Assistance device, assistance method, and assistance program

Abstract

To make it easy to determine a constraint condition for optimization that uses an objective function, an assistance device (1) comprises a variable extraction unit (11) that extracts a variable that corresponds to a feature quantity for the objective function from training data used to generate the objective function, a reference information generation unit (12) that generates reference information for determining a constraint condition for an optimization calculation that uses the objective function on the basis of the extracted variable, and an information presentation unit (13) that presents the reference information.

Description

Technical Field

[0001] The present invention relates to a support device and the like that support the determination of constraint conditions in optimization calculations using an objective function.

Background Art

[0002] Optimization calculations using an objective function are used in various technical fields. For example, Patent Document 1 below discloses a technique related to optimal control for operating in an optimal state by optimization using an objective function in industrial processes such as petroleum refining processes and petrochemical processes.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] Optimization calculations using an objective function, including the technique disclosed in Patent Document 1, are performed under given constraint conditions. Therefore, an operator who performs optimization using an objective function needs to set constraint conditions. However, since knowledge about the objective function to be used is required to set appropriate constraint conditions, there have been many operators who feel difficulty in determining the constraint conditions.

[0005] One aspect of the present invention has been made in view of the above problems, and an example of the object is to provide a technique that facilitates the determination of constraint conditions in optimization using an objective function.

Means for Solving the Problems

[0006] The support device according to one aspect of the present invention includes a variable extraction unit that extracts a variable corresponding to a feature amount of the objective function from learning data used for generating the objective function, a reference information generation unit that generates reference information for determining a constraint condition in an optimization calculation using the objective function based on the extracted variable, and an information presentation unit that presents the reference information.

[0007] The support method according to one aspect of the present invention includes at least one processor extracting a variable corresponding to a feature amount of the objective function from learning data used for generating the objective function, generating reference information for determining a constraint condition in an optimization calculation using the objective function based on the extracted variable, and presenting the reference information.

[0008] The support program according to one aspect of the present invention causes a computer to function as a variable extraction unit that extracts a variable corresponding to a feature amount of the objective function from learning data used for generating the objective function, a reference information generation unit that generates reference information for determining a constraint condition in an optimization calculation using the objective function based on the extracted variable, and an information presentation unit that presents the reference information.

Advantages of the Invention

[0009] According to one aspect of the present invention, it is possible to facilitate the determination of a constraint condition in optimization using an objective function.

Brief Description of the Drawings

[0010]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

[0011] [Exemplary Embodiment 1] The first exemplary embodiment of the present invention will be described in detail with reference to the drawings. This exemplary embodiment is a basic form of the exemplary embodiments described later.

[0012] (Configuration of the Support Device) The configuration of the support device 1 according to this exemplary embodiment will be described with reference to FIG. 1. FIG. 1 is a block diagram showing the configuration of the support device 1 according to this exemplary embodiment. As shown in the figure, the support device 1 includes a variable extraction unit 11, a reference information generation unit 12, and an information presentation unit 13.

[0013] The variable extraction unit 11 extracts variables corresponding to the feature amounts of the objective function from the learning data used for the generation of the objective function. Further, the reference information generation unit 12 generates reference information for determining constraint conditions in the optimization calculation using the objective function based on the variables extracted by the variable extraction unit 11. Then, the information presentation unit 13 presents the reference information generated by the reference information generation unit 12.

[0014] As described above, in the support device 1 according to the present exemplary embodiment, a variable extraction unit 11 that extracts a variable corresponding to a feature amount of the objective function from learning data used for generating the objective function, and a reference information generation unit 12 that generates reference information for determining a constraint condition in an optimization calculation using the objective function based on the variable extracted by the variable extraction unit 11, and an information presentation unit 13 that presents the reference information generated by the reference information generation unit 12 are adopted. Therefore, according to the support device 1 according to the present exemplary embodiment, an effect that the determination of the constraint condition in the optimization using the objective function can be facilitated is obtained.

[0015] (Support program) The functions of the above-described support device 1 can also be realized by a program. The support program according to the present exemplary embodiment causes a computer to function as a variable extraction unit 11 that extracts a variable corresponding to a feature amount of the objective function from learning data used for generating the objective function, a reference information generation unit 12 that generates reference information for determining a constraint condition in an optimization calculation using the objective function based on the variable extracted by the variable extraction unit 11, and an information presentation unit 13 that presents the reference information generated by the reference information generation unit 12. According to this support program, an effect that the determination of the constraint condition in the optimization using the objective function can be facilitated is obtained.

[0016] (Flow of the support method) The flow of the support method according to the present exemplary embodiment will be described with reference to FIG. 2. FIG. 2 is a flowchart showing the flow of the support method according to the present exemplary embodiment. Note that the execution subject of each step in this Support method may be a processor included in the support device 1, or may be a processor included in another device, or may be processors provided in different devices for each step.

[0017] In S11, at least one processor extracts variables corresponding to the feature amounts of the objective function from the learning data used for generating the objective function. Next, in S12, at least one processor generates reference information for determining the constraint conditions in the optimization calculation using the objective function based on the variables extracted in S11. Then, in S13, at least one processor presents the reference information generated in S12.

[0018] As described above, in the support method according to this exemplary embodiment, at least one processor extracts variables corresponding to the feature amounts of the objective function from the learning data used for generating the objective function, generates reference information for determining the constraint conditions in the optimization calculation using the objective function based on the extracted variables, and presents the generated reference information. Such a configuration is adopted. Therefore, according to the support method according to this exemplary embodiment, the effect of being able to facilitate the determination of the constraint conditions in the optimization using the objective function can be obtained.

[0019] 〔Exemplary Embodiment 2〕 (Overview) Based on FIG. 3, the overview of the support device 2 according to Exemplary Embodiment 2 will be described. FIG. 3 is a diagram for explaining the overview of the support device 2 according to Exemplary Embodiment 2 of the present invention. The support device 2 is a device that supports the optimization calculation using the objective function, and more specifically, is a device that facilitates the determination of the constraint conditions used in the optimization calculation.

[0020] Note that although FIG. 3 shows an example in which the support device 2 is a desktop personal computer, this is merely an example. The support device 2 can also be realized by a general-purpose computer such as a personal computer or a smartphone, or can be realized as a device mainly having the function of supporting the determination of the constraint conditions, and can be a stationary device or a portable device.

[0021] The support device 2 extracts variables corresponding to the feature amounts of the objective function from the learning data used for the generation of the objective function. Note that when the support device 2 uses the variables included in the learning data as the feature amounts of the objective function as they are, the support device 2 extracts those variables, and when calculating the feature amounts of the objective function from one or more variables included in the learning data, the support device 2 extracts one or more variables used for the calculation of the feature amounts.

[0022] The learning data shown in FIG. 3 is data in which state data and action data are associated with each other, and the action data indicates the action to be taken in the state indicated by the state data. The learning data may be generated based on past actions (actions that can be said to be appropriate actions) and the states at the times when those actions were taken. In this case, it can also be said that the learning data is history data indicating the history of decision-making regarding the action.

[0023] Here, it is assumed that the objective function is generated by inverse reinforcement learning. Therefore, it is sufficient to prepare the number of learning data required for inverse reinforcement learning. Since multiple learning data are usually used in inverse reinforcement learning, the support device 2 extracts variables from each learning data. Note that the objective function is not limited to being generated by inverse reinforcement learning, and any function generated using learning data may be used. For example, the objective function may be generated by a known machine learning method such as density ratio estimation.

[0024] Various types of data can be applied as the action data. For example, when creating a schedule using the above objective function, data indicating a schedule suitable for the state indicated by the state data may be used as the action data. For example, when creating an employee work schedule, data indicating the workers for each time period as shown in FIG. 3 may be used as the action data.

[0025] Further, for example, when performing resource allocation or matching using the above objective function, data indicating the allocation or combination of resources that conforms to the state indicated by the state data may be used as the action data. Similarly, when performing sequencing, optimal path search, or automatic control using the above objective function, data indicating the sequence, path, or control content that conforms to the state indicated by the state data may be used as the action data.

[0026] Also, the state data only needs to indicate the state when the action indicated by the action data is performed, and the state may be represented by a plurality of variables (St1, St2, …) as shown in the figure. For example, when creating an employee work schedule, variables indicating the cumulative overtime hours of each employee or variables indicating the maximum working hours per day may be used as the state data. The support device 2 extracts such variables from the learning data. Note that the state data may directly become a feature amount of the objective function, or the feature amount of the objective function may be calculated using the variables.

[0027] Next, the support device 2 generates reference information for determining the constraint conditions in the optimization calculation using the objective function based on the extracted variables. Then, the support device 2 presents the generated reference information to an operator (the person who determines the constraint conditions) who operates the support device 2. The reference information can be generated based on the extracted variables and only needs to be a reference for determining the constraint conditions. Also, the mode of presenting the reference information is not particularly limited.

[0028] In the example of FIG. 3, the support device 2 causes the display device 5 to display, as the reference information, the statistical amounts calculated for the variables included in the state data extracted from the learning data. If the statistical amounts of the variables extracted from the learning data are known, it becomes easier to judge appropriate constraint conditions, so according to the support device 2, the determination of the constraint conditions can be facilitated. In the example of FIG. 3, the operator determines the constraint condition that variable 1 is less than or equal to the threshold value th1 (variable 1 ≦ th1) with reference to the statistical amount of variable 1.

[0029] By performing an optimization calculation using the constraint conditions determined in this way and the objective function generated using the learning data, it is possible to estimate the actions to be taken in any state. For example, when the action data included in the learning data indicates the work schedule of an employee, the optimization calculation may be performed using the state data indicating each state that affects the work schedule as input data. Thereby, it is possible to estimate the schedule to be adopted in the said state.

[0030] The form of the objective function is not particularly limited. For example, when generating an objective function in inverse reinforcement learning, a mathematical formula that multiplies and adds weight values to each feature quantity may be used as the objective function. In the example of FIG. 3, the weight values of variables 1 to 3, which are feature quantities, are α to γ. It can be said that each feature quantity indicates a factor that affects the action or a perspective that is emphasized in the decision-making of the action. Also, the weight value multiplied by each feature quantity indicates how much that factor or perspective is emphasized, and is learned using the learning data.

[0031] The optimization calculation may be performed by an optimization solver. As the optimization solver, for example, general application programs such as IBM ILOG CPLEX, Gurobi Optimizer, and S CIP can be used. Note that the support device 2 may also perform either or both of the generation of the objective function and the optimization calculation.

[0032] (Configuration of the support device) Based on FIG. 4, the configuration of the support device 2 according to this exemplary embodiment will be described. FIG. 4 is a block diagram showing the configuration of the support device 2. As shown in the figure, the support device 2 includes a control unit 20 that comprehensively controls each part of the support device 2, and a storage unit 21 that stores various data used by the support device 2. Further, the support device 2 includes a communication unit 22 for the support device 2 to communicate with other devices, an input unit 23 that receives input of various data to the support device 2, and an output unit 24 for the support device 2 to output various data. Further, the control unit 20 includes a data acquisition unit 201, a reception unit 202, a variable extraction unit 203, a reference information generation unit 204, and an information presentation unit 205.

[0033] The data acquisition unit 201 acquires learning data used for generating the objective function. The method of acquiring the learning data is not particularly limited. For example, the data acquisition unit 201 may acquire the learning data input via the input unit 23, or may acquire the learning data from another device via the communication unit 22.

[0034] The reception unit 202 receives the designation of variables used for generating the reference information. For example, the reception unit 202 may receive the designation of variables by displaying a screen as shown in FIG. 5. FIG. 5 is a diagram showing an example of a reception screen for designating variables. In the illustrated reception screen, variable names and check boxes are displayed in association with each other. The operator can check the check boxes via the input unit 23, and the reception unit 202 uses the variables corresponding to the checked check boxes as the variables used for generating the reference information.

[0035] The variable extraction unit 203 extracts variables corresponding to the feature amounts of the objective function from the learning data acquired by the data acquisition unit 201. For example, as in the example of FIG. 3, when learning data in which state data and action data are associated is acquired, the variable extraction unit 203 may extract each variable included in the state data. Note that when the variables included in the learning data are directly used as the feature amounts of the objective function, the variable extraction unit 203 may extract such variables. Further, when the feature amounts of the objective function are calculated from one or more variables included in the learning data, the variable extraction unit 203 may extract one or more variables used for the calculation of the feature amounts.

[0036] The reference information generation unit 204 generates reference information for determining constraint conditions in the optimization calculation using the objective function based on the variables extracted by the variable extraction unit 203. Note that when the reception unit 202 has received a designation of variables to be used for the generation of the reference information, the reference information generation unit 204 generates the reference information based on the variables designated to be used for the generation of the reference information among the variables extracted by the variable extraction unit 203.

[0037] As described above, the support device 2 includes the reception unit 202 that receives the designation of variables, and the reference information generation unit 204 generates the reference information based on the designated variables. Therefore, according to the support device 2 according to the present exemplary embodiment, in addition to the effects exhibited by the support device 1 according to the first exemplary embodiment, an effect can be obtained that reference information generated based on the variables can be presented to an operator who has designated variables of interest.

[0038] As described above, the reference information may be, for example, a statistic of the extracted variables. Specifically, the reference information may be at least any one of the average value, median value, maximum value, minimum value, mode value, variance, deviation, and standard deviation of the extracted variables. For example, when the average value of "St1" included in the state data in the example of FIG. 3 is used as the reference information, the reference information generation unit 204 may use, as the reference information, a value obtained by dividing the sum of the values of "St1" extracted by the variable extraction unit 203 from each learning data by the total number of the extracted "St1".

[0039] The information presentation unit 205 presents the reference information generated by the reference information generation unit 204 to the operator. The presentation mode is not particularly limited. For example, the information presentation unit 205 may display the reference information on a display device, output it as audio through an audio output device, or output it as a print through a printing device. Further, the device for outputting the reference information may be included in the support device 2 or may be a device external to the support device 2. For example, the information presentation unit 205 may output the reference information to the output unit 24, or may output it to a display device 5 external to the support device 2 as in the example of FIG. 3.

[0040] When the reference information is a statistic, the information presentation unit 205 may present a numerical value indicating the statistic or may present a graph indicating the statistic. FIG. 6 is a diagram showing an example of a screen in which the statistic calculated as the reference information is represented by a bar graph. For example, by representing the variance of each variable in such a bar graph, it becomes possible to easily grasp a variable with a large variance, and it is also possible to set the constraint conditions for determining the range of such a variable more broadly.

[0041] In addition to the reference information, the information presentation unit 205 may present a weight value corresponding to each variable. The weight value is determined by learning using learning data. Since the weight value corresponding to each variable indicates how much importance is attached to that variable, it is useful as a reference when determining what constraint conditions to set for each variable.

[0042] (Use of mathematical expressions) Further, the reference information generation unit 204 may calculate the statistic of the value obtained by substituting the variable extracted by the variable extraction unit 203 into a predetermined mathematical expression instead of the statistic of the variable. For example, in general optimization calculations, the following may be used as constraint conditions for the variable x. Note that a to c are constants. ax + b > c...(1) ax + b < c...(2) ax + b = c...(3) Therefore, the reference information generation unit 204 may substitute the variable x extracted by the variable extraction unit 203 into a mathematical formula such as "ax + b" and calculate the statistic of the value as reference information. Thereby, reference information that is useful when creating the above-described constraint conditional expression can be presented to the operator.

[0043] Also, when calculating the feature amount of the objective function from one or more variables included in the learning data, the statistic of the feature amount may be calculated. In this case, the variable extraction unit 203 extracts one or more variables used for calculating the feature amount, and the reference information generation unit 204 calculates the feature amount from each learning data using the variable and calculates the statistic thereof.

[0044] Also, for example, when formulating a production plan for a plurality of items by the objective function, the variable extraction unit 203 may extract a variable indicating the production quantity of each item. In this case, the reference information generation unit 204 may calculate the value obtained by summing the production quantities of each item, that is, the statistic for the total production quantity of all items, and use this as the reference information. In this case, if the number of items is n (n is an integer of 3 or more), the predetermined mathematical formula is "production quantity of item 1 + production quantity of item 2 +... + production quantity of item n". Thereby, reference information that is useful when using the total production quantity of all items as a constraint condition can be presented to the operator.

[0045] As described above, the reference information generation unit 204 may calculate the statistic of the variable extracted by the variable extraction unit 203 or the value obtained by substituting the variable into a predetermined mathematical formula, and use the calculated statistic as reference information.

[0046] If the statistic of the variable extracted from the learning data is known, it becomes easier to determine a reasonable constraint condition. The same applies to the statistic of the value obtained by substituting the variable extracted from the learning data into a predetermined mathematical formula. Therefore, according to the above-described configuration in which the statistic of the variable extracted from the learning data or the value obtained by substituting the variable into a predetermined mathematical formula is used as reference information and presented, the determination of the constraint condition can be facilitated.

[0047] (Switching of learning data to be extracted) The variable extraction unit 203 may classify the learning data acquired by the data acquisition unit 201 into a plurality of categories, and extract variables from the learning data of each category. In this case, the reference information generation unit 204 generates reference information corresponding to each category based on the variables extracted from the learning data of each category, and the information presentation unit 205 presents those reference information for each category.

[0048] As a result, in addition to the effect exhibited by the support device 1 according to the exemplary embodiment 1, an effect can be obtained that useful information can be provided when an operator determines which category of learning data to use for generating an objective function. Note that the above reference information only needs to be a reference for determining constraint conditions, and for example, it may be the above-described statistic, or may be distribution information described in the exemplary embodiment 3 described later.

[0049] For example, when the learning data is data related to the work schedule of an employee in the past year, the variable extraction unit 203 may classify the learning data by category for each predetermined period. For example, when the variable extraction unit 203 classifies the learning data monthly, the reference information generation unit 204 calculates a statistic (for example, an average value of working hours) corresponding to each month based on the variables extracted from the learning data for each month, and the information presentation unit 205 presents the calculated statistics monthly. Note that the information presentation unit 205 may present the monthly statistics to the operator by, for example, displaying them on one screen, or only display the statistics for some months on one screen and allow the operator to switch the target months to be displayed. As a result, the operator can, for example, avoid using the learning data of months with significantly different statistic values compared to other months for generating the objective function. And thereby, it becomes possible to generate a general-purpose objective function using average learning data.

[0050] (Flow of processing) The flow of the process (support method) executed by the support device 2 will be described with reference to FIG. 7. FIG. 7 is a flowchart of the support method according to the exemplary embodiment 2.

[0051] In S21, the data acquisition unit 201 acquires learning data used for generating the objective function. Subsequently, in S22, the variable extraction unit 203 extracts variables from the learning data acquired in S21. The variables extracted in S22 are variables corresponding to the feature amounts of the objective function generated using the learning data acquired in S21.

[0052] In S23, the reception unit 202 receives the designation of variables. For example, the reception unit 202 may present the variable names of the respective variables extracted in S22 to the operator by displaying a reception screen as shown in FIG. 5, and receive the designation of variables by the operator. Thereby, the variables used for generating the reference information are narrowed down to those designated by the operator.

[0053] Note that the narrowing down of the variables used for generating the reference information may be performed by the variable extraction unit 203. For example, the variable extraction unit 203 may acquire the weight values of the objective function corresponding to the extracted respective variables, and use, as the variables for generating the reference information, the variables whose weight values are equal to or greater than a predetermined threshold value. Thereby, it is possible to generate the reference information by narrowing down to the variables having a great influence in the optimization calculation. Further, for example, the variable extraction unit 203 may use, as the variables for generating the reference information, the variables included in the constraint conditions set in the past, the variables having a high usage frequency in the constraint conditions set in the past, or variables of the same type as those variables. For example, if information indicating the variables used in the determined constraint conditions and their types is accumulated, the variable extraction unit 203 can narrow down the variables using the information. Further, the reception unit 202 may cause the operator to designate, from among the variables narrowed down in this way, the variables used for generating the reference information.

[0054] In S24, the reference information generation unit 204 calculates the statistic of the variables designated in S23. As described above, the statistic is reference information for determining the constraint conditions in the optimization calculation using the objective function generated using the learning data acquired in S21. Then, in S25, the information presentation unit 205 presents the statistic calculated in S25 to the operator, and thereby the process of FIG. 7 ends.

[0055] 〔Exemplary Embodiment 3〕 (Configuration of Support Device) Based on FIG. 8, the configuration of the support device 3 according to this exemplary embodiment will be described. FIG. 8 is a block diagram showing the configuration of the support device 3. As shown in the figure, the support device 3 includes a control unit 30 that comprehensively controls each part of the support device 3, and a storage unit 31 that stores various data used by the support device 3. Further, the support device 3 includes a communication unit 32 for the support device 3 to communicate with other devices, an input unit 33 that receives various data inputs to the support device 3, and an output unit 34 for the support device 3 to output various data. Further, the control unit 30 includes a data acquisition unit 301, a reception unit 302, a variable extraction unit 303, a reference information generation unit 304, an information presentation unit 305, and a determination unit 306. Note that the data acquisition unit 301 and the variable extraction unit 303 have the same configuration as the data acquisition unit 201 and the variable extraction unit 203 in Exemplary Embodiment 2.

[0056] The reference information generation unit 304 generates distribution information indicating the distribution of the variables extracted by the variable extraction unit 303. The distribution information is reference information for determining the constraint conditions in the optimization calculation using the objective function generated using the learning data acquired by the data acquisition unit 301. That is, the reference information generation unit 304 generates the distribution information as reference information. Note that also in the generation of the distribution information, as described in the item of "Use of Mathematical Formulas", a value obtained by substituting the variable into a predetermined mathematical formula may be used instead of the extracted variable.

[0057] If the distribution of the variables extracted from the learning data or the distribution of the values obtained by substituting the variables into a predetermined mathematical formula is known, it becomes easier to judge appropriate constraint conditions. Therefore, according to the above configuration of generating and presenting the information indicating the distribution of the variables as reference information, the determination of the constraint conditions can be facilitated.

[0058] The reception unit 302 receives the specification of the threshold value. The reception of the specification of the threshold value will be described in the item of "Example of Distribution Information Display Screen" below. Also, similar to the reception unit 202 of the exemplary embodiment 2, the reception unit 302 may also receive the specification of the variables used for generating the reference information.

[0059] The determination unit 306 determines whether or not the constraint conditions are satisfied based on the threshold value received by the reception unit 302 and the variables extracted by the variable extraction unit 303. The details of this determination will be described in the item of "Example of Distribution Information Display Screen" below.

[0060] The information presentation unit 305 presents the reference information generated by the reference information generation unit 304 to the operator. Also, the information presentation unit 305 presents the relevant information related to the variables determined by the determination unit 306 not to satisfy the constraint conditions to the operator. The relevant information will be described in the item of "Example of Distribution Information Display Screen" below.

[0061] (Example of Distribution Information Display Screen) When the presentation mode of the distribution information is display output, the information presentation unit 305 may display, for example, a display screen as shown in FIG. 9. FIG. 9 is a diagram showing an example of the distribution information display screen. FIG. 9 shows two display screen examples, A1 and A2. Both the display screen examples A1 and A2 display the distribution information representing the distribution of variables as a scatter diagram. Of course, the distribution information may be any information indicating the distribution of variables and is not limited to a scatter diagram. For example, a frequency distribution diagram of variables or the like may be used as the distribution information, or numerical information such as a combination of the average value and variance value of variables may be used as the distribution information.

[0062] When the distribution information is a scatter diagram, the information presentation unit 305 selects the variable to be the vertical axis and the variable to be the horizontal axis of the scatter diagram from among the variables extracted by the variable extraction unit 303. Note that the variable to be the vertical axis and the variable to be the horizontal axis may be specified by the operator. In this case, the reception unit 302 may receive the specification of the variables by displaying a screen listing the selectable variables as shown in FIG. 5, for example.

[0063] The information presentation unit 305 that has selected the variable for the vertical axis and the variable for the horizontal axis draws the coordinate plane defined by the selected two axes, and draws points corresponding to the values of the variables on the coordinate plane. As a result, a scatter diagram as shown in display screen examples A1 and A2 is drawn.

[0064] For example, when the variable for the vertical axis is the cumulative working hours in a predetermined period and the variable for the horizontal axis is the cumulative number of vacation days acquired in a predetermined period, the horizontal axis x in FIG. 9 m is the cumulative number of vacation days acquired, and the vertical axis x n is the cumulative working hours, and the information presentation unit 305 draws the coordinate plane. Then, the information presentation unit 305 plots points corresponding to the cumulative working hours and the cumulative number of vacation days acquired by each worker on the coordinate plane. In this case, each point on the coordinate plane indicates the cumulative working hours and the cumulative number of vacation days acquired by each worker.

[0065] Also, the display screen examples A1 and A2 include a slider bar SB1 and a slider SL. The slider bar SB1 and the slider SL are objects for accepting the specification of a threshold value. The threshold value is a threshold value used as a constraint condition for the variable extracted by the variable extraction unit 303. For example, the threshold value may define at least either the upper limit or the lower limit of the variable, or may define at least either the upper limit or the lower limit of the value calculated using the variable.

[0066] More specifically, the slider bar SB1 indicates the range of the threshold value that can be specified, and the position of the slider SL on the slider bar SB1 indicates the threshold value. Then, the operator can operate the slider SL with the cursor CU1 to change the position of the slider SL on the slider bar SB1, and thereby can specify the threshold value. Then, the reception unit 302 identifies the threshold value specified from the position of the slider SL on the slider bar SB1.

[0067] When the reception unit 302 receives the specification of the threshold value in this way, the determination unit 306 determines whether or not the constraint conditions are satisfied based on the threshold value received by the reception unit 302 and the variable extracted by the variable extraction unit 303. Then, the information presentation unit 305 makes the presentation mode of the variable in the scatter diagram correspond to the result of the determination.

[0068] For example, when setting a constraint condition that the overtime hours of each worker are equal to or less than the threshold value, the information presentation unit 305 may display the slider bar SB1 and the slider SL for receiving the specification of the overtime hour threshold value. Further, the reception unit 302 identifies the threshold value specified from the position of the slider SL on the slider bar SB1, and the determination unit 306 determines whether or not the overtime hours are equal to or less than the specified threshold value for each point plotted in the scatter diagram. Then, the information presentation unit 305 makes the presentation modes of the points where the overtime hours are equal to or less than the specified threshold value, that is, the points satisfying the constraint conditions, and the points exceeding the threshold value, that is, the points not satisfying the constraint conditions, in the scatter diagram different.

[0069] For example, the information presentation unit 305 may be able to identify whether or not the constraint conditions are satisfied by the color of the points as in the display screen examples A1 and A2 of FIG. 9. For example, in the display screen example A1, the points P1 and P2 have different colors, which indicates that either P1 or P2 satisfies the constraint conditions and the other does not. Here, it is assumed that the point P1 does not satisfy the constraint conditions and the point P2 satisfies the constraint conditions.

[0070] Note that how to display the points satisfying the constraint conditions and how to display the points not satisfying the constraint conditions may be determined in advance. Further, the information presentation unit 305 may indicate whether or not the constraint conditions are satisfied by, in addition to the color, the saturation, brightness, gradation value, filling pattern, shape of the points (such as circles and squares), or the size of the points.

[0071] As described above, in the display screen example A1, the threshold value can be changed by the slider SL. The display screen example A2 is an example of the screen when the slider SL is moved to the right (the direction in which the threshold value is increased) from the state of the display screen example A1. By increasing the threshold value, in the display screen example A2, the number of points that satisfy the constraint condition, that is, points with the same display mode as the point P2, has increased.

[0072] In this way, the reception unit 302 receives the specification of the threshold value in the constraint condition, and the determination unit 306 determines whether or not the constraint condition is satisfied based on the received threshold value and the extracted variable. Then, the information presentation unit 305 makes the presentation mode of the variable in the scatter diagram, which is the distribution information, according to the determination result of the determination unit 306. Thereby, according to the support device 3, in addition to the effect exhibited by the support device 1 according to the exemplary embodiment 1, the operator can easily recognize which variable satisfies the constraint condition when any threshold value is set, and the effect that the determination of the constraint condition including the threshold value can be facilitated is obtained. The same applies when presenting the distribution information of the values obtained by substituting the extracted variables into a predetermined mathematical formula.

[0073] Also, in the display screen example A2, the message M1 is displayed in association with the point P1. The message M1 corresponds to the worker named "N" for the point P1, and indicates that the overtime of the worker exceeds the specified threshold value. In this way, the information presentation unit 305 may present the relevant information related to the variable determined not to satisfy the constraint condition to the operator. Also, the information presentation unit 305 may present the relevant information in the same manner when presenting the distribution information of the values obtained by substituting the extracted variables into a predetermined mathematical formula.

[0074] If it is known what variables are determined not to satisfy the constraint condition, it becomes easier to judge appropriate constraint conditions. Therefore, according to the above-described configuration for presenting the relevant information related to the variable determined not to satisfy the constraint condition, the effect that the constraint condition can be determined with reference to the relevant information is obtained.

[0075] Note that the related information only needs to be information related to variables determined not to satisfy the constraints, and is not limited to the example in FIG. 9. For example, the information presentation unit 305 may present, as related information, the date when the overtime hours exceeded the threshold, how much the overtime hours exceeded the threshold, or the work system and job type of Mr. A.

[0076] (Processing flow) The flow of the process (support method) executed by the support device 3 will be described based on FIG. 10. FIG. 10 is a flowchart of the support method according to the exemplary embodiment 3.

[0077] In S31, the data acquisition unit 301 acquires learning data used for generating the objective function. Subsequently, in S32, the variable extraction unit 303 extracts variables corresponding to the feature amounts of the objective function from the learning data acquired in S31. Then, in S33, the reception unit 202 receives a designation of a variable. These processes are the same as the processes of S21 to S23 in FIG. 7. Note that when generating a scatter diagram as distribution information, in S33, the reception unit 202 may also receive a designation regarding which variables are used for each axis of the scatter diagram.

[0078] In S34, the reference information generation unit 304 generates distribution information indicating the distribution of the variables designated in S33. As described above, the distribution information is reference information for determining the constraints in the optimization calculation using the objective function generated using the learning data acquired in S31, and may be, for example, a scatter diagram as shown in FIG. 9.

[0079] In S35, the determination unit 306 determines whether or not the variables extracted in S32 satisfy the constraints. For example, when the threshold of the constraints can be specified by the slider bar SB1 shown in FIG. 9, the determination unit 306 may determine that variables below the threshold indicated by the slider bar SB1 satisfy the constraints, and determine that variables having values exceeding the threshold do not satisfy the constraints.

[0080] In S36, the information presentation unit 305 presents to the operator a presentation mode of the variables in the distribution information generated in S34 with the determination result in S35 reflected therein. Further, the information presentation unit 305 generates related information according to the determination result in S35 and presents it to the operator. Note that the related information may be presented, for example, triggered by an operation of the operator or the like.

[0081] In S37, the reception unit 202 determines whether the threshold value has been changed. For example, the reception unit 202 determines that the threshold value has been changed when the slider bar SB1 shown in FIG. 9 is operated. If it is determined as YES in S37, the process returns to S35, and the determination unit 306 makes the determination described above based on the changed threshold value. On the other hand, if it is determined as NO in S37, the process proceeds to S38.

[0082] In S38, the reception unit 202 determines whether to end the presentation of the reference information (here, the distribution information). The conditions for ending the presentation of the reference information may be determined in advance. For example, the reception unit 202 may determine to end the presentation when it receives an operation of the operator to end the presentation of the reference information. If it is determined as NO in S38, the process returns to S37. On the other hand, if it is determined as YES in S38, the process in FIG. 10 ends.

[0083] [Modification Example] The execution entity of each process described in the above exemplary embodiment is arbitrary and is not limited to the above example. For example, a support system having the same functions as the support devices 1 to 3 can be constructed by a plurality of devices capable of communicating with each other. For example, by dispersing and providing each block shown in FIGS. 1, 4, and 8 to a plurality of devices, a support system having the same functions as the support devices 1 to 3 can be constructed. Further, for example, each process in the flowcharts shown in FIGS. 2, 7, or 10 can also be executed by being shared among a plurality of information processing devices (or processors).

[0084] [Example of Realization by Software] Some or all of the functions of the support devices 1 to 3 may be realized by hardware such as an integrated circuit (IC chip), or may be realized by software.

[0085] In the latter case, the support devices 1 to 3 are realized, for example, by a computer that executes instructions of a program, which is software for realizing each function. An example of such a computer (hereinafter referred to as computer C) is shown in FIG. 11. Computer C includes at least one processor C1 and at least one memory C2. A program (support program) P for operating computer C as any one of the support devices 1 to 3 is recorded in memory C2. In computer C, processor C1 reads and executes program P from memory C2, thereby realizing each function of any one of the support devices 1 to 3.

[0086] As the processor C1, for example, a CPU (Central Processing Unit), GPU (Graphic Processing Unit), DSP (Digital Signal Processor), MPU (Micro Processing Unit), FPU (Floating point number Processing Unit), PPU (Physics Processing Unit), TPU (Tensor Processing Unit), quantum processor, microcontroller, or a combination thereof can be used. As the memory C2, for example, a flash memory, HDD (Hard Disk Drive), SSD (Solid State Drive), or a combination thereof can be used.

[0087] Incidentally, the computer C may further include a RAM (Random Access Memory) for expanding the program P during execution or temporarily storing various data. The computer C may further include a communication interface for transmitting and receiving data to and from other devices. The computer C may further include an input / output interface for connecting input / output devices such as a keyboard, a mouse, a display, and a printer.

[0088] Also, the program P can be recorded on a non-transitory tangible recording medium M readable by the computer C. As such a recording medium M, for example, a tape, a disk, a card, a semiconductor memory, or a programmable logic circuit can be used. The computer C can acquire the program P via such a recording medium M. Also, the program P can be transmitted via a transmission medium. As such a transmission medium, for example, a communication network or a broadcast wave can be used. The computer C can also acquire the program P via such a transmission medium.

[0089] 〔Supplementary Note 1〕 The present invention is not limited to the above-described embodiments, and various modifications are possible within the scope shown in the claims. For example, embodiments obtained by appropriately combining the technical means disclosed in the above-described embodiments are also included in the technical scope of the present invention.

[0090] 〔Supplementary Note 2〕 Some or all of the above-described embodiments can also be described as follows. However, the present invention is not limited to the aspects described below.

[0091] (Supplementary Note 1) A support device comprising: variable extraction means for extracting, from learning data used for generating an objective function, a variable corresponding to a feature amount of the objective function; reference information generation means for generating reference information for determining constraint conditions in an optimization calculation using the objective function based on the extracted variable; and information presentation means for presenting the reference information.

[0092] (Appendix 2) The support device according to Appendix 1, wherein the reference information generation means calculates a statistic of the extracted variable or a value obtained by substituting the variable into a predetermined mathematical formula, and uses the calculated statistic as the reference information.

[0093] (Appendix 3) The support device according to Appendix 1, wherein the reference information generation means generates, as the reference information, distribution information indicating a distribution of the extracted variable or a value obtained by substituting the variable into a predetermined mathematical formula.

[0094] (Appendix 4) A reception means for receiving a specification of a threshold value in the constraint conditions, and a determination means for determining whether or not the extracted variable satisfies the constraint conditions based on the received threshold value and the extracted variable. The information presentation means presents a presentation mode of the variable in the distribution information or a value obtained by substituting the variable into a predetermined mathematical formula according to the result of the determination. The support device according to Appendix 3.

[0095] (Appendix 5) The support device according to Appendix 4, wherein the information presentation means presents related information related to the variable determined not to satisfy the constraint conditions or a value obtained by substituting the variable into a predetermined mathematical formula.

[0096] (Appendix 6) The support device according to any one of Appendices 1 to 5, further comprising reception means for receiving a specification of the variable, and wherein the reference information generation means generates the reference information based on the specified variable.

[0097] (Appendix 7) The reference information generation means generates the reference information corresponding to each category based on the variables extracted from the learning data classified into a plurality of categories, and the information presentation means presents the reference information for each category. The support device according to any one of Appendices 1 to 6.

[0098] (Appendix 8) A support method including: one or more processors extracting variables corresponding to feature amounts of the objective function from learning data used for generation of the objective function; generating reference information for determining constraint conditions in an optimization calculation using the objective function based on the extracted variables; and presenting the reference information.

[0099] (Appendix 9) A support program that causes a computer to function as a variable extraction means for extracting variables corresponding to feature amounts of an objective function from learning data used for generation of the objective function, a reference information generation means for generating reference information for determining constraint conditions in an optimization calculation using the objective function based on the extracted variables, and an information presentation means for presenting the reference information.

[0100] [Appendix Item 3] Some or all of the above-described embodiments can also be expressed as follows. A support device including at least one processor, the processor performing a process of extracting variables corresponding to feature amounts of an objective function from learning data used for generation of the objective function, a process of generating reference information for determining constraint conditions in an optimization calculation using the objective function based on the extracted variables, and a process of presenting the reference information.

[0101] Note that this support device may further include a memory, and the memory may store a support program for causing the processor to execute the process of extracting the variables, the process of generating the reference information, and the process of presenting the reference information. Further, this support program may be recorded on a non-transitory tangible recording medium readable by a computer.

Explanation of Symbols

[0102] 1 Support Device 11 Variable Extraction Unit 12 Reference Information Generation Unit 13 Information Presentation Unit 2 Support Device 202 Reception Unit 203 Variable Extraction Unit 204 Reference Information Generation Unit 205 Information Presentation Unit 3 Support Device 302 Reception Unit 303 Variable Extraction Unit 304 Reference Information Generation Unit 305 Information Presentation Unit 306 Judgment Unit

Claims

1. A variable extraction means for extracting variables corresponding to feature quantities of an objective function from learning data used for generating the objective function; a reference information generating means for generating reference information for determining constraint conditions in an optimization calculation using the objective function based on the extracted variables; and an information presenting means for presenting the reference information.

2. The support device according to claim 1 , wherein the reference information generating means calculates statistics of the extracted variables or values ​​obtained by substituting the variables into a predetermined formula, and sets the calculated statistics as the reference information.

3. The support device according to claim 1 , wherein the reference information generating means generates, as the reference information, distribution information indicating a distribution of the extracted variables or a distribution of values ​​obtained by substituting the extracted variables into a predetermined mathematical expression.

4. A receiving means for receiving a threshold value designated in the constraint condition; a determination means for determining whether or not the constraint condition is satisfied based on the received threshold value and the extracted variable, The support device according to claim 3 , wherein the information presenting means presents the variables in the distribution information or values ​​obtained by substituting the variables into a predetermined formula in a manner that corresponds to a result of the determination.

5. The support device according to claim 4 , wherein the information presenting means presents related information relating to the variable determined not to satisfy the constraint condition or a value obtained by substituting the variable into a predetermined mathematical expression.

6. A reception means for receiving the designation of the variable, The support device according to claim 1 , wherein the reference information generating means generates the reference information based on the designated variable.

7. the reference information generating means generates the reference information corresponding to each of the plurality of categories based on the variables extracted from the learning data classified into the plurality of categories; The support device according to claim 1 , wherein the information presenting means presents the reference information for each of the categories.

8. At least one processor Extracting variables corresponding to feature quantities of the objective function from learning data used to generate the objective function; generating reference information for determining constraint conditions in an optimization calculation using the objective function based on the extracted variables; and presenting the reference information.

9. Computer, a variable extraction means for extracting variables corresponding to features of the objective function from the learning data used for generating the objective function; a reference information generating means for generating reference information for determining constraint conditions in an optimization calculation using the objective function based on the extracted variables; and and a support program that functions as an information presentation means for presenting the reference information.