JP2019125127A - Demand predicting device, demand predicting method, and program - Google Patents

Abstract

To provide a demand predicting device, a demand predicting method, and a program that predict a demand for a consumable item which is utilized in a product and which can be replaced.SOLUTION: A demand predicting device that predicts a demand for a consumable item which is utilized in a product and which can be replaced includes: a demand prediction model storing unit that stores a demand prediction model on the basis of a first model representing a demand curve for the consumable item, and a second model representing a change rate in time in the number of operating products; a parameter calculating unit which estimates a parameter contained in the demand prediction model stored in the demand prediction model storing unit on the basis of the actual number of sold products and of the actual number of sold consumable items, and which calculates a parameter estimation value; a parameter storing unit that stores the parameter estimation value calculated by the parameter calculating unit; and a prediction value calculating unit that calculates a prediction value for the number of demanded consumable items on the basis of the demand prediction model stored in the demand prediction model storing unit, the parameter estimation value stored in the parameter storing unit, and a planned number of sales products in future.SELECTED DRAWING: Figure 1

Description

  The present invention relates to a demand forecasting device, a demand forecasting method, and a program.

  With respect to a product (hereinafter also referred to as a main body) that uses replaceable consumables, there is a technique for predicting the demand for the consumables. For example, Patent Document 1 below describes a technology for performing demand forecasting of consumables by incorporating the number of actual operation of the main unit into prediction. In addition, it is described in the following patent document 1 that the said main body operation | use number is confirmed when a service engineer visits a customer by maintenance.

Japanese Patent Application Publication No. 2003-263300

  However, it was a heavy human burden for the service engineer to confirm the actual number of operation of the main unit. In particular, if the main body related to the consumables for demand forecasting is a product that exists in many homes including household appliances such as vacuum cleaners and air conditioners, a service engineer visits each home to frequently operate the main body It was difficult to confirm.

  Therefore, the present invention has been made in view of the above problems, and an object of the present invention is to provide a novel and improved method capable of predicting the demand for consumables while reducing the human burden. A demand forecasting apparatus, a demand forecasting method, and a program are provided.

  In order to solve the above problems, according to an aspect of the present invention, there is provided a demand forecasting device for forecasting demand for replaceable consumables used in a product, wherein the first model represents a demand curve of the consumables. And a demand forecasting model storage unit for storing a demand forecasting model based on a second model representing a time change rate of the number of operating products, and the number of sales records of the products and the sales results of the consumables. A parameter calculating unit configured to estimate a parameter included in the demand forecasting model stored in the demand forecasting model storage unit, and calculating a parameter estimated value; and storing the parameter estimated value computed by the parameter computing unit. A parameter storage unit, the demand forecasting model stored in the demand forecast model storage unit, the parameter estimated value stored in the parameter storage unit, and a future Based on product sales plan number, a predicted value calculation unit for calculating a predicted value of demand number of the consumables, the demand prediction apparatus comprising a are provided.

  The predicted value calculation unit inputs the estimated value of the parameter, the number of sales records of the product, the number of sales records of the consumables, and the number of planned product sales plans in the future in the demand prediction model, and sets the predicted values. It may be calculated.

  In the second model, the time rate of change of the number of operating products may be expressed by subtracting a predetermined percentage of the number of operating products from the number of newly shipped products.

  The first model may be a model based on the number of operated products, the newly shipped products, and the simultaneous purchase ratio of the consumables at the time of product purchase.

  The first model and the second model may be expressed by mathematical expressions, and the demand forecasting model may be expressed as a simultaneous equation of the first model and the second model.

The demand forecasting model includes an expected value A (t) of the cumulative shipment number of the product at a predetermined time t, an expected value B (t) of the number of products operated at the predetermined time t, and the predetermined value t The expected value C (t) of the number of shipments of the consumables, the parameter p representing the rate of suspension of use of the product, the parameter q representing the rate of utilization of the consumables, simultaneous purchase of the consumables at the time of purchase of the product Formula (F) which is a simultaneous equation of the first model expressed as formula (F-1) and the second model expressed as formula (F-2) using the parameter r representing a ratio It may be expressed.

  The demand forecasting model may be selected based on the user's operation.

  Further, in order to solve the above problems, according to another aspect of the present invention, there is provided a demand forecasting method for forecasting a demand for replaceable consumables used in a product, the demand forecasting method representing the demand curve of the consumables Storing a demand forecasting model based on a first model and a second model representing a time change rate of the number of operating products of the product, and based on the number of sales results of the product and the number of sales results of the consumables; Estimating a parameter included in the demand prediction model stored in the demand prediction model storage unit, calculating a parameter estimation value, and storing the parameter estimation value calculated by the parameter calculation unit; Based on the demand forecasting model stored in the demand forecasting model storage unit, the estimated parameter value stored in the parameter storage unit, and the number of planned product sales plans in the future, It includes calculating a predicted value of goods demand number, a demand prediction method is provided.

  In order to solve the above problems, according to another aspect of the present invention, there is provided a program for causing a computer to realize a demand forecasting function of replaceable consumables used for a product, the program comprising: A function of storing a demand forecasting model based on a first model representing a demand curve and a second model representing a time change rate of the number of operating products, the number of sales results of the product and the number of sales results of the consumables Based on the estimation of the parameters included in the demand forecasting model stored in the demand forecasting model storage unit, and the function of computing the parameter estimation value, and storing the parameter estimation value computed by the parameter computing unit Function, the demand forecasting model stored in the demand forecasting model storage unit, the parameter estimated value stored in the parameter storage unit, and a future product Based on the planned number of sales, a program for realizing a function for calculating a predicted value of demand number of the consumable is provided.

  As described above, according to the present invention, it is possible to predict the amount of demand for consumables while reducing the human burden.

It is a block diagram showing an example of functional composition of demand forecasting device 1 concerning one embodiment of the present invention. FIG. 16 is a table showing an example of data on the number of main sales results stored in the main sales sales number accumulation unit 102. FIG. FIG. 16 is a table showing an example of data on the number of consumables sales records stored in the consumables sales record number accumulation unit 104. FIG. It is a table which shows an example of data of the main sales plan quantity stored in the main sales plan quantity storage unit 106. It is a flowchart figure which shows the operation | movement outline | summary of the demand prediction apparatus 1 which concerns on this embodiment. It is a flowchart figure which shows the more detailed flow of model construction step S10. It is a table | surface which shows an example of the main body sales performance number in which the tabulation process was performed. It is a table | surface which shows an example of the consumables sales number in which the tabulation process was performed. It is a flowchart figure which shows the more detailed flow of consumables demand prediction step S20. It is a table | surface which shows an example of the data of the main body number couple | bonded. It is a table showing an example of a predicted value of the number of demand for consumables output by the output unit 122. It is an explanatory view showing hardware constitutions of demand forecasting device 1 concerning an embodiment of the present invention.

  The present invention will now be described more fully with reference to the accompanying drawings, in which exemplary embodiments of the invention are shown. In the present specification and the drawings, components having substantially the same functional configuration will be assigned the same reference numerals and redundant description will be omitted.

  Further, in the present specification and the drawings, a plurality of components having substantially the same functional configuration may be distinguished by attaching different alphabets to the same reference numerals. However, when it is not necessary to distinguish each of a plurality of components having substantially the same functional configuration, only the same reference numerals will be given.

<< 1. Background >>
First, the background which led to creation of the demand-forecasting device by one embodiment of the present invention is explained.

  Conventionally, with regard to products (body) that use replaceable consumables, the amount of demand related to the consumables is manually predicted empirically by, for example, past sales results, market trends, number of sales plans of the body, etc. It was Such a manual prediction may have low accuracy because the predicted value is greatly influenced by the skill and experience of the individual in addition to the heavy human burden.

  In addition, there is also a technology for forecasting the demand for consumables as in Patent Document 1 above, but it is necessary for the service engineer to confirm the number of actual operation of the main unit used for demand forecast in this technology at the time of maintenance. Burden was heavy.

  In order to reduce such a human burden, instead of the service engineer directly confirming the main body operating number, it is conceivable to estimate the main body operating number using a mathematical model. For example, it is considered possible to estimate the demand for consumables by estimating the number of operating main bodies using a fixed model based on the average failure interval and Weibull distribution, and using the estimated number of operating main bodies. .

  However, when estimating the number of operating main bodies using such a model, it is not considered about the rate at which utilization of the main body is stopped for reasons other than failure. Therefore, when the number of main units operating is estimated using such a model, it is possible that the estimated number of operating units will deviate from the actual number of units operating over time, and as a result, the demand forecast for consumables There is a risk that the accuracy may decline.

  Therefore, the inventor of the present invention has made the above-mentioned situation one point of view and has come to create an embodiment of the present invention. A demand forecasting apparatus according to an embodiment of the present invention directly uses a demand forecasting model based on a first model that represents a demand curve of consumables and a second model that represents a time change rate of the number of operating main bodies. To clarify the relationship between the number of main units that can not be observed and the number of demand for consumables. According to the present embodiment described below, since the number of operating main units at the time of prediction is estimated based on the past consumption records of consumables, the number of operating units estimated even if time passes It is possible to suppress the difference between the and the actual number of operating units. As described above, according to the present embodiment, it is possible to predict the number of operating main bodies with higher accuracy, and as a result, it is possible to perform the demand forecast of consumables with high accuracy while reducing the human burden. it can.

  Hereinafter, in the present embodiment, a demand forecasting device for realizing the above effects will be described in more detail.

<< 2. Configuration >>
FIG. 1 is a block diagram showing an example of the functional configuration of a demand forecasting device 1 according to an embodiment of the present invention. The demand forecasting apparatus 1 according to the present embodiment is an apparatus for forecasting a demand for a product that uses replaceable consumables, and more specifically, an information processing apparatus for forecasting demand for the consumables. The demand prediction device 1 may calculate a predicted value of the demand for the number of consumables, and the predicted value may be used as, for example, a planned number of sales of the consumable (the number of future sales).

  As shown in FIG. 1, the demand forecasting device 1 according to the present embodiment includes a main sales data storage unit 102, a consumable sales sales number storage unit 104, a main sales plan number storage unit 106, an acquisition unit 108, and a data processing unit. The information processing apparatus 110 includes an input unit 112, a demand prediction model storage unit 114, a parameter calculation unit 116, a parameter storage unit 118, a prediction value calculation unit 120, and an output unit 122.

  The main unit sales record number accumulation unit 102 has data of the number of main units that are products using consumables targeted for demand forecast so far (hereinafter referred to as main unit sales record number) sold so far (in the past) It is accumulated. FIG. 2 is a table showing an example of data on the main sales result number stored in the main sales record number storage unit 102.

  The format of data accumulated in main unit sales record accumulation unit 102 is not limited to the example shown in FIG. For example, the date and time of sale for each main unit may be recorded, or a value obtained by totaling the number of main units sold for each time unit (for example, month, day, etc.) for which prediction is desired to be performed is recorded. It is also good.

  The consumables sales track record storage unit 104 stores data of the number of consumables for which demand forecasting has been performed (hereinafter, also referred to as consumables sales track record) sold so far (in the past). FIG. 3 is a table showing an example of data on the number of consumables sales records stored in the consumable sales sales number storage unit 104.

  Note that the format of data stored in the consumables sales track record storage unit 104 is not limited to the example shown in FIG. For example, as in the example of the main body described above, the date and time of sale sold for each consumable item may be recorded, or the consumable item sold for each time unit (for example, month, day, etc.) for which prediction is desired The value which totaled the number of may be recorded.

  In the main unit sales plan number accumulation unit 106, data of the planned number of main unit sales (hereinafter also referred to as main unit sales plan number) in a future period targeted for demand forecasting by the demand forecasting device 1 is accumulated. FIG. 4 is a table showing an example of data on the planned main-unit sales stored in the planned sales data unit 106. However, the format of the data stored in the planned sales data storage unit 106 is not limited to the example shown in FIG.

  The acquiring unit 108 receives the main unit sales record number, the consumables sales record number, and the main unit sales from each of the main unit sales record unit 102, the consumable sales unit count storage unit 104, and the main unit sales plan unit 106 described above. Data on the planned number is acquired and provided to the data processing unit 110.

  The data processing unit 110 performs predetermined processing on the data acquired by the acquisition unit 108, and provides the data to the parameter calculation unit 116 and the predicted value calculation unit 120. For example, the data processing unit 110 may use time units (e.g., a month or a day) to make a forecast on each data of the main sales record number, the consumable sales sales number, and the main sales plan number acquired by the acquisition section 108. Perform tallying process to total the number of each). In addition, the data processing unit 110 performs a combining process of combining data of the number of sold main units for which the aggregation process has been performed and data of the number of planned main sales for which the aggregation process is performed. The data processing unit 110 provides the parameter calculation unit 116 with data of the number of actual sales of main units subjected to the aggregation processing and data of the numbers of actual sales of consumables subjected to the aggregation processing. Further, the data processing unit 110 provides the predicted value calculation unit 120 with data of the combined number of main bodies and data of the number of consumables sales records subjected to the aggregation processing, which are obtained by the combination processing.

  The input unit 112 inputs a demand forecasting model (for example, a formula) for the demand forecasting device 1 to forecast the demand for consumables. The demand forecasting model input by the input unit 112 is input to the demand forecasting model storage unit 114 and stored in the demand forecasting model storage unit 114. In addition, although the demand forecasting model will be described later in detail, the demand forecasting model according to the present embodiment is a first model that represents a demand curve of consumables and a second model that represents a time change rate of the number of operating main units. It may be a mathematical model based on

  For example, the function of the input unit 112 may be realized by an interface that accepts user operations such as a mouse, a keyboard, a touch panel, and a button, or an input of a demand forecasting model may be performed based on the user's operation. In such a case, for example, one demand prediction model may be selected and input from among a plurality of demand prediction models prepared in advance by the user operation, or the demand prediction model is created and input by the user operation It is also good.

  Further, the input unit 112 may receive data on a demand forecasting model prepared in advance from an external device, or read it from a configuration (not shown) of the demand forecasting apparatus 1 and may input it to the demand forecasting model storage unit 114 .

  The demand prediction model storage unit 114 stores the demand prediction model input by the input unit 112. The function of the demand prediction model storage unit 114 may be realized by, for example, a volatile memory that temporarily stores data, or may be realized by a non-volatile memory.

  The parameter calculation unit 116 estimates an unknown parameter included in the demand prediction model stored in the demand prediction model storage unit 114, and calculates a parameter estimation value. Although the details of the demand forecasting model will be described later, the demand forecasting model stored in the demand forecasting model storage unit 114 includes at least one or more unknown parameters as coefficients or terms, and the parameter calculating unit 116 estimates such unknown parameters.

  The parameter calculation unit 116 may calculate the estimated value of the parameter included in the demand prediction model based on the number of actual sales of main body and the number of actual sales of consumables. As described above, the data processing unit 110 performs data processing for totaling the number of units for each unit of time for which the data processing unit 110 performs prediction, and the parameter calculation unit 116 May be provided.

  In addition, the parameter calculation unit 116 may use, for example, a parameter that reduces the difference between the number of consumables sold calculated by the demand forecasting model based on the number of actual units sold and the actual number of consumables sold. It may be calculated as an estimated value. With such a configuration, the demand forecasting model to which the parameter estimated value is applied becomes more consistent with the number of actual sales of main units and the number of actual sales of consumables.

  The parameter storage unit 118 stores the parameter estimation value calculated by the parameter calculation unit 116. The function of the parameter storage unit 118 may be realized by, for example, a volatile memory that temporarily stores data, or may be realized by a non-volatile memory.

  The forecast value calculation unit 120 includes a demand forecast model stored in the demand forecast model storage unit 114, an estimated parameter value stored in the parameter storage unit 118, a main sales record, a consumable sales sales count, and a main sales forecast number. Based on the forecast value of the number of consumables required is calculated. Note that, as described above, data on the number of actual sales of consumables is provided to the predicted value calculation unit 120 after the data processing unit 110 performs each processing of totaling the number of units for each time unit for which prediction is performed. It is also good. Further, as described above, the data processing unit 110 performs data processing for totaling the number of units for each time unit for performing prediction by the data processing unit 110, and further predicts the data of the actual number of actual units sold and the planned number of main units sales as combined data. It may be provided to the value calculator 120.

  For example, the forecast value calculation unit 120 adds the above-described parameter estimation value, the main sales record number, the consumable sales sales number, and the main sales plan number to the demand prediction model stored in the demand prediction model storage section 114. Calculate the forecast value of the number of consumables required. The prediction value calculation unit 120 may calculate the prediction value of the number of demand for consumables for each time unit (for example, month, day, etc.) for which it is desired to make a preset prediction. Furthermore, the period targeted for prediction by the prediction value calculation unit 120 (the prediction target period) may be the same as the period in which data of the main sales plan number is stored in the main sales plan storage unit 106.

  The output unit 122 outputs the predicted value of the demand for the consumables calculated by the predicted value calculation unit 120. For example, the output unit 122 may be configured to include a monitor, and in such a case, the output unit 122 displays and outputs a predicted value of the number of demand for consumables. Alternatively, the output unit 122 may output the predicted value of the demand number of consumables to the storage in the demand forecasting apparatus 1 or an external device as a file of a file format designated in advance.

<< 3. Operation >>
<3-1. Operation outline >>
Heretofore, the functional configuration example of the demand prediction device 1 according to the present embodiment has been described. Subsequently, an operation example of the demand prediction device 1 according to the present embodiment will be described. FIG. 5 is a flowchart showing an operation outline of the demand prediction device 1 according to the present embodiment.

  As shown in FIG. 5, first, the demand prediction apparatus 1 according to the present embodiment estimates an unknown parameter included in the demand prediction model, and constructs a demand prediction model (S10). In addition, below, step S10 is also called a model construction | assembly step, and the detail of a model construction | assembly step is later mentioned with reference to FIGS. 6-8.

  Subsequently, the demand prediction apparatus 1 according to the present embodiment uses the demand prediction model constructed in step S10 to forecast the demand for consumables (S20). In addition, below, step S20 is also called a consumables demand forecasting step, and the details of the consumables demand forecasting step will be described later with reference to FIGS.

  The process shown in FIG. 5 may be repeated as appropriate. For example, after the demand forecast of consumables is performed by the process shown in FIG. 5, the forecast target period before the forecast target period in the demand forecast ends. The process shown in FIG. 5 may be performed again to forecast the demand for the consumables later. Further, FIG. 5 shows an example in which the consumables demand prediction step S20 is performed once following the model construction step S10, but once the model construction step S10 has been completed, the consumables demand prediction step S20 is completed. It can be done many times. That is, the period during which the model construction step S10 is repeated may be longer than the period during which the consumable item demand prediction step S20 is repeated.

  In the above, the outline of operation of demand forecasting device 1 concerning this embodiment was explained. Subsequently, the model construction step S10 and the consumable item demand prediction step S20 shown in FIG. 5 will be sequentially described in more detail.

<3-2. Model building step >>
FIG. 6 is a flowchart showing a more detailed flow of the model construction step S10. First, the demand forecasting model is input by the input unit 112 (S102). For example, the demand forecasting model may be input by the user operating the interface of the keyboard and selecting one demand forecasting model from among a plurality of demand forecasting models prepared in advance. Alternatively, data on a demand forecasting model prepared in advance based on a predetermined format may be received from an external device or read from a not shown configuration of the demand forecasting device 1 and may be input.

  The demand forecasting model input in step S102 is stored in the demand forecasting model storage unit 114 (S104).

  Subsequently, the acquiring unit 108 acquires data (actual data) on the main sales record number and the consumable sales sales number from each of the main sales record number accumulation portion 102 and the consumable sales sales number storage portion 104 ( S106).

  Subsequently, the data processing unit 110 performs an aggregation process of totaling the number of units for each time unit (for example, month or day) for which it is desired to perform prediction, on each of the actual data acquired in step S106 (S108). FIG. 7 is a table showing an example of the number of sold main bodies on which the tabulation processing has been performed. Further, FIG. 8 is a table showing an example of the number of consumables sales to which the tabulation processing has been performed. The examples shown in FIG. 7 and FIG. 8 are obtained by performing the tabulation process in units of time on a daily basis with respect to the number of actual sales of main body and the number of actual sales of consumables shown in FIG. 2 and FIG. It is It should be noted that the total number of main unit sales record numbers stored in main unit sales record number accumulation unit 102 and consumables sales record number storage unit 104, and the number of consumables sales record units are totaled in units of time for which it is desired to predict in advance. If yes, the process of step S108 may be skipped.

  As shown in FIG. 6, in the subsequent step S110, the parameter calculation unit 116 determines unknown parameters included in the demand forecasting model, based on the number of main unit sales records subjected to the tabulation processing in step S108 and the number of consumables sales records. Calculate the estimated value of

  Here, the details of the demand forecasting model according to the present embodiment will be described. As described above, the demand forecasting model according to the present embodiment is a mathematical model based on the first model representing the demand curve of the consumables and the second model representing the time rate of change of the number of operating main bodies. For example, when the first model and the second model are each represented as an equation, the demand prediction model is a simultaneous equation of the first model and the second model, or a model obtained by modifying the simultaneous equation, It is also good.

  For example, the first model and the second model can be respectively expressed as differential equations, and in such a case, the demand forecasting model is a simultaneous differential equation in which the first model and the second model are simultaneous, or the simultaneous differential equation It may be a model obtained by transforming equation Formula (1-1) and Formula (1-2) shown below are examples of the first model and the second model expressed as differential equations, respectively, and the demand forecasting model is the formula (1) -1) and Formula (1-2) can be expressed as Formula (1) which is a simultaneous differential equation.

  In the above equation (1), A (t) is an expected value of the cumulative number of shipments of the main body at a predetermined time t, and B (t) is an expected value of the number of operating main bodies at time t. (T) is an expected value of the number of shipments of consumables at time t. In addition to the idea that the time rate of change of the expected value C (t) of the number of shipments of consumables increases linearly according to the number of operating units of the main unit, the above equation (1) It is a demand forecasting model based on the idea that it is expressed by the sum of the increase due to shipment of and the steady decay.

  Moreover, in said Formula (1), p, q, and r are unknown parameters. p represents the rate of use suspension of the main body, and 0 <p <1. Also, q represents the use rate of consumables, and q> 0. Also, r represents the ratio of simultaneous purchase of consumables at the time of main unit purchase, and r> 0.

  As in Equation (1-1), the first model representing the demand curve of consumables is a model in accordance with the idea that the number of consumables in demand is based on the current number of main units in operation. Furthermore, in the first model representing the demand curve of consumables as in equation (1-1), consumables purchased simultaneously with the newly shipped main body may be considered. In equation (1-1), the temporal change rate dC (t) / dt of the expected value C (t) of the number of shipments of consumables is the expected value B (t) of the usage rate q of the consumables and the number of operating main bodies. Expressed by the sum of the first term, which is the product of, the simultaneous purchase ratio r of consumables at the time of purchase of the main unit, and the second term, which is the product of the number of main units dA (t) / dt shipped newly. It is done.

  The second model, equation (1-2), is an equation for estimating the expected value B (t) of the number of active units in the unobserved body. In equation (1-2), the temporal change rate dB (t) / dt of the expected value B (t) of the operating main body is calculated based on the number of main bodies dA (t) / dt newly shipped. It is estimated by subtracting a predetermined ratio of In Equation (1-2), the predetermined ratio of the number of operating main units is expressed as the product of the ratio p of usage suspension of the main unit and the expected value B (t) of the number of operating main units.

  Further, the demand forecasting model is a formula (2-1), which is another example of the first model and the second model, and the formula as shown in the following formula (2), instead of the formula (1) above. (2-2) may be expressed as a simultaneous differential equation.

  In the above equation (2), D (t) is the number of main bodies having left at time t. Further, therefore, A (t) -D (t) in the above equation (2) means an expected value of the number of operating main bodies, and corresponds to B (t) in the above equation (1). Therefore, in the above equation (2), as with the above equation (1), the time change rate of the expected value C (t) of the number of shipments of consumables increases linearly in accordance with the number of operating main bodies. It is an expression that represents another mechanism of the same mechanism as equation (1).

  In step S110 shown in FIG. 6, the parameter calculation unit 116 calculates unknown parameters p, q, and r in, for example, equation (1) or equation (2). An example of a method of calculating the unknown parameters p, q, and r by the parameter calculation unit 116 will be described below. In addition, although the case where a demand-forecasting model is represented like Numerical formula (1) is demonstrated to an example below, even if a demand-forecasting model is represented like Numerical formula (2), it can calculate similarly is there.

  First, the equation (3) below is obtained by solving the equation (1) for the expected value B (t) of the number of operating main bodies.

  Further, the equation (4) below can be obtained by solving the equation (1) for the expected value C (t) of the number of shipments of consumables.

  Here, non-stationary Poisson process is assumed for events such as the shipment of consumables and the shipment of the main body, and it is also assumed that the expected value A (t) of the cumulative total number of main bodies is distributed according to the Bass model. Since the equation (1) is assumed to be the intensity function, the likelihood function is as the following equation (5).

Here, ΔA (t) = A (t + 1) −A (t), θ is a parameter of A (t), and η is a parameter of B (t) and C (t). Here, A (t) does not depend on the parameters of B (t) and C (t). Therefore, first determine the parameter of A (t) from the shipment number x _t of the main body in the first stage, and then give the parameter of A (t) in the second stage, from the shipment number y _{t of} consumables A two-step method can be adopted in which the parameters of (t) and C (t) are determined.

  The estimation equation to be maximized in the first stage is expressed as the following equation (6).

  Also, the estimation equation to be maximized in the second stage is expressed as the following equation (7).

  The parameter calculation unit 116 uses, for example, an optimization method such as the Newton method to calculate stepwise parameter estimation values that maximize the right sides of the above equations (6) and (7). Here, maximizing means to make the value as large as possible, and may include, for example, local maximizing using an optimization method.

  In addition, the parameter calculation unit 116 may calculate the parameter estimated value by sequentially maximizing the above Equations (6) and (7) one by one, or by repeatedly maximizing the parameter estimated value. It may be calculated. In the case of repeated maximization, for example, it may be repeated a predetermined number of times, or may be repeated until the difference due to the repetition falls below a predetermined threshold.

  Returning to FIG. 6, the description will be continued. As described above, the parameter calculated by the parameter calculation unit 116 in step S110 is stored in the parameter storage unit 118 (S112).

<3-3. Consumables Demand Forecast Step>
FIG. 9 is a flowchart showing a more detailed flow of the consumables demand prediction step S20. First, the acquiring unit 108 includes the main sales record number, the consumable sales sales number, and the main body sales from each of the main sales sales number storage 102, the consumable sales sales number storage 104, and the main sales plan number storage 106. Data on the planned number is acquired (S202).

  Subsequently, the data processing unit 110 performs, for each of the data acquired in step S202, an aggregation process of totaling the number of units for each time unit (for example, month or day) for which prediction is desired (S204).

  Subsequently, the data processing unit 110 combines the data of the main sales target number for which the aggregation processing has been performed in step S204 and the data of the main sales target number for which the aggregation processing has been performed (S206). FIG. 10 is a table showing an example of the combined number of main body data.

  Subsequently, the forecast value calculation unit 120 calculates a forecast value of the demand for the consumables based on the demand forecast model, the parameter estimation value, the actual sales volume of the main unit, the actual sales volume of the consumables, and the planned number of main sales. S208).

  In step S208, the predicted value calculation unit 120 reads from the demand prediction model storage unit 114 the demand prediction model stored in the demand prediction model storage unit 114 in step S104 of the above-described model construction step S10. Further, in step S208, the predicted value calculation unit 120 reads from the parameter storage unit 118 the parameter estimated value calculated in step S110 of the above-described model construction step S10 and stored in the parameter storage unit 118 in step S112. Further, in step S208, the predicted value calculation unit 120 generates, from the data processing unit 110, data processed for the main sales target number, the consumable sales sales number, and the main sales target number in steps S204 and S206. It may be provided.

  In step S208, in the demand forecast model obtained as described above, the forecast value calculation unit 120 estimates the parameter estimates obtained as described above, the main sales figures, the consumable sales figures, and the main sales plans. To calculate the forecasted value of the demand for consumables.

  Subsequently, the predicted value of the calculated number of demand for consumables is output by the output unit 122 (S210). FIG. 11 is a table showing an example of predicted values of the demand for the number of consumables output by the output unit 122.

<< 4. Modified example >>
Heretofore, an embodiment of the present invention has been described. In the following, some variations of the above embodiment will be described. In addition, each modification explained below may be applied to the above-mentioned embodiment independently, and may be applied to the above-mentioned embodiment by combination. Further, each modification may be applied instead of the configuration described in the above embodiment, or may be additionally applied to the configuration described in the above embodiment.

<4-1. Modification 1>
Although Formula (1) and Formula (2) were demonstrated as an example of a demand forecasting model in the said embodiment, this technique is not limited to the example which concerns. An example of another demand forecasting model is described below.

  For example, the demand forecasting model may be expressed as Formula (8) below by removing the second term on the right side of Formula (1-1) representing the first model in Formula (1).

  Similarly, in the equation (2), the demand forecasting model can be expressed as the following equation (9) by removing the second term on the right side of the equation (2-1) representing the first model. Good.

  Moreover, in the above, although the demand-forecasting model was represented by simultaneous differential equation, this technique is not limited to the example which concerns. For example, the demand forecasting model is a simultaneous difference obtained by simultaneously establishing Equation (10-1) and Equation (10-2), which are other examples of the first model and the second model, as Equation (10) below. It may be expressed by an equation.

  The above equation (10) is another expression of the same mechanism as the equation (1) and the equation (2).

  In addition, the demand forecasting model is not limited to the example described above. For example, a formula obtained by modifying a simultaneous equation in which the first model and the second model described above are combined as a demand forecast model is also considered to be a demand forecast model based on the first model and the second model. obtain. A mathematical expression transformed in advance to a format suitable for calculating the parameter estimated value and the predicted value of the demand for the consumables may be input as a demand forecast model and stored in the demand forecast model storage unit 114. Moreover, although the demand forecasting model showed the example represented as numerical formula above, this technique is not limited to the example which concerns, It may be represented by the other format which the demand forecasting apparatus 1 can process.

<4-2. Modification 2>
Although the above-mentioned embodiment explained composition of demand forecasting device 1 with reference to Drawing 1, composition of an embodiment concerning this art is not limited to the example concerned.

  For example, it is desirable that the data stored in the main sales record storage unit 102, the consumable sales sales count storage unit 104, and the main sales plan number storage unit 106 described above be updated as needed, and the demand forecasting device 1 It may be configured to enable such data updating.

  For example, the demand prediction apparatus 1 includes a communication unit (not illustrated in FIG. 1), and the communication unit includes new data on the main sales record number, the consumable sales sales number, and the main sales plan number from an external server not shown. It may be received. Then, new data received by the communication unit may be stored in main unit sales record storage unit 102, consumable sales record number storage unit 104, and main unit sales plan number storage unit 106.

  Alternatively, the user may be able to input new data regarding the number of actual sales of main units, the number of actual sales of consumables, and the number of planned sales of main units using the input unit 112 shown in FIG. Then, new data input by the user using the input unit 112 may be stored in the main unit sales record storage unit 102, the consumable sales record number storage unit 104, and the main unit sales plan number storage unit 106.

  Although FIG. 1 shows an example in which the demand forecasting device 1 includes the main sales record storage unit 102, the consumable sales sales number storage unit 104, and the main sales plan storage unit 106, the present invention is not limited to this example. For example, the main unit sales record storage unit 102, the consumables sales record storage unit 104, and the main unit sales plan storage unit 106 may be provided in an external server. Also, in such a case, the acquisition unit 108 acquires (receives) data from the main sales record storage unit 102, the consumable sales sales count storage unit 104, and the main sales plan storage unit 106 provided in the external server. May be

  Further, the configuration shown in FIG. 1 may be distributed and provided in a plurality of devices. For example, a model construction apparatus provided with a configuration for executing the model construction step S10 described with reference to FIG. 6 and a configuration for executing the consumables demand prediction step S20 described with reference to FIG. 9 are provided. An information processing system including the consumables demand forecasting device can also be realized.

  In such a case, the model construction apparatus is, for example, the main sales record storage unit 102, the consumable sales sales count storage unit 104, the acquisition unit 108, the data processing unit 110, the input unit 112, and the demand prediction model storage unit 114 illustrated in FIG. The function corresponding to the parameter calculation unit 116 and the parameter storage unit 118 may be provided. Also, in such a case, the consumables demand forecasting device is, for example, the main sales data storage unit 102, the consumable sales sales number storage unit 104, the main sales plan number storage unit 106, the acquisition unit 108, and the data processing unit shown in FIG. A function corresponding to the demand prediction model storage unit 114, the parameter storage unit 118, the prediction value calculation unit 120, and the output unit 122 may be provided. Then, the demand forecasting model and the parameter estimation value may be provided (sent) from the model construction device to the consumables demand forecasting device.

<< 5. Hardware configuration >>
The embodiments of the present invention have been described above. The information processing such as the aggregation processing, the coupling processing, the parameter calculation processing, and the prediction value calculation processing described above is realized by the cooperation of the software and the hardware of the demand prediction device 1. Below, the example of hardware constitutions of demand forecasting device 1 which is an information processor concerning an embodiment of the present invention is explained.

  FIG. 12 is an explanatory view showing a hardware configuration of the demand forecasting device 1 according to the embodiment of the present invention. As illustrated in FIG. 12, the demand prediction apparatus 1 includes a central processing unit (CPU) 1001, a read only memory (ROM) 1002, a random access memory (RAM) 1003, an input device 1004, and an output device 1005. , Storage apparatus 1006, and communication apparatus 1007.

  The CPU 1001 functions as an arithmetic processing unit and a control unit, and controls the overall operation in the demand prediction device 1 according to various programs. The CPU 1001 may be a microprocessor. The ROM 1002 stores programs used by the CPU 1001, calculation parameters, and the like. The RAM 1003 temporarily stores programs used in the execution of the CPU 1001, parameters that appropriately change in the execution, and the like. These are mutually connected by a host bus configured of a CPU bus and the like. The functions of the data processing unit 110, the demand prediction model storage unit 114, the parameter calculation unit 116, the parameter storage unit 118, the prediction value calculation unit 120, etc. are realized mainly by the cooperation of the CPU 1001, the ROM 1002 and the RAM 1003 and the software. Be done.

  The input device 1004 generates an input signal based on an input by a user such as a mouse, a keyboard, a touch panel, a button, a microphone, a switch, and a lever, and an input control circuit that generates an input signal based on the input by the user. And so on. The user of the demand forecasting device 1 can input various data to the demand forecasting device 1 and instruct processing operations by operating the input device 1004. The input device 1004 corresponds to the input unit 112.

  The output device 1005 includes, for example, a display device such as a liquid crystal display (LCD) device, an OLED device and a lamp. Furthermore, the output device 1005 includes an audio output device such as a speaker and headphones. For example, the display device displays a captured image, a generated image, and the like. On the other hand, the audio output device converts audio data and the like into audio and outputs the audio. The output device 1005 corresponds to the output unit 122.

  The storage device 1006 is a device for storing data. The storage device 1006 may include a storage medium, a recording device that records data in the storage medium, a reading device that reads data from the storage medium, and a deletion device that deletes data recorded in the storage medium. The storage device 1006 stores programs executed by the CPU 1001 and various data. The storage device 1006 corresponds to the main unit sales track record accumulation unit 102, the consumables sales record count storage unit 104, and the main unit sales plan count storage unit 106.

  The communication device 1007 is, for example, a communication interface configured of a communication device or the like for connecting to a communication network. Also, the communication device 1007 may include a wireless local area network (LAN) compatible communication device, a long term evolution (LTE) compatible communication device, a wire communication device performing wired communication, or a Bluetooth (registered trademark) communication device.

<< 5. End >>
As described above, according to the embodiment of the present invention, it is possible to predict the amount of demand for consumables with high accuracy while reducing the human burden.

  Although the preferred embodiments of the present invention have been described in detail with reference to the accompanying drawings, the present invention is not limited to such examples. It is obvious that those skilled in the art to which the present invention belongs can conceive of various changes or modifications within the scope of the technical idea described in the claims. Of course, it is understood that these also fall within the technical scope of the present invention.

  For example, the steps in the above embodiment do not necessarily have to be processed chronologically in the order described as the flowchart diagram. For example, each step in the process of the above embodiment may be processed in an order different from the order described as the flowchart diagram, or may be processed in parallel.

  Further, according to the above embodiment, it is possible to provide a computer program for causing hardware such as the CPU 1001, the ROM 1002, and the RAM 1003 to exhibit the same function as each configuration of the demand prediction device 1 described above. Also provided is a recording medium having the computer program recorded thereon.

DESCRIPTION OF SYMBOLS 1 Demand forecasting device 102 Main unit sales record number storage unit 104 Consumables sales record number storage unit 106 Main unit sales plan number storage unit 108 Acquisition unit 110 Data processing unit 112 Input unit 114 Demand prediction model storage unit 116 Parameter calculation unit 118 Parameter storage unit 120 Predicted Value Calculation Unit 122 Output Unit

Claims (9)

A demand forecasting device for forecasting demand for replaceable consumables used in products, comprising:
A demand forecasting model storage unit storing a demand forecasting model based on a first model representing a demand curve of the consumables and a second model representing a time change rate of the number of operating products;
A parameter calculation unit that estimates a parameter included in the demand prediction model stored in the demand prediction model storage unit based on the number of sales results of the product and the number of sales results of the consumables, and calculates a parameter estimation value When,
A parameter storage unit that stores the parameter estimated value calculated by the parameter calculation unit;
Based on the demand forecasting model stored in the demand forecasting model storage unit, the estimated parameter value stored in the parameter storage unit, and the number of future product sales plans, the forecasted value of the demand for the consumables is calculated A predicted value calculation unit to
A demand forecasting device comprising:   The predicted value calculation unit inputs the estimated value of the parameter, the number of sales records of the product, the number of sales records of the consumables, and the number of planned product sales plans in the future in the demand prediction model, and sets the predicted values. The demand forecasting device according to claim 1 which calculates.   In the second model, the time change rate of the number of operating products is represented by subtracting a predetermined ratio of the number of operating products from the number of the newly shipped products. Or the demand forecasting device according to 2.   The first model is a model based on the number of products operated, the number of products newly shipped, and the ratio of simultaneous purchase of the consumables at the time of product purchase. The demand forecasting device according to claim 1.   The said 1st model and the said 2nd model are represented by numerical formula, The said demand-forecasting model is expressed as a simultaneous equation of the said 1st model and the said 2nd model, Any one of Claims 1-4 The demand forecasting device described in. The demand forecasting model includes an expected value A (t) of the cumulative shipment number of the product at a predetermined time t, an expected value B (t) of the number of products operated at the predetermined time t, and the predetermined value t The expected value C (t) of the number of shipments of the consumables, the parameter p representing the rate of suspension of use of the product, the parameter q representing the rate of utilization of the consumables, simultaneous purchase of the consumables at the time of purchase of the product Formula (F) which is a simultaneous equation of the first model expressed as formula (F-1) and the second model expressed as formula (F-2) using the parameter r representing a ratio The demand forecasting device according to claim 5, which is represented.

  The demand forecasting device according to any one of claims 1 to 6, wherein the demand forecasting model is selected based on a user's operation. A demand forecasting method for forecasting demand for replaceable consumables used in a product, comprising:
Storing a demand forecasting model based on a first model representing a demand curve of the consumables and a second model representing a time change rate of the number of operating products.
Estimating a parameter included in the stored demand forecast model based on the number of sales results of the product and the number of sales results of the consumable item to calculate a parameter estimation value;
Storing the calculated parameter estimates;
Calculating a forecast value of the demand for the consumable item based on the stored demand forecasting model, the stored parameter estimation value, and the number of future product sales plans;
Demand forecasting methods, including: A program for causing a computer to realize a demand forecasting function of replaceable consumables used for a product,
A function of storing a demand forecasting model based on a first model representing a demand curve of the consumable item and a second model representing a time change rate of the number of operating products.
A function of estimating a parameter included in the stored demand forecasting model based on the number of sales results of the product and the number of sales results of the consumable to calculate a parameter estimation value;
A function of storing the calculated parameter estimated value;
A function of calculating a predicted value of the demand for the consumables based on the stored demand forecasting model, the stored estimated value of the parameter, and the number of planned product sales in the future;
Program for realizing

Images