Logic model construction support device, logic model construction support method, and logic model construction support system

Abstract

The present invention addresses the problem of supporting construction of a logic model that can ensure validity. This logic model construction support device displays a directed graph representing a logic model to be used for evaluating a social impact index for evaluating an activity of an organization and having a plurality of hierarchies represented by edges connecting nodes to each other, receives a selection of a node in the displayed directed graph, displays details of the selected node and edges connected to the node, changes the weighting coefficients corresponding to the displayed edges according to an operation of a user, creates a first quantitative value of the social impact index of the logic model before the weighting coefficients are changed, a second quantitative value of the social impact index of the logic model after the weighting coefficients are changed, and variation information relating to the difference between the first quantitative value and the second quantitative value, and displays the first quantitative value, the second quantitative value, and the variation information.

Description

Logic model construction support device, logic model construction support method, and logic model construction support system

[0001] The present invention relates to a technology for supporting a predetermined activity such as an organizational activity, and to a technology for supporting the construction of a logic model that uses a social impact index to evaluate the activity.

[0002] In recent years, it has become important for various organizations, such as local governments, companies, and financial institutions, to evaluate the impact of their policies, operations, investments, and other activities on society. For example, the Sustainable Development Goals (SDGs) have been established to address global social issues. Organizations, including companies and local governments, are placing importance on efforts to achieve these goals in their business operations, management, and activities. The SDGs call for achieving sustainable development in a balanced and integrated manner across the three dimensions of society, the environment, and the economy. Therefore, the activities of various organizations, including corporate businesses, must also consider the impact and effects on these three dimensions. To do this, it is necessary to predict the impact of business operations and other activities along the value axes of society, the environment, and the economy, and to carry out activities to improve value.

[0003] To achieve these activities, environmental and economic indicators—such as carbon dioxide emissions, which belong to environmental value, and corporate profits, which belong to economic value—can be objectively quantified through carbon dioxide measurements and corporate activity reports, and this is already being implemented. Meanwhile, social value includes objective indicators such as inequality and fairness, but also subjective aspects such as the continuity of local culture through regional development, the quality of life of local residents, and their sense of happiness. Furthermore, subjective indicators must take into account changes over time. For example, even if the subjective attachment to a landscape decreases with the construction of a facility, it is likely that this will ease as people become accustomed to it. Against this backdrop, social impact indicators are increasingly being used to evaluate the social impact of policies and investments as social value in local government policy decisions and financial institutions' investment and lending decisions.

[0004] Therefore, it is necessary to compare the differences in social impact indicators between multiple policy candidates and to predict changes in social impact indicators over time. Logic models are often used to evaluate social impact indicators.

[0005] Patent Literature 1 has been proposed as a technology related to the above background. Patent Literature 1 discloses a technology that can "stimulate discussion," which is an example of an activity. More specifically, Patent Literature 1 describes that "the creation support server 10 traces the links from a solution node indicating a solution to a problem node indicating a problem back through the links from a solution node indicating a solution to a problem node among multiple nodes included in a tree diagram generated using nodes indicating opinions created by each user to derive a solution from a problem under discussion and links indicating relationships between the nodes, and selects each node included between the solution node and the problem node, calculates the contribution of each node to deriving the solution using the type of each selected node and the type of link, and identifies nodes whose contribution is equal to or greater than a threshold."

[0006] JP 2017-68704 A

[0007] In Patent Document 1, the contribution of each node in a tree diagram to the discussion is evaluated. Here, the validity of the above-mentioned logic model must be ensured. If the validity is insufficient, the social impact prediction results cannot be trusted when the logic model is constructed, and there is a risk of discrepancies in the prediction results (different from what was expected). Furthermore, when constructing a logic model, the base logic model is constructed by trial and error while being modified. Therefore, the objective of the present invention is to ensure the validity of a logic model used to evaluate social impact indicators constructed by trial and error.

[0008] The present invention, which aims to solve the above-mentioned problems, is a logic model construction support device having a processor that executes a program and a storage device that stores the program, wherein the processor executes, in accordance with the program, a logic model construction process that constructs a directed graph having multiple hierarchies represented by edges connecting nodes, a selection process that accepts selections for nodes in the constructed directed graph, a modification process that modifies weighting coefficients corresponding to edges connecting to the selected nodes, a simulation process that creates variation information regarding a first quantitative value of a social impact index of the logic model before the weighting coefficient is modified, a second quantitative value of the social impact index of the logic model after the weighting coefficient is modified, and the difference between the first quantitative value and the second quantitative value, and a variation information output process that outputs the variation information.

[0009] Another aspect of the logic model construction support device of the present invention is a logic model construction support device having a processor that executes a program and a storage device that stores the program, wherein the processor executes, in accordance with the program, a logic model construction process that constructs a directed graph having multiple hierarchies represented by edges connecting nodes, a display process that displays the constructed directed graph, a selection process that accepts selections of nodes for the displayed directed graph, a detail display process that displays details of the selected nodes and the edges connecting to the nodes, a modification process that modifies weighting coefficients corresponding to the displayed edges, a simulation process that creates variation information regarding a first quantitative value of a social impact index of the logic model before the weighting coefficient is modified, a second quantitative value of the social impact index of the logic model after the weighting coefficient is modified, and the difference between the first quantitative value and the second quantitative value, and a variation information display process that displays the variation information.

[0010] The present invention also includes a logic model building support method using a logic model building support device, a program that causes the device to function as a computer, or a logic model building support system that includes a storage medium that stores the program, and the logic model building support device.

[0011] According to the present invention, the validity of a logic model can be ensured when the logic model is constructed.

[0012] 1 is an explanatory diagram showing an example of a system configuration of a logic model building support system 100 according to a first embodiment. FIG. 2 is a functional block diagram of a logic model building support device according to the first embodiment. FIG. 3 is a hardware configuration diagram of a computer that realizes each device according to the first embodiment. FIG. 4 is a flowchart showing a logic model building support process according to the first embodiment. FIG. 5 is a diagram showing a logic model used in the first embodiment. FIG. 6 is an example of a screen display at step S32 in the first embodiment. FIG. 7 is an example of a screen display at step S34 in the first embodiment. FIG. 8 is an example of a screen display at step S35 in the first embodiment. FIG. 9 is an example of a screen display at step S37 in the first embodiment. FIG. 10 is a flowchart showing a logic model building support process according to the second embodiment. FIG. 11 is a graph showing a node influence degree in the second embodiment. FIG. 12 is an example of a screen display at step S32-1 in the second embodiment. FIG. 13 is a flowchart of a social impact index quantification process according to the third embodiment. FIG. 14 is a diagram showing nodes selected in a node selection process S303 according to the third embodiment. FIG. 15 is a diagram showing a logic model in which edges are connected according to the third embodiment. FIG. 16 is a diagram showing a logic model in which weights are added to each edge according to the third embodiment. FIG. 17 is a diagram showing an extracted portion of a logic model in which weights are added to each edge according to the third embodiment. FIG. 10 is a diagram illustrating an example of parameter settings of a driving node according to Example 3. FIG. 11 is a diagram for explaining a quantification process according to Example 3. FIG. 12 is a diagram illustrating parameters of a driving node according to Example 3. FIG. 13 is a diagram illustrating a graph showing the number of occurrences of action requests according to Example 3. FIG. 14 is a diagram illustrating a simulation result in a simulation process S306 according to Example 3.

[0013] An embodiment of the present invention will be described below. In this embodiment, the logic model construction support device supports decision-making, such as the formulation of policies and consensus building for such policies, as an activity. To this end, the logic model construction support device includes a processor that executes a program and a storage device that stores the program. The processor, in accordance with the program, represents a logic model used to evaluate social impact indicators for evaluating organizational activities, and executes the following operations: a display process that displays a directed graph having multiple hierarchies represented by edges connecting nodes constituting the logic model; a selection process that accepts selections of nodes for the displayed directed graph; a detailed display process that displays details of the selected nodes and the edges connected to the selected nodes; a modification process that modifies weighting coefficients corresponding to the displayed edges in response to user operations; a simulation process that generates first quantitative values ​​of social impact indicators of the logic model before the weighting coefficient is modified, second quantitative values ​​of the social impact indicators of the logic model after the weighting coefficient is modified, and variation information related to the difference between the first quantitative value and the second quantitative value; and a variation information display process that displays the first quantitative value, the second quantitative value, and the variation information.

[0014] This ensures validity in the construction of a logic model. Note that "ensuring validity" means that the performance of predicting social impacts and the like using the constructed logic model satisfies predetermined conditions. Here, the performance of prediction can be confirmed by checking whether an index value indicating the accuracy of the prediction result is equal to or greater than a threshold value. Below, examples showing specific examples of this embodiment will be described.

[0015] First, the background of Example 1 will be described. Example 1 is an example of an implementation in which a logic model is constructed in a local government or the like with the aim of improving the social impact index, which is a long-term goal. This example supports the formulation of measures to improve the social impact index of a target area. In other words, this example supports the realization of better measures. To this end, by calculating the extent to which the social impact index will change due to a measure through the logic model, it is possible to provide, for example, a measure with the highest improvement prediction.

[0016] To achieve this, it is important that the logic model is convincing to members of local governments, etc. Therefore, in this embodiment, it is possible to build a highly convincing logic model by presenting the fluctuation values ​​of the social impact index when building the logic model. The configuration for this is explained below.

[0017] 1 is an explanatory diagram showing an example of the system configuration of a logic model building support system 100 according to a first embodiment. This system configuration is also common to the second and third embodiments described below. In FIG. 1, the logic model building support system 100 includes a plurality of first information terminals 101-1 to 101-n, a second information terminal 102, and a logic model building support device 103, which are connected to one another via a network 105. The network 105 can be realized by the Internet, a LAN (Local Area Network), a WAN (Wide Area Network), or the like.

[0018] The first information terminals 101-1 to 101-n are information terminals used by participant 1 to participant n (n is an integer equal to or greater than 1; hereinafter collectively referred to as participants), respectively, and acquire information on the logic model structure envisioned by the participants and indicators they consider important, and transmit this information to the logic model construction support device 103. For example, this can be realized by an information processing device (computer) such as a personal computer or a communication device like a smart device. Note that a participant refers to someone who takes part in the organization's activities (policies). For example, if the organization is a local government, this would be residents or general employees.

[0019] The second information terminal 102 is an information terminal used by a decision maker such as a local government, and displays simulation results including the logic model being constructed and quantitative values ​​of the social impact indicators from the logic model construction support device 103. For example, it can be realized by an information processing device (computer) such as a personal computer or a communication device such as a smart device. Note that the decision maker refers to a person who decides the activities (policies) of an organization. For example, if the organization is a local government, it is the head or a senior official.

[0020] The logic model construction assistance device 103 also receives the structures of the participants' logic models and indicators that they consider important from at least one of the first information terminals 101-1 to 101-n. The logic model construction assistance device 103 then executes the following processes.

[0021] A construction process that constructs a directed graph A selection process that accepts selections for nodes in the directed graph A modification process that modifies the weighting coefficients corresponding to the edges connected to the selected nodes A simulation process that creates variation information regarding the first quantitative value of the social impact index of the logic model before the weighting coefficient is modified, the second quantitative value of the social impact index of the logic model after the weighting coefficient is modified, and the difference between the first quantitative value and the second quantitative value A variation information output process that outputs the variation information The logic model construction support device 103 can further perform an output process that outputs the constructed directed graph and a detailed output process that outputs details of the selected nodes and the edges connected to those nodes.

[0022] Here, the variation information output process, output process, and detailed output process include each display process that displays the output target. Furthermore, the various display processes described in this specification are examples of output processes. Therefore, other output process modes, such as notification and transmission, can also be adopted in the present invention. Furthermore, the variation information output process may also output a first quantitative value and a second quantitative value. The logic model construction support device 103 may also be realized as a social impact index quantification device that calculates and quantifies a social impact index.

[0023] Furthermore, the logic model building support device 103 transmits quantifiable logic model data 1041 and execution results of information processing (for example, quantitative values ​​of social impact indexes) to the second information terminal 102. The database 104 stores data used by the logic model building support device 103 (logic model data 1041 and social impact index data 1042). The database 104 may be realized as a storage device of the logic model building support device 103, or may be connected to the logic model building support device 103 via the network 105. The logic model building support device 103 may also be realized by an information processing device (computer) such as a server.

[0024] Next, a logic model construction support device that executes the main processing of Example 1 will be described. Fig. 2 is a functional block diagram of a logic model construction support device 103 according to Example 1. In Fig. 2, the logic model construction support device 103 includes a communication unit 1031, a directed graph generation unit 1032, a simulation unit 1033, and a storage unit 1040. Here, the storage unit 1040 in Fig. 2 corresponds to the database 104 in Fig. 1, and the logic model construction support device 103 is configured to include the storage unit 1040, but the storage unit 1040 may be configured separately from the logic model construction support device 103.

[0025] Each unit will be described below. The communication unit 1031 connects to the network 105 and communicates with the first information terminals 101-1 to 101-n and the second information terminal 102. The directed graph generation unit 1032 generates a directed graph having multiple hierarchies represented by edges connecting nodes in a logic model.

[0026] The simulation unit 1033 also calculates the quantitative value of the social impact index for the policy based on the generated directed graph, and calculates the fluctuation value of the quantitative value of the social impact index.

[0027] Furthermore, the storage unit 1040 stores logic model data 1041 and social impact index data 1042. Here, the logic model data 1041 includes logic models, including a logic model that is the basis for construction, a logic model under construction that is modified in response to operations on the logic model, and a constructed logic model.

[0028] The social impact index data 1042 indicates the social impact index of the logic model, and includes its quantitative value and its fluctuation value. The logic model data 1041 and the social impact index data 1042 may be treated as a single piece of information. In other words, these may be information in which the social impact index is associated with each piece of logic model data.

[0029] Next, the hardware configuration of each device in Example 1 will be described. Fig. 3 is a hardware configuration diagram of a computer that realizes each device according to Example 1. The computer 200 shown in Fig. 3 is common to the first information terminals 101-1 to 101-n, the second information terminal 102, and the logic model construction support device 103. Note that the following describes the configuration specific to the logic model construction support device 103 that executes the main processing of this example, and the configuration shared by each device.

[0030] The computer 200 has a processor 201, a storage device 202, an input device 203, an output device 204, and a communication interface (communication IF) 205. These are connected to one another by a bus 206. Note that if the logic model construction assistance apparatus 103 is implemented as a server, the input device 203 and output device 204 can be omitted. Each component will be described below. The processor 201 executes calculations in the computer 200 in accordance with a program stored in the storage device 202. The storage device 202 is composed of a work area for the processor 201 and a storage medium or storage device (hereinafter simply referred to as a storage medium). The work area can be implemented as so-called memory. The storage medium can be implemented as a non-temporary or temporary recording medium that stores various programs and data. The storage device 202 can be configured with, for example, a ROM (Read Only Memory), a RAM (Random Access Memory), a HDD (Hard Disk Drive), or a flash memory.

[0031] Here, a directed graph generation program, a simulation execution program, and a variation value calculation program are stored as programs in the storage device 202 of the logic model construction assistance device 103. These programs cause the processor 201 to execute the processes of the directed graph generation unit 1032 and the simulation unit 1033 in FIG.

[0032] 2. Therefore, the storage medium can be realized as a device separate from the logic model building support device 103.

[0033] The input device 203 accepts operations from users such as participants and decision makers. The input device 203 can be realized, for example, by a keyboard, mouse, touch panel, numeric keypad, scanner, or microphone. The output device 204 outputs data such as the processing results of the processor 201. The output device 204 can be realized, for example, by a display, printer, or speaker. The communication IF 205 connects to the network 105, communicates with other devices, and transmits and receives logic data, various instructions, and data. In the logic model construction support device 103, the communication IF 205 corresponds to the communication unit 1031 in FIG. 2. This concludes the description of the configuration of the first embodiment, but FIGS. 2 and 3, like FIG. 1, are also common to the second and third embodiments described below.

[0034] Next, we will explain social impact indicators and logic models in detail. While there is generally no clear definition of social impact indicators, in this example, they indicate the social impact generated by activities such as policies. The degree of this social impact can be expressed as a numerical value (quantitative value). Social impact indicators target issues such as equality, livelihood, health, nutrition, poverty, safety, and justice. Specific examples from local areas also include subjective indicators such as residents' work attitudes and awareness of the SDGs. Social impact is also increasingly being used in finance, such as ESG investment and social impact bonds, to consider social and environmental impacts in investment decisions.

[0035] There are several methods for evaluating social impact indicators, one of which is the use of logic models. Logic models can be expressed as a network model (e.g., a directed graph) consisting of nodes and edges. Here, nodes represent various indicators, and edges represent the causal relationships between the indicators. Logic models are also conceptual tools for clarifying the logic of social programs, and specifically, they can clarify how invested resources will logically affect the target social impact indicators. A typical logic model is often composed of five main layers. Each layer is explained below.

[0036] The first layer, the input layer, includes both resources and constraints for the policy. Resources are used when the policy is implemented, and include not only human resources, material resources, and financial resources, but also cultural resources. It is important to utilize these resources and generate effective social impact indicators within the constraints.

[0037] The second layer, the activity layer, is the policy or the actions and activities for implementing the policy, including processes, work and actions.

[0038] The third layer, the Output Layer, is the direct result of the activities carried out by the implementers. The results include business activities such as products and services provided to the beneficiaries and ongoing support for the project.

[0039] The fourth layer, the outcome layer, is the ripple effect of the output layer above, and refers to the indicators necessary to improve the social impact indicators. The ripple effects of "results" often include subjective indicators such as increased motivation, satisfaction, and recognition. "Returns" can sometimes be divided into short-term and long-term effects.

[0040] The fifth layer, the impact layer (social impact layer), is the final goal shared by the implementer and all stakeholders, and is an indicator of improvement toward solving social problems. While broadly speaking, issues such as reducing disparities are addressed in the SDGs, some organizations may set more local goals. In some cases, output or outcome indicators may be used as substitutes for social impact indicators. In such cases, intermediate indicators are considered as targets.

[0041] The main focus of existing logic models is to logically clarify the ripple effects, and qualitative and subjective indicators are included in the outcome layer as indicators of these ripple effects. These indicators are necessary given the purpose of social impact indicators, and are considered effective in assessing the validity of the logic.

[0042] Therefore, in this embodiment, a logic model that allows the social impact index to be quantified is used. Specifically, the following process is performed.

[0043] (1) First, a weighting factor (numerical value) is added to the edges of an existing logic model (directed graph) to calculate a quantitative value of the social impact index for the activities in the logic model.

[0044] (2) Next, the weight coefficients of the existing logic model are manipulated to modify the existing logic model (directed graph), thereby generating a new directed graph that represents the modified logic model. The modified logic model can be expressed as a logic model under construction.

[0045] (3) Next, a quantitative value of the social impact index for the activity in the logic model under construction represented by the new directed graph is calculated. This quantitative value changes according to the manipulated weight coefficient.

[0046] (4) Next, the difference (amount of change) between the quantitative value of the social impact index in the existing logic model and the quantitative value of the social impact index in the logic model being constructed is calculated as a variation value. (5) Then, by presenting the variation value to participants and decision makers, the relationship between the edge and the quantitative value of the social impact index can be visualized, providing a highly convincing logic model. The presented information is not limited to variation values ​​(numerical values), but can be treated as variation information related to the difference. For example, it may include specific numerical values, detailed information indicating the content of the change, or an outline showing its summary. For example, an outline may be output to the first information terminals 101-1 to 101-n (participants), and detailed information may be output to the second information terminal 102 (decision maker). Furthermore, different information may be included for each of the first information terminals 101-1 to 101-n (participants). Furthermore, the detailed information may include information other than the variation value.

[0047] The above-described existing logic models and the like are stored in the storage unit 1040 as logic model data 1041. In addition, the calculated quantitative values ​​and fluctuation information are stored in the storage unit 1040 as social impact index data 1042. For this reason, the logic model is constructed in advance.

[0048] Furthermore, at least two of the first information terminals 101-1 to 101-n and the second information terminal 102 can share the screens shown in FIGS. 6 to 9 and 12, which will be described later, and can perform operations on those screens.

[0049] Next, the above-mentioned process will be described in detail. The configuration of the main processing entity of this process will be described with reference to Fig. 2. Fig. 4 is a flowchart showing the logic model construction support process according to the first embodiment. This flowchart is based on the premise that the storage unit 1040 has stored an existing logic model as logic model data 1041 in advance.

[0050] 4, the directed graph generation unit 1032 reads an existing logic model from the logic model data 1041 in the storage unit 1040 in response to an instruction from either the first information terminals 101-1 to 101-n or the second information terminal 102. Here, the logic model may be read automatically by the directed graph generation unit 1032 at a predetermined interval or the like. Furthermore, if the storage unit 1040 is implemented as a device separate from the logic model construction support device 103, the communication unit 1031 will receive the existing logic model. Furthermore, in this step, an existing logic model created in either the first information terminals 101-1 to 101-n or the second information terminal 102 may be used.

[0051] Here, we will explain the logic model used in this embodiment, such as the existing logic model read in step S31. Figure 5 is a diagram showing an example of the logic model used in Example 1. In Figure 5, for ease of explanation, the logic model is expressed as four layers, but any number of layers is acceptable. Also, nodes in the input layer are represented as driving nodes, nodes in the impact layer are represented as social impact indicators, and nodes in the activity layer to outcome layer are represented as intermediate nodes. Note that this diagram can also be treated as a directed graph as is.

[0052] In addition, the edges connecting each node in Figure 5 are shown with solid and dashed lines. Of these, solid lines represent positive weight coefficients, and dashed lines represent negative weight coefficients. An edge with a positive weight coefficient has a positive effect from the source side of the edge (the base of the arrow) to the destination side of the edge (the tip of the arrow). In other words, if the effect of the node on the source side increases, the effect of the node on the destination side is expected to increase. For example, in the case of a policy to improve academic ability, if the node on the source side represents "motivation to study" and the node on the destination side represents "test scores," it is expected that an improvement in "motivation to study" will lead to an improvement in "test scores."

[0053] A negative weighting coefficient represents the inverse relationship of the positive weighting coefficient. Weighting coefficients can be normalized to real numbers between -1 and +1, and the absolute value of the coefficient represents the magnitude of the effect.

[0054] Next, returning to FIG. 4 , the description of the processing of the first embodiment will be continued. In step S32 of FIG. 4 , the directed graph generation unit 1032 executes a logic model construction process to create, that is, construct, a directed graph that represents the read logic model. At this time, the directed graph may be displayed. Note that this display is an example of output, and output includes various modes other than display. Note that the displays in the following embodiments are also examples of output, and are not limited to these.

[0055] In this step, the directed graph generation unit 1032 notifies at least one of the first information terminals 101-1 to 101-n and the second information terminal 102 of the directed graph via the communication unit 1031. As a result, the first information terminals 101-1 to 101-n and the second information terminal 102 display the notified directed graph, allowing at least one of the participants and decision makers to confirm its contents. The information terminals to which the directed graph is notified (displayed) may be limited as follows. For example, the information terminal may be an information terminal used by a user who wishes to modify the logic model to increase the persuasiveness of the logic model, or at least one of the first information terminals 101-1 and the second information terminal 102. The directed graph is an example of a representation format for a logic model, and the target of notification or display may be anything that indicates the logic model. For this reason, the logic model may be notified, and the first information terminals 101-1 to 101-n and the second information terminal 102 may create and display a directed graph. The logic model may also be displayed in a different display format.

[0056] An example of a display screen is shown in Fig. 6. In Fig. 6, a display screen 501 has a logic model display section 502 that displays the logic model and a social impact display section 503. The logic model display section 502 displays the notified directed graph, i.e., a configuration diagram of the logic model. In Fig. 6, the directed graph is displayed as a graph structure of nodes and edges, but it may also be displayed as a list, such as a node list or edge list. In the node list, a list of each node is displayed, and the display information for each node includes the connected edge and the connected node.

[0057] The social impact display section 503 also displays each social impact index in the logic model indicated by the directed graph in the logic model display section 502. That is, the social impact index predicted value (SI predicted value), the social impact index predicted value (reference value) (SI predicted value (reference value)), and the social impact index predicted value (variation value) (SI predicted value (variation value)) are displayed.

[0058] The SI forecast value displays the quantitative value of the social impact index of the logic model after correction in the subsequent steps (S33 to S36). In FIG. 6 , since the data is before correction, the SI forecast value (reference value), which will be described later, is displayed as is. The SI forecast value (reference value) displays the quantitative value of the social impact index of the logic model shown by the directed graph in the logic model display section 502, i.e., the quantitative value before correction. The SI forecast value (variation value), which is an example of variation information, displays the variation value, which is the difference (amount of change) between the previous SI forecast value and the SI forecast value (reference value). This difference may be a variation rate. The variation rate is the value obtained by dividing the difference by the SI forecast value (reference value). Note that in the example of FIG. 6 , since the data is before correction in S33 to S36, the variation value is 0. Note that in this embodiment, two indicators are used as social impact indicators. These indicators correspond to social impact index 1 (SI1) and social impact index 2 (SI2), respectively. The display screen 501 in FIG. 6 may include only the logic model display section 502 .

[0059] Next, returning to Fig. 4, the description of the processing of the first embodiment will be continued. In S33 of Fig. 4, a node selection process is executed. To do this, first, the first information terminal 101-1 or the second information terminal 102 accepts, from a participant or decision maker, a selection of a node that the participant or decision maker wishes to change from among the nodes displayed in the logic model display unit 502. Next, the first information terminal 101-1 or the second information terminal 102 notifies the logic model construction assistance device 103 of the selected node via the network 105. Then, the directed graph generation unit 1032 accepts the selected node via the communication unit 1031.

[0060] Furthermore, in step S34, detailed display processing of edges and nodes is performed in accordance with the node selection in step S33. To this end, the directed graph generation unit 1032 identifies detailed information about the selected node from the logic model data 1041. This detailed information includes the edges to which the node is connected and their weight coefficients. Next, the directed graph generation unit 1032 notifies the first information terminal 101-1 and the second information terminal 102 of this detailed information using the communication unit 1031. Then, the first information terminal 101-1 and the second information terminal 102 perform display including the detailed information.

[0061] Furthermore, in step S35, a change process is performed to change the weight coefficients of the edges displayed in step S34 in response to an operation from a participant or decision maker. To do this, the first information terminal 101-1 and the second information terminal 102 accept the changed weight coefficients from the participant or decision maker. Next, the first information terminal 101-1 and the second information terminal 102 notify the logic model construction assistance device 103 of the changed weight coefficients via the network 105. Next, the directed graph generation unit 1032 accepts the changed weight coefficients via the communication unit 1031. Then, the directed graph generation unit 1032 changes the weight coefficients of the logic model data 1041 to the accepted weight coefficients.

[0062] In step S36, the simulation unit 1033 executes a simulation process using the changed weighting coefficients. This simulation process includes calculation of the SI prediction value and the fluctuation value using the changed weighting coefficients. The simulation process also includes creation of fluctuation information.

[0063] Here, a specific example of steps S33 to S36 will be described. Here, an example will be described in which a node labeled intermediate node 2-2 is selected in the logic model represented by the displayed directed graph. First, in step S33, a selection process is executed to accept selection by clicking or tapping on the node in question on the logic model display unit 502. Next, as described above, in step S34, the selected node and the edges corresponding to this node are displayed. In this step, details of the selected node are displayed on the selected node screen 604 shown in FIG. 7. The displayed contents are the nodes and edges that connect to the selected node. In this screen example, the selected node is intermediate node 2-2. Therefore, the nodes and edges that connect to intermediate node 2-2 are the edge between intermediate node 1-1 and number 1, the edge between intermediate node 1-2 and number 2, the edge between intermediate node 1-3 and number 3, and the edge between intermediate node 1-4 and number 4.

[0064] In addition, the weighting coefficients for the corresponding edges are displayed. In this example screen, edge number 1 is 0.25, edge number 2 is -0.25, edge number 3 is -0.25, and edge number 4 is 0.25. Here, it is assumed that the weighting coefficients are allocated so that the sum of the absolute values ​​is 1.

[0065] Next, in step S35, the first information terminals 101-1 to 101-n and the second information terminal 102 execute a weighting factor change process to accept changes to the weighting factors corresponding to edges from participants or decision makers. As a result, as shown in FIG. 8 , the weighting factor values ​​in the edge number and weighting factor display section 701 are input and changed using an input device 203, such as a keyboard, attached to the first information terminals 101-1 to 101-n and the second information terminal 102. In the example screen shown in FIG. 8 , the weighting factor for edge number 1 is changed to 0.40, the weighting factor for edge number 2 is changed to −0.10, the weighting factor for edge number 3 is changed to −0.40, and the weighting factor for edge number 4 is changed to 0.10. Thereafter, the first information terminal 101-1 and the second information terminal 102 accept selections from participants or decision makers by clicking the apply button 702, for example. As a result, the changed weighting factors are sent to the logic model construction support device 103 via the network 105. Then, in step S36, the quantitative values ​​of the social impact indexes are recalculated in accordance with the change in edge, that is, the fluctuation information is created, including the calculation of the SI predicted values ​​and fluctuation values.

[0066] Next, in step S37, a fluctuation information display process is executed for the results of the simulation process in step S36. To this end, the simulation unit 1033 notifies the first information terminal 101-1 and the second information terminal 102 of the simulation results, including at least the SI predicted value and the fluctuation value, using the communication unit 1031. The first information terminal 101-1 and the second information terminal 102 then display the notified SI predicted value and the fluctuation value. An example of this screen is described below. As shown in FIG. 9 , the notified fluctuation amount is displayed in the social impact display unit 803. In the social impact display unit 803, the SI1 predicted value and the SI2 predicted value are displayed after being changed to 0.113 and 0.113, respectively, due to recalculation of the quantitative value of the social impact index following the edge change. Additionally, the SI1 change amount (variation value) and SI2 change amount (variation value) are displayed as +0.019 and +0.019, respectively, which are the differences between the predicted value and the SI1 predicted value (before change) and the SI2 predicted value (before change). In this example screen, the SI1 and SI2 predictions have changed in the same way, but they may change differently depending on how the logic model is changed.

[0067] Next, in step S38, it is determined whether the participants or decision makers are satisfied with the changed social impact index. To do this, first, the first information terminal 101-1 or the second information terminal 102 receives input from the participants or decision makers indicating whether they are satisfied. Then, the first information terminal 101-1 or the second information terminal 102 notifies the logic model construction support device 103 of the input content indicating whether they are satisfied via the network 105. Next, the simulation unit 1033 of the logic model construction support device 103 receives the notified input content via the communication unit 1031. Then, if the simulation unit 1033 is satisfied, it transitions to step S39. On the other hand, if the participants are not satisfied, it transitions to step S33.

[0068] In step S39, the simulation unit 1033 sends the changed logic model to the first information terminal 101-1 and the second information terminal 102 via the communication unit 1031 and saves the changed logic model in the database 104 as logic model data 1041. On the other hand, if the user is not satisfied, the process returns to step S303 and repeats the subsequent processes.

[0069] In this way, users such as participants and decision makers can refer to the quantitative values ​​of social impact indicators obtained through calculation processing and the fluctuations in the quantitative values ​​of social impact indicators when modifying the corresponding logic model. This helps users to build logic models that are satisfactory to them, and is expected to enable users to efficiently decide on measures with high social impact indicators. Here, "satisfying the user" is an example of "ensuring validity," and can be confirmed by receiving input from the user indicating their consent or by meeting the indicator values ​​set by the user.

[0070] In comparison with Example 1, Example 2 is an example in which a function is added that takes into account the node influence degree and supports node selection performed by users such as participants and decision makers when modifying a logic model. Note that the configuration of each device in Example 2 is also the same as in Example 1. Therefore, the processing flow of Example 2 will be described. FIG. 10 is a flowchart showing logic model construction support processing according to Example 2. Example 2 adds a step of calculating the node influence degree (step S901) to the flowchart of Example 1 in FIG. 4. Furthermore, step S32 is changed (S32-1) in accordance with the addition of step S901. In FIG. 10, steps common to Example 1 are denoted by the same reference numerals.

[0071] Below, we will explain step S901, which is different from Example 1. In this example, when a user modifies a logic model, the user must select a node to modify. However, when a logic model has a large number of nodes, there is a problem in that it is unclear which node to select from in order to obtain a more convincing logic model. Therefore, the user must select nodes in order by exhaustive selection. In Example 2, the node influence degree is calculated (step S901), and the user is presented with nodes that have a greater influence on the quantitative value of the social impact index, thereby assisting the user in selecting nodes when modifying the logic model.

[0072] In step S901, the digraph generation unit 1032 calculates the node influence degree, which indicates the degree of influence of a node on a social impact index, as follows. Here, the calculation target is node A. The digraph generation unit 1032 fixes each weight coefficient, except for the weight coefficient of the edge connected to node A, to a predetermined value. This predetermined value is, for example, a representative value such as an average value. The digraph generation unit 1032 then changes the quantitative value of node A. This change is performed continuously, for example, between 0 and 1. The digraph generation unit 1032 records the quantitative value of the social impact index when the quantitative value of node A is changed, and compares the maximum and minimum values, or the reference value and minimum value, among them. The reference value here is a predetermined numerical value, such as 0.5, if the amount of change (variation value) from the quantitative value is between 0 and 1. Furthermore, the digraph generating unit 1032 uses the following (Equation 1) to set the minimum value or the reference value as a reference and determines the ratio of the compared difference values ​​as a fluctuation value.

[0073] Influence degree node A = (reference value of quantitative value of social impact index - minimum value of quantitative value of social impact index) / (minimum value of quantitative value of social impact index) (Equation 1) The directed graph generation unit 1032 performs this process for each intermediate node and compares the influence degree. Here, since the difference value of the reference value is 0, the influence degree is also defined as 0. Figure 11 shows the influence degree of each node in the logic model used in this example. Figure 11 is a graph showing the node influence degree in Example 2. This graph shows the influence degree for each intermediate node. Furthermore, if there are multiple social impact indexes, a representative value such as the average or total of the influence degree on each social impact index is used as the influence degree. In the example of Figure 11, it can be seen that the node with the greatest influence degree on the social impact index is intermediate node 2-2. The node with the second greatest influence degree is intermediate node 1-1, and the node with the third greatest influence degree is intermediate node 1-4. When modifying a logic model, it is efficient to start with the node with the greatest (or least) influence degree.

[0074] For this reason, in step S32-1, a display process is performed that takes into account the degree of influence. For example, the directed graph generation unit 1032 creates a directed graph including the degree of influence and its ranking, and the first information terminal 101-1 and the second information terminal 102 display this. FIG. 12 shows an example in which the degree of influence is presented on a directed graph that represents a logic model. In this figure, the nodes ranked 1st to 3rd (highest) in terms of influence degree are presented to the user. However, it is also possible to display the entire ranking, or lower rankings, or to display the numerical value of the degree of influence instead of the ranking. Furthermore, the above-mentioned higher-ranking nodes may be distinguished from the others or displayed in a limited manner.

[0075] As described above, by calculating the degree of influence and presenting it on the logic model, the user can refer to it as a candidate for selecting a node to modify when modifying the logic model.

[0076] In Example 3 and Example 4 described below, examples of the above-mentioned quantification method will be described. Therefore, in this example, the difference in social impact index between a large number of policy candidates is quantitatively predicted (quantified) and support is provided for the consideration of policies that maximize the social impact index. In this case, a directed graph representing a logic model for the policy candidates is generated, and the details of this graph will be described below. In this example and Example 4, the generation of the directed graph is referred to as the construction of a logic model. However, as in Examples 1 and 3, the directed graph is one form of logic model, and the logic model may be expressed in other forms. Therefore, the directed graph generation unit 1032 can also be referred to as a logic model construction unit.

[0077] FIG. 13 is a flowchart of the social impact index quantification process according to Example 3. This social impact index quantification process is broadly divided into two processing blocks: a logic model construction process S301 for enabling quantification, and a simulation process S302. First, the quantification-oriented logic model construction process S301 corresponds to the directed graph generation process described above and includes a node selection process S303, an edge weight prediction process S304, and a driving node parameter setting process S305. The node selection process S303 selects nodes required for the logic model using data obtained through workshops with IT support, etc. Furthermore, the edge weight prediction process S304 predicts and adds weights to edges based on the selection for quantification. This completes the construction of a logic model. Then, the driving node parameter setting process S305 sets parameters for nodes (referred to as driving nodes) that enable simulation of the logic model created in S304, thereby constructing a quantifiable logic model. The logic model construction process S301 is performed using the first information terminal 101 and data acquired from participants (inputs indicating that the indicators are important).

[0078] Furthermore, in the simulation process S302, the logic model constructed in S301 is quantified and evaluated for policy formulation. To this end, the simulation process S302 includes a simulation process S306 and an analysis process S307. The simulation process S306 executes a time evolution process starting from a driving node added to the logic model, thereby executing a quantification simulation of the social impact index.

[0079] Furthermore, analysis processing S307 analyzes the results of simulation processing S306. Details of each processing will be explained in order below. In the following explanation, a specific example will be taken as an example of a power saving request when natural energy is introduced in a simple region. In this case, the social impact index will be "power saving acceptance" and the measure will be "power saving request." Note that hereinafter, the social impact index will also be simply referred to as "index." Note that in order to quantify the social impact index, the introduction of driving nodes and the addition of weights to edges will be added to the logic model. Focusing on these, the processing of this embodiment will be explained according to the flowchart of FIG. 13.

[0080] First, in the node selection process S303, the directed graph generation unit 1032 selects indicators for each of the five layers in the logic model. In the logic model of this embodiment, the five layers are considered to be categorized into two layers with different behaviors from a quantification perspective. The first layer (input / activity / output) combines the input layer, activity layer, and output layer. Similarly, the second layer (outcome / social impact) combines the outcome layer and impact layer. This process involves cleaning each of the above layers, i.e., the nodes.

[0081] FIG. 14 is a diagram showing the nodes selected (picked up) in the node selection process S303. The layers illustrated in FIG. 14 are a first layer 401 and a second layer 402. FIG. 14 also illustrates examples of indices in the first layer 401 and the second layer 402. The indices illustrated in the first layer 401 are as follows: Power generation facility size (solar power) 403 Power generation facility size (nano-hydro) 404 Degree of power saving request 405 Number of power saving requests (weekday, tight) 406 Number of power saving requests (weekday, not tight) 407 Number of power saving requests (holiday, tight) 408 Number of power saving requests (holiday, not tight) 409 Of course, other indices can be added or deleted. In addition, the indexes shown as examples of the second layer 402 are environmental consciousness 410, dissatisfaction with requests to save electricity (negative node) 411, and power saving acceptance 412. Indicators other than these can also be added or deleted from the second layer 402. Here, dissatisfaction with requests to save electricity 411 is noted as a negative node.

[0082] Here, when the index value becomes negative, it indicates an improvement, and the positive and negative direction of improvement is opposite to that of other indexes, such as environmental consciousness 410. For example, when these indexes are quantified, the index "environmental consciousness" 410 becomes positive as the value increases, because environmental consciousness increases. On the other hand, the index "dissatisfaction with requests to save electricity" 411 becomes negative as the value increases, because dissatisfaction increases. This makes post-processing simulations complicated, so the polarity of negative nodes is reversed, indicating that dissatisfaction decreases as the value increases.

[0083] The node in question is marked as a negative node (for example, this can be considered equivalent to converting the indicator into "satisfaction with power saving requests" and making it a positive node). Also, for indicators where it is difficult to determine the direction of improvement, it is necessary to decide in advance whether to label them as such.

[0084] Here, the first layer selects indicators (nodes) of activities related to power generation facilities and power saving requests. These power saving requests are assumed to request a certain amount of power reduction when there is a shortage of renewable energy power. In addition, since the occurrence of power saving requests depends on the amount of renewable energy generated, the scale of power generation facilities and the degree of power saving requests are used as indicators.

[0085] The power saving request is also divided into several categories depending on the situation. In this example, it is divided by a combination of the day of the week and whether it is urgent or not. First, the day of the week is divided into Monday to Friday (weekdays) or Saturday and Sunday (holidays). In addition, the outdoor temperature is divided into four indicators by combining whether it indicates a high pressure (above 35°C or below 5°C) or a low pressure (other than the above conditions).

[0086] These indices, i.e., nodes, can be selected by creating a database of related keywords in advance and selecting the necessary indices from the database 104, or by selecting them in a workshop or the like. The directed graph generation unit 1032 then selects each indices in accordance with operations on each information terminal and predetermined rules. These indices become nodes in the logic model. The workshop includes acquiring data from various terminals using the logic model building support system 100.

[0087] In edge weight prediction processing S304, the directed graph generation unit 1032 connects the nodes selected in S303 with edges and predicts the weight of each edge. Here, connections at edges are originally intended to logically express causal relationships, but in the case of quantification, a calculation formula may be inserted between nodes.

[0088] An example of this is shown in FIG. 15. FIG. 15 is a diagram illustrating a logic model in which edges are connected according to this embodiment. In the example of FIG. 15, a process 513 for calculating power supply and demand is inserted into the input / activity / output layer. The reason for this insertion is that fluctuations in the amount of power generated by power generation facilities can affect the frequency of power saving requests. To calculate this effect in detail, the power supply and demand calculation can be realized by simulation. Alternatively, for a simple calculation, a method can be used in which the frequency of power saving requests is determined in advance (for example, two power saving requests occur in an average month) without taking the power generation facilities into consideration.

[0089] Here, there are four edges 514 input to index 510, four edges 515 input to index 511, and two edges 516 input to index 512. In Fig. 15, edge 515 is an edge input to a negative node (index 511), and is therefore depicted with a dashed line to clearly indicate that it is different from a positive node.

[0090] Furthermore, a logic model in which weights (numbers from -1 to 1) have been assigned (predicted) by the directed graph generation unit 1032 to each edge (514, 515, 516) in FIG. 15 is shown in FIG. 16. Furthermore, a diagram in which four edges 614 of index 610 are extracted from the logic model in FIG. 16 is shown in FIG. 17. For index 710, the weight assigned to the edge input from index 706 is 0.0 (figure number 717). Furthermore, the weight assigned to the edge input from index 707 is +0.66 (figure number 718), and the weight assigned to the edge input from index 708 is 0.0 (figure number 719). Furthermore, the weight assigned to the edge input from index 709 is +0.34 (figure number 720).

[0091] In this embodiment, the directed graph generation unit 1032 adds (predicts) weights in accordance with the following rules: 1. The absolute value of the sum of the weights of the edges input to one index node is "1." 2. Nodes to which edges with negative weights are input are separated from nodes to which edges with positive weights are input. It is desirable that the directed graph generation unit 1032 outputs the logic models shown in FIGS. 14 to 16 to each information terminal using the communication unit 1031. As a result, each information terminal can display a directed graph representing the logic model shown in FIGS. 14 to 16 on the output device 204.

[0092] Furthermore, the directed graph generating unit 1032 determines and adds weight values ​​to edges based on the values ​​(preference order) of the policy decision maker or interested parties. This example will be described with reference to FIG.

[0093] In FIG. 17 , four nodes are connected to the indicator 710 "environmental awareness" from top to bottom: indicator 706 "power saving request, weekday, tight," indicator 707 "power saving request, weekday, not tight," indicator 708 "power saving request, weekend, tight," and indicator 709 "power saving request, weekend, not tight." This shows the logic by which environmental awareness changes depending on the four power saving requests. The weight indicates the degree of influence. This degree of influence can be derived objectively (statistical correlation value), but here we assume that it is given in subjective order of preference. A weight of "0.0" indicates no influence or is not considered, and an absolute large weight is assigned to the order of greatest influence. A negative influence is assigned a negative weight.

[0094] For example, in the example of FIG. 17 , the digraph generation unit 1032 determines that the influence of the indicator 707 "power saving request, weekday, not tight" (weight +0.66, diagram number 718) is greatest. The digraph generation unit 1032 also determines that the next most important indicator is the indicator 706 "power saving request, weekend, not tight" (weight +0.34, diagram number 720). The digraph generation unit 1032 then determines that the rest are irrelevant (assuming that a power saving request during times of tightness would not be perceived positively). If these are expressed in order of preference, the degree of the power saving request to "environmental awareness" is "power saving request, weekday, not tight" > "power saving request, weekend, not tight."

[0095] As mentioned above, the sum of the weights is set to "1," so the directed graph generation unit 1032 can estimate (predict) the weights from the order of preference using linear allocation or the like. In the example of FIG. 17, the weight is 0.0 + 0.66 + 0.0 + 0.34 = 1. The directed graph generation unit 1032 can also add weights to other nodes in the same way. In other words, if an edge input to a negative node has a negative weight, the sum is set to "-1." Furthermore, as per rule 2 above, negative nodes and positive nodes are considered separately, and if an edge with a positive weight is input to a certain index node, an edge with a negative weight is not input.

[0096] Next, the parameter setting process S305 for the driving node will be described. Here, the driving node is a node required for a quantifiable logic model, and is a definition not found in conventional logic models. The digraph generation unit 1032 selects, as the driving node, a node from the input / activity / output layer that is to be considered as a candidate measure. In this embodiment, it is desirable to select as a candidate measure the indicators "power generation facility" 403, 404 and the indicator "power saving request level" 405 shown in FIG. 14 when they are changed. For this reason, the digraph generation unit 1032 identifies the indicators "power generation facility size (solar power)" 403, "power generation facility size (nanohydro)" 404, and "power saving request level" 405 as driving nodes. Alternatively, if calculations are not performed within the input / activity / output layer, the digraph generation unit 1032 sets the following indicators as driving nodes: Indicator “power saving request, weekday, tight” 406 Indicator “power saving request, weekday, not tight” 407 Indicator “power saving request, weekend, tight” 408 Indicator “power saving request, weekend, not tight” 409 In addition, the directed graph generation unit 1032 can also exclude the indicator “power generation facility size (solar power)” 403, the indicator “power generation facility size (nano-hydro)” 404, and the indicator “degree of power saving request” 405.

[0097] Here, it is necessary to set a time series parameter for the driving node in advance. The time series parameter is, for example, a value for each month from January to December. For the indicator "power generation facility size (photovoltaic)" 403, the time series parameter can be the system capacity (kW) of photovoltaic power generation. Generally, the system capacity of a power generation facility does not change over time, so the same number (e.g., 10 kW) is assigned as the time series parameter from January to December.

[0098] On the other hand, when the indicator "power saving request, weekday, tight" 406 is set as a driving node, the number of power saving requests in January, the number of power saving requests in February, etc. are set as time-series parameters. Fig. 18 shows an example of parameter setting for a driving node according to this embodiment. In this example, the parameter setting is assumed to be for one year in monthly units.

[0099] Also, the upper diagram 801 in FIG. 18 is an example of a photovoltaic power generation system capacity, and the lower diagram 802 in FIG. 18 is an example of the number of times a power saving request has occurred. Furthermore, since the output of the driving node may have different units, it needs to be normalized to a numerical value such as 0 to 1 as necessary. Here, any function can be used for normalization, but basically it is sufficient as long as it reflects the degree of influence on the logic in the subsequent stage. For example, one method is to normalize using the maximum value among driving nodes with the same unit. That is, in the parameter setting S305 of the driving node, the directed graph generation unit 1032 sets (specifies) the time series parameters that it determines as described above.

[0100] Through the logic model construction process S301 consisting of the above steps S303 to S305, the directed graph generation unit 1032 constructs a quantifiable logic model. The directed graph generation unit 1032 stores the constructed quantifiable logic model in the storage unit 1040 as logic model data 1041.

[0101] Next, the simulation process S302 will be described. First, in simulation processing S306, the simulation unit 1033 executes a time-evolution simulation for the quantifiable logic model. This simulation is executed with a minimum time step equivalent to one hour, which can be changed to a monthly unit, etc., as necessary. Furthermore, for each hourly step, the simulation unit 1033 transfers values ​​to the connected nodes according to the parameters set in the driving nodes. Then, the simulation unit 1033 executes predetermined processing in the connected nodes according to the transferred values. As a result, quantification processing is executed for the nodes. Note that this simulation process S302 is preferably the same as the simulation processing (step S36) in Examples 1 and 2. Therefore, it is preferable that the simulation process S302 be executed in the simulation processing (step S36) in Examples 1 and 2.

[0102] Here, using Fig. 19, the quantification process by the simulation unit 1033 will be described using the node "environmental awareness" shown in Fig. 17 as an example. This node shown in Fig. 19 has four inputs, so n = 4 in the calculation formula 920 shown as a function in Fig. 19. The simulation unit 1033 determines the index value of this node (S, S indicates the index value at a certain time step t) based on this calculation formula. Here, the initial value S is set to 0.

[0103] However, if the initial value is known in advance through a questionnaire or the like, a value other than 0 can be specified as the initial value. The simulation unit 1033 executes the processing of this calculation formula for the nodes in the outcome / social impact layer at a certain time step to calculate the index value of each node. Furthermore, the simulation unit 1033 advances the time step over time and updates the index value of each node (St+1, which is the index value for time step t+1 after time step t). The simulation unit 1033 then repeatedly processes the time step a predetermined number of times to obtain the index value of each node after the predetermined number of times. In this example, the time step is set to one month, which corresponds to one year, so the predetermined number is 12 times (12 months).

[0104] Furthermore, the simulation unit 1033 can compare the differences in index values ​​between different measures by simulating the same logic model while changing the parameters of the driving nodes. In this way, the simulation unit 1033 can also search for a measure that maximizes the index value. Since the above index values ​​include the social impact index (in this example, power saving acceptance), which is the target node for quantification, this simulation makes it possible to quantify the social impact index. Furthermore, the social impact index quantified in this way can be analyzed by the simulation unit 1033 between different measures. This will be explained in S307 below. Note that S307 may be executed by an evaluation unit separate from the simulation unit 1033.

[0105] Here, a quantification example will be described, showing specific values ​​for this example. In this example, the data required to set the parameters of the driving node are the power generation potential of natural energy and the amount of power consumed by residents. It is also assumed that a power saving request will be made at a certain rate when natural energy is insufficient compared to demand, and that if there is still a shortage after the power saving request, power will be purchased from the electricity market, etc. In this example, the system capacity of the photovoltaic power generation facility, the effective head of the small hydroelectric power generation facility, and the degree of power saving request are set as parameters of the driving node.

[0106] The simulation unit 1033 also estimates the amount of power generated from the scale of each power generation facility, and performs supply and demand calculations in advance from the amount of power consumed by residents, to calculate time-series parameters for the occurrence of power saving requests. This calculation is performed in advance in this embodiment, but it is desirable to perform it during the simulation process S302.

[0107] The parameters of the driving node are shown in Figure 20. As shown in Figure 20, there are 36 possible combinations of measures. For the sake of explanation, a measure number is assigned to each measure in Figure 20. Let measure number 1 be (0kW, 0m, 30%), and measure number 2 be (10kW, 0m, 30%). In other words, the numbers are assigned in the order of solar power, small hydropower, and power saving request level. Also, measure number 36 is (20kW, 2m, 0%).

[0108] 21 and 22 show graphs showing the pre-calculated number of times that action requests occurred for each action number. As shown in FIG. 14, there are four indicators for action requests: number of power-saving requests (weekday, tight) 406, number of power-saving requests (weekday, not tight) 407, number of power-saving requests (holiday, tight) 408, and number of power-saving requests (holiday, not tight) 409. FIG. 21 shows the number of power-saving requests (weekday, tight) 406 and number of power-saving requests (weekday, not tight) 407 per year for each of the 36 different actions. FIG. 22 also shows the number of power-saving requests (holiday, tight) 408 and number of power-saving requests (holiday, not tight) 409.

[0109] It is desirable that the simulation unit 1033 outputs the graphs shown in Figures 11 and 12 to each information terminal using the communication unit 1031. As a result, each information terminal can display the graphs shown in Figures 21 and 22 on the output device 204. In addition, in this process, the simulation unit 1033 uses the weights shown in Figure 16 as the weights of each edge of the quantifiable logic model.

[0110] Next, the simulation results of the simulation process S306 are shown in Fig. 23. Fig. 23 shows the values ​​of each node of the quantifiable logic model, "environmental awareness," "dissatisfaction with power saving requests (negative node)," and the social impact index, "power saving acceptance," for each measure. Since "dissatisfaction with power saving requests" is a negative node, it is shown as a positive value in Fig. 23. However, unlike other indexes, a positive value indicates a worsening trend. Therefore, it is considered to be essentially a negative value.

[0111] Furthermore, looking at the node "Power Saving Acceptance" for policy numbers 1 to 9, we can see that this index is decreasing. On the other hand, the scale of solar and small hydroelectric power generation facilities is increasing for policy numbers 1 to 9. In other words, this logic model shows that as the scale of power generation increases, power saving acceptance decreases. Intuitively, one would think that the introduction of renewable energy would increase power saving acceptance, but the opposite effect was found.

[0112] In the analysis process S307, the simulation unit 1033 analyzes the simulation results for the above. As a result, the simulation unit 1033 identifies that the cause is the "environmental awareness" and "dissatisfaction with the power saving request" nodes in the previous stage. This logic model is based on the subjective assumption that "environmental awareness" increases when a power saving request is made. However, conversely, this suggests that if there is an abundance of power generation facilities, there will be no power shortage and environmental awareness will weaken.

[0113] On the other hand, even a slight lack of power leads to a certain level of dissatisfaction. Therefore, as a result of the analysis process S307, increasing the power generation scale and reducing the number of power saving requests will actually result in a lowering of the social impact index unless the number is zero. Therefore, in this embodiment, when the power saving request level is not 0, the measures that maximize the social impact index are measure numbers 19 and 20, and their respective index values ​​are -19.19. Through the analysis process S307, the simulation unit 1033 searches for such maximum values. Here, these maximum values ​​are relative to each index. Therefore, the simulation unit 1033's search for the maximum value means performing a relative evaluation to build consensus on each measure. Furthermore, the simulation unit 1033 can identify the measure that exhibits the maximum value as the measure that improves the social impact index.

[0114] Then, the simulation unit 1033 uses the communication unit 1031 to output the measures that will improve the social impact index to the second information terminal 102. As a result, the second information terminal 102 can display the measures on the output device 204. Therefore, suggestions for improvement measures can be made to the decision maker. The evaluation unit 1034 may also output the measures that will improve the social impact index to the first information terminal 101. As a result, the first information terminal 101 can display the measures on the output device 204. Therefore, it is also possible to present the above-mentioned suggestions to the participants. In this embodiment, the third information terminal used by the supporter may be connected to the network 105. In this case, it is possible to display the measures that will improve the social impact index on the third information terminal. This concludes the explanation of Example 3.

[0115] Example 4 is an example of a case where a project or company is evaluated when making ESG investments as a measure. Note that in this example, processing is executed with the same configuration as in Example 1. In Example 4, the driving node may be an index related to the project or company that is the investment target. For example, the number of employees in a region or the percentage of the investment that is returned to the region (regional distribution rate) may be used.

[0116] The outcome / social impact layer uses indicators related to the SDGs or those considered important by the investing company. For example, subjective indicators such as CO2 emissions (or reductions) and employee job satisfaction can also be used. The logic model's edge connections and weights can be constructed according to the investing company's standards. The resulting quantifiable logic model is then used to run simulations by varying the parameters of the driving nodes that affect the investment's impact, and the social impact indicators are quantified. This allows the system to propose to investment decision makers the investment business or corporate structure (e.g., investment distribution rate) that will maximize the social impact indicator.

[0117] This concludes the description of the fourth embodiment of the present invention. Note that according to the third and fourth embodiments, it is possible to quantify the social impact index numerically. Furthermore, by taking into account the temporal changes in the subjective index and comparing the numerical values ​​of the social impact index for each of a large number of candidate policies, it is possible to achieve a relative evaluation between policies. This makes it possible to support consensus building on policies.

[0118] 100...Logic model construction support system, 101...First information terminal, 102...Second information terminal, 103...Logic model construction support device, 104...Database, 105...Network, 1031...Communication unit, 1032...Directed graph generation unit, 1033...Simulation unit, 1034...Evaluation unit, 1041...Logic model data, 1042...Social impact index data, 200...Computer, 201...Processor, 202...Storage device, 203...Input device, 204...Output device, 205...Communication IF

Claims

1. A logic model construction support device having a processor that executes a program and a storage device that stores the program, wherein the processor executes, in accordance with the program, a logic model construction process that constructs a directed graph having multiple hierarchies represented by edges connecting nodes together; a selection process that accepts selections for nodes in the constructed directed graph; a modification process that changes weighting coefficients corresponding to the edges connecting to the selected nodes; a simulation process that creates variation information relating to a first quantitative value of a social impact index of the logic model before the weighting coefficient is modified, a second quantitative value of the social impact index of the logic model after the weighting coefficient is modified, and the difference between the first quantitative value and the second quantitative value; and a variation information output process that outputs the variation information.

2. A logic model construction assistance device according to claim 1, wherein the variation information is a variation value indicating the difference.

3. A logic model construction support device as described in claim 1, further comprising a communication interface for connecting via a network to a first information terminal used by a participant in a specified activity and a second information terminal used by a decision maker for said activity, and wherein said directed graph, details of said edges, said first quantitative value, said second quantitative value and said variation information are output on said first information terminal and said second information terminal.

4. A logic model construction support device according to claim 3, wherein the first information terminal outputs an outline of the variation information, and the second information terminal outputs detailed information of the variation information.

5. A logic model construction support device according to claim 1, wherein the processor, in accordance with the program, further executes an output process for outputting a node influence degree indicating the degree of influence on a social impact index at the node.

6. A logic model construction support device as described in claim 1, wherein the directed graph comprises a first group of nodes and a second group of nodes in which parameters related to changes over time are set, the second group of nodes calculates output data for the set parameters, and the first group of nodes inputs output data from the second group of nodes connected via the edges, executes specified processing, and outputs quantified data of a social impact index.

7. A logic model construction support device according to claim 1, wherein the fluctuation information output process outputs the first quantitative value and the second quantitative value.

8. A logic model construction support method using a logic model construction support apparatus having a processor that executes a program and a storage device that stores the program, wherein the processor executes, in accordance with the program, a logic model construction process that constructs a directed graph having multiple hierarchies represented by edges connecting nodes; a selection process that accepts selections for nodes for the constructed directed graph; a modification process that changes a weighting coefficient corresponding to the edge connecting to the selected node; a simulation process that creates variation information relating to a first quantitative value of a social impact index of the logic model before the weighting coefficient is modified, a second quantitative value of the social impact index of the logic model after the weighting coefficient is modified, and a difference between the first quantitative value and the second quantitative value; and a variation information output process that outputs the variation information.

9. A logic model construction support method according to claim 8, wherein the fluctuation information is a fluctuation value indicating the difference.

10. A logic model construction support method as described in claim 8, wherein the logic model construction support device further has a communication interface for connecting via a network to a first information terminal used by a participant in a specified activity and a second information terminal used by a decision maker for the activity, and the directed graph, details of the edges, the first quantitative value, the second quantitative value and the variation information are output on the first information terminal and the second information terminal.

11. A logic model construction support method as defined in claim 10, wherein the first information terminal outputs an outline of the variable information, and the second information terminal outputs detailed information of the variable information.

12. A logic model construction support method as described in claim 8, further comprising the processor executing, in accordance with the program, an output process for outputting a node influence degree indicating the degree of influence on a social impact index at the node.

13. A logic model construction support method as described in claim 8, wherein the directed graph comprises a first group of nodes and a second group of nodes in which parameters related to changes over time are set, the second group of nodes calculates output data for the set parameters, and the first group of nodes inputs output data from the second group of nodes connected via the edges, executes specified processing, and outputs quantified data of a social impact indicator.

14. A logic model construction support method as defined in claim 1, wherein the fluctuation information output process outputs the first quantitative value and the second quantitative value.

15. A logic model construction support device having a processor that executes a program and a storage device that stores the program, wherein the processor executes, in accordance with the program, a logic model construction process that constructs a directed graph having multiple hierarchies represented by edges connecting nodes; a display process that displays the constructed directed graph; a selection process that accepts a selection of a node for the displayed directed graph; a detailed display process that displays details of the selected node and the edges connecting to the node; a modification process that changes a weighting coefficient corresponding to the displayed edge; a simulation process that creates variation information regarding a first quantitative value of a social impact index of a logic model before the weighting coefficient is changed, a second quantitative value of a social impact index of a logic model after the weighting coefficient is changed, and a difference between the first quantitative value and the second quantitative value; and a variation information display process that displays the variation information.

16. A logic model construction support system having a first information terminal used by a participant in a specified activity, a second information terminal used by a decision maker for the activity, and a logic model construction support device, wherein the logic model construction support device has a processor that executes a program and a storage device that stores the program, and the processor executes, in accordance with the program, a logic model construction process that constructs a directed graph having multiple hierarchies represented by edges connecting nodes together, a selection process that accepts selections for nodes for the constructed directed graph, a modification process that changes a weighting coefficient corresponding to the edge connecting to the selected node, a simulation process that creates variation information relating to a first quantitative value of a social impact index of the logic model before the weighting coefficient is modified, a second quantitative value of the social impact index of the logic model after the weighting coefficient is modified, and the difference between the first quantitative value and the second quantitative value, and a variation information output process that outputs the variation information.