demand probability function in terms of a sold amount less than a supplied amount and computing a total probability that a sold amount becomes equal to or more than a supplied amount among probabilities to be computed by the potential demand probability function in terms of a sold amount being equal to a supplied amount; andcomputes the value of the parameter maximizing a value of the likelihood function using the inputted history as a sample in order to generate the potential demand probability function.

5. The system according to claim 4, further comprising a selecting means for selecting a set of a supplied amount and a sold amount in case that the sold amount is less than the supplied amount by at least a difference value M among sets of sold amounts ni and supplied amounts mi(0≦i≦I) included in the inputted history, wherein the function generating means: represents the conditional probability function P′ in an equation

6. The system according to claim 1, wherein the function generating means: represents the conditional probability function, relative to a sold amount and a supplied amount that are metric variables, as a function computing a same probability density as the potential demand probability function in terms of a sold amount less than a supplied amount, and represents as an integration value obtained by integrating the probability density computed by the potential demand probability function in case that a sold amount becomes equal to or more than a supplied amount, in terms of a sold amount equal to a supplied amount; andcomputes a value of the parameter maximizing a value of the likelihood function of the conditional probability function relative to the inputted history.

7. The system according to the claim 6, further comprising a selecting means for selecting, among the inputted histories, a set of a sold amount and a supplied amount in case that the sold amount is less than the supplied amount by at least a predetermined difference value, wherein the function generating means: represents the conditional probability function, by means of the potential demand probability function, as an equation for computing a probability density of a contingent probability under a condition that a sold amount is less than a supplied amount by at least the difference value; andcomputes a value of the parameter maximizing the value of the likelihood function of the conditional probability function relative to the selected history.

8. The system according to claim 1, wherein: the input means further accepts an input of a history of a selling price of the product, in association with the history of the sold amount and a supplied amount;the function generating means represents the conditional probability function depending on a selling price as a potential demand probability function including a parameter depending on a selling price, and computing a value of the parameter maximizing a value of the likelihood function of the conditional probability function relative to the history of the sold amount and the supplied amount; andthe supplied amount computing means includes a first computing means for computing a supplied amount of the product maximizing a profit by a sale of the product, based on the generated potential demand probability function as well as a preliminarily given selling price and the supplying price; a second computing means for computing the selling price maximizing a profit by a sale of the product to provide the same to the first computing means, based on the generated potential demand probability function, the supplied amount and the supplying price computed by the first computing means; and an output means for causing the first computing means and the second computing means to alternately compute the selling price and the supplied amount, and under the condition that the selling price and the supplied amount have converged to a predetermined range, outputting the converged selling price and the supplied amount.

9. The system according to claim 1, wherein: the input means accepts, for each of a plurality of sellers, an input of a history of a supplied amount of the product to the seller and a history of a sold amount of the product in the seller;the function generating means, for each of the plurality of sellers, generates the potential demand probability function of the seller as an upward convex function by proposing a maximum likelihood of the parameter relative to the conditional probability function of the seller; andthe supplied amount computing means includes:a determining means for determining whether or not a total value of the supplied amount maximizing the potential demand probability function generated for each seller is within a specified range predetermined as a range of a value that a total of supplied amounts can take;a changing means for computing a expected profit being decreased when a supplied amount to each of the sellers is changed in order to bring the total value of the supplied amounts close to the specified range on condition that the total value of the supplied amounts is outside the specific range, and selecting a seller who has the smallest expected profit to be decreased, in order to decrease a supplied amount to the seller; andan output means for causing the changing means to further change a supplied amount on condition that the total value of the supplied amounts is outside the specified range regardless of the decrease in the sold amount by the changing means, and outputting each supplied amount including the changed supplied amount to each of the sellers on condition that the total value of supplied amounts has entered within the specified range by the change of the supplied amount by the changing means.

10. A method that supports accumulation of a product, the method comprising the steps of: accepting an input of a history of a supplied amount of a product and a history of a sold amount that is a quantity of the product sold from the supplied amount of the product to form an input history;representing a conditional probability function showing probability distribution of the sold amount in case that the sold amount is restricted by the supplied amount by means of a potential demand probability function including a parameter showing probability distribution of the sold amount when the sold amount is not restricted by the supplied amount and computing a value of the parameter maximizing a value of a likelihood function of the conditional probability function using the input history as a sample, in order to generate the potential demand probability function; andcomputing a supplied amount of the product maximizing a profit by a sale of the product, based on the generated potential demand probability function as well as a predetermined selling price and supplying price of the product, and outputting the computed supplied amount as a quantity of the product to be accumulated.

11. A program product stored in a storage medium that causes an information processing apparatus to function as a system supporting an accumulation of a product, the program product causing the information processing apparatus to function as: an input means for accepting an input of a history of a supplied amount of the product and a history of a sold amount that is a quantity of the product sold from the supplied amount of the product to form an input history;a function generating means (i) for representing a conditional probability function showing probability distribution of a sold amount in case that the sold amount is restricted by the supplied amount by means of a potential demand probability function including a parameter showing probability distribution of the sold amount when the sold amount is not restricted by the supplied amount and (ii) computing a value of the parameter maximizing a value of a likelihood function of the conditional probability function using the input history as a sample in order to generate the potential demand probability function; anda supplied amount computing means for computing a supplied amount of the product maximizing a profit by a sale of the product, based on the generated potential demand probability function as well as predetermined selling price and supplying price of the product, and outputting the computed supplied amount as a quantity of the product to be accumulated.

12. A method for providing a service that supports accumulating a product, the method comprising the steps of: accepting a supplied history of a supplied amount of the product and a sold history of a sold amount that is a quantity of the product sold from the supplied amount of the product, and inputting the supplied history and the sold history into a system to form an input history;representing a conditional probability function showing probability distribution of the sold amount in case that the sold amount is restricted by the supplied amount by means of a potential demand probability function including a parameter showing probability distribution of the sold amount when the sold amount is not restricted by the supplied amount and computing a value of the parameter maximizing a value of a likelihood function of the conditional probability function using the input history as a sample in order to generate the potential demand probability function; andcomputing a supplied amount of the product maximizing a profit by a sale of the product, based on the generated potential demand probability function as well as a predetermined selling price and supplying price of the product, and causing the system to output the computed supplied amount as a quantity of the product to be accumulated.