Policy effectiveness estimation system and policy effectiveness estimation method

Abstract

To estimate effect of a measure for reducing a consumption power amount.SOLUTION: Provided is a measure effect estimation system, comprising: a request reception unit for receiving a request for estimating effect of a measure; a model creation unit for creating a model based on information relating to an effectiveness indicator during a second period included in a first period during which the effect of the measure is estimated, work information of a target facility, and measure management information; a measure analysis dedicated data generation unit for, based on information relating to an effectiveness indicator and work information at an existing facility, information relating to the effectiveness indicator and work information of the target facility during the second period, and information relating to the measure, generating information relating to the effectiveness indicator and work information during the first period; and a measure effect prediction unit for estimating the effect of the measure, using the information relating to the effectiveness indicator and work information during the first period and the model, based on information relating to the effectiveness indicator in a case when the measure is implemented and a case when the measure is unimplemented during the first period.SELECTED DRAWING: Figure 1

Description

[[Technical Field]]

[0001] The present invention relates to a measure effect estimation system and a measure effect estimation method for estimating the effect of a measure in a facility such as a store. [[Background Art]]

[0002] In order to reduce power consumption in businesses and households, it is desirable to be able to analyze current power consumption, estimate the effect of various measures before introducing them, and introduce highly effective measures. Therefore, in a short period before the introduction of the measure, power data, business operation data of companies, operation data of households, and other data serving as external factors that affect power consumption are collected. Based on the collected short-term data, the amount of power when no measure is implemented and when the measure is implemented is estimated, and the amount of power that can be reduced is calculated. Here, business operation data of companies, household operation data, and data on external factors are collectively referred to as business data. Since power consumption may be affected by seasons and other factors, it is assumed that estimation of the effect of power reduction achieved by a measure over a long period, for example, one year, is performed using power data and business data for more than the past one year.

[0003] Conventionally, Patent Document 1 and Patent Document 2 are known as techniques for predicting the effect of a measure.

[0004] Patent Document 1 discloses an energy prediction system including: an analysis condition setting unit that sets a comparison time and a comparison number of days for performing cluster analysis and selects a comparison target from classified load patterns; a time-series data setting unit that sets current day time-series data and past time-series data based on actually measured values classified into the selected load patterns; and a predicted value calculation unit that selects a plurality of pieces of past time-series data with high similarity, weights and adds the actually measured values from a predetermined time of the selected past time-series data to a time a predetermined prediction time ahead according to the similarity, and calculates a predicted value of an energy load from the predetermined time of the prediction day to a prediction time a prediction time ahead.

[0005] Patent Document 2 discloses a data integration analysis system that automatically generates a large number of explanatory variables by pre-defining three operators—condition, target, and operation—which are variable generation conditions, for data input to the analysis system. [Prior art documents] [Patent Documents]

[0006] [Patent Document 1] International Publication No. 2016 / 136323 [Patent Document 2] Japanese Patent Publication No. 2014-081750 [Overview of the project] [Problems that the invention aims to solve]

[0007] For example, a company may consider implementing measures to reduce electricity consumption at a particular facility (hereinafter, this facility will be referred to as a store, for example). To predict the effectiveness of these measures, it is necessary to estimate the amount of electricity consumption that can be reduced by these measures, which requires long-term electricity data and operational data used to estimate the effectiveness of each measure. Therefore, it becomes necessary to estimate the effectiveness of the measures over a long period, for example, one year, when only short-term measurement data has been collected before the implementation of the measures, for example, about one month.

[0008] Conventionally, the technology disclosed in Patent Document 1 requires historical data from the same equipment when predicting annual power consumption. The technology disclosed in Patent Document 2 requires the use of data from the own store when managing POS data, behavioral data, and customer information data, and creating explanatory variable data by defining variable generation conditions in advance.

[0009] Before implementing a measure, if the actual measurement data for the target store is only available for a short period compared to the period over which the effects of the measure are to be estimated, one possible approach is to use performance data from other stores. When using electricity and operational data from other stores, it is necessary to use data from the appropriate store, as electricity consumption trends and the degree to which operational data influence electricity consumption differ from store to store.

[0010] The present invention aims to enable the estimation of the effectiveness of measures to reduce power consumption by utilizing performance data from other stores, even when only short-term measurement data is available for a target store where the measures are to be implemented. [Means for solving the problem]

[0011] A typical example of a means for solving the problems of the present invention is as follows: In other words, a policy effectiveness estimation system for estimating the effectiveness of a policy regarding effectiveness indicators in a facility, comprising a computer having a calculation unit that performs calculation processing and realizes the following functional units, and a storage unit accessible by the calculation unit, comprising: a request receiving unit that receives requests for estimation of the effectiveness of the policy in a target facility where the policy is implemented; a model creation unit that creates a model for performing analysis necessary for the policy in the target facility based on information regarding the effectiveness indicators of the target facility in a second period included in a first period for estimating the effectiveness of the policy, business information of the target facility, and policy management information regarding the policy; a policy analysis data generation unit that generates information regarding the effectiveness indicators of the target facility in the first period based on information regarding the effectiveness indicators of an existing facility and business information, information regarding the effectiveness indicators of the target facility in the second period, and information regarding the policy; and using the information regarding the effectiveness indicators of the target facility in the first period and business information of the target facility in the first period, and the model, calculates information regarding the effectiveness indicators of the target facility when the policy is implemented and when it is not implemented in the first period, and based on the calculated information regarding the effectiveness indicators of the target facility when the policy is implemented and when it is not implemented in the first period, By using information and operational information regarding the performance indicators of existing facilities to supplement the information and operational information regarding the performance indicators of the target facilities during the second period, A policy effect estimation system characterized by having a policy effect prediction unit that estimates the effect of the aforementioned policy. [Effects of the Invention]

[0012] According to the present invention, even if only short-term measurement data is available for a target store where the effectiveness of a measure is to be estimated, it becomes possible to estimate the long-term effectiveness of the measure by using data collected and accumulated at other stores. [Brief explanation of the drawing]

[0013] [Figure 1] This figure shows an example of the configuration of a policy effectiveness estimation system. [Figure 2] This figure shows an example of a hardware configuration for running a policy effectiveness estimation system. [Figure 3] It is a diagram showing an example of an overall flowchart. [Figure 4] It is a diagram showing a configuration example of a measure management table. [Figure 5] It is a diagram showing an example of a flowchart for power data creation processing. [Figure 6] It is a diagram showing a configuration example of a performance data management table. [Figure 7] It is a diagram showing a configuration example of a store attribute management table. [Figure 8] It is a diagram showing an example of a flowchart for business data creation processing. [Figure 9] It is a diagram showing an example of an input / output screen of a measure effect estimation system. [Figure 10] It is a diagram showing a configuration example of data managed in a graph database of a measure effect estimation system. MODES FOR CARRYING OUT THE INVENTION

[0014] Example 1 FIG. 1 is a conceptual diagram of a measure effect estimation system 101 according to Example 1 of the present invention. In the following description, a measure refers to a measure introduced for reducing power consumption at a certain facility (hereinafter, the facility is referred to as a store, as an example). Further, a computer used by a user of the measure effect estimation system 101 is a client 102.

[0015] The policy effectiveness estimation system 101 estimates the effectiveness of policies regarding effectiveness indicators at a facility. The policy effectiveness estimation system 101 operates on a server. The policy effectiveness estimation system 101 has the following parts. Specifically, the policy effectiveness estimation system 101 has a request reception unit 103 that receives requests for policy effectiveness estimation from clients 102, i.e., requests for estimating the effectiveness of a policy at a facility where a certain policy is implemented (hereinafter, "requests for policy effectiveness estimation" will be simply referred to as "requests"). Furthermore, the policy effectiveness estimation system 101 has a model creation unit 104 that creates a model for performing the analysis necessary for the policy in the request (hereinafter, "policy in the request" will be simply referred to as "policy") using the actual measurement data from the target store read in the request. Furthermore, the policy effectiveness estimation system 101 has a model management unit 105 that saves and manages the created model. Furthermore, the policy effectiveness estimation system 101 has a performance data management unit 106 that manages power data and various business data from existing stores where the policy has already been implemented. Furthermore, the policy effect estimation system 101 has a management database 107 that manages information such as information related to policies and information related to stores. Furthermore, the policy effect estimation system 101 has a policy analysis data generation unit 108 that generates data for policy effect estimation. Furthermore, the policy effect estimation system 101 has a policy effect prediction unit 109 that predicts the amount of electricity after the policy is implemented using the policy effect estimation data. Furthermore, the policy effect estimation system 101 has an economic effect calculation unit 110 that calculates the economic effect of the policy based on the policy effect prediction. Furthermore, the policy effect estimation system 101 has an interface 111 that provides an interface between the policy effect estimation system 101 and the outside of the policy effect estimation system.

[0016] The facilities that are the subject of estimation for the effects of a policy are simply called target facilities. If the facility is a store, the target facility is called a target store. While target facilities are those that will be the target of the policy in the future, existing facilities, including those that have already been subjected to the policy, are called existing facilities. If the facility is a store, the existing facilities are called existing stores.

[0017] Regarding a request from the client 102, information included in the request such as actually measured power data at a target store, which is the store targeted for estimating the effect of a measure, is input via the interface 111. The data input here is stored in a memory or a secondary storage device. On the other hand, as a different configuration, a target store actually measured data management unit 112 that manages actually measured data at the target store may be provided. That is, the target store actually measured data management unit 112 may be provided separately from the group of the measure effect estimation system 101 shown in FIG. 1. Even in this case, the measure effect estimation system 101 may include the target store actually measured data management unit 112. Further, in the request, an access destination to the actually measured data at the target store, for example, a URL (Uniform Resource Locator), may be specified, and the measure effect estimation system 101 may read the actually measured data via the request receiving unit 103 in accordance with the specification. The target store actually measured data management unit 112 may be provided in the measure effect estimation system 101. The target store actually measured data management unit 112 manages power data and various types of business data before the measure is introduced at the target store. The acquisition period of the actually measured power data may be a predetermined period, for example, one month. The power data and the business data may be data in the same period.

[0018] For the power data at each of the target store and existing stores, power may be measured by a power sensor provided in a main power device of the store to generate the power data of the store. Alternatively, power may be measured by an individual power sensor provided for each electric device used in the store (hereinafter, "electric device" may be simply referred to as "device"), to generate power data related to the store. When a power sensor provided in the main power device is used, power data separated for each device may be generated by using a method (called a separation model) of separating main power into power for each device via the main power device. Here, an algorithm such as a factorial hidden Markov model or a neural network may be used for the separation model.

[0019] Measures to reduce electricity consumption in stores may include equipment-specific measures such as "reviewing the on / off operation of cooking equipment" or "reviewing the operation of air conditioning." Here, "reviewing the on / off operation of cooking equipment" could involve reviewing the timing of turning cooking equipment on or off, or the duration of being on or off. Similarly, "reviewing the operation of air conditioning" could involve reviewing the operation control of air conditioning equipment, such as turning it on or off according to the outside temperature, the inside temperature, and the set temperature of the air conditioning. In this case, the data necessary to estimate the effectiveness of the measures could include, for example, in the case of reviewing the on / off operation of cooking equipment, data on the electricity consumed by cooking equipment, and operational data such as the number of items cooked for each type of food using the store's cooking equipment.

[0020] Figure 2 shows an example of a hardware configuration in which the policy effectiveness estimation system 101 is configured using a server 200. The server 200 may have one or more central processing units (CPUs) 201, memory 202, secondary storage devices 203 such as hard disks, input / output interfaces 204 that control input from a keyboard and mouse and output information to a display, and a network interface 205 that connects to a network. Each part of the policy effectiveness estimation system 101 is loaded into the memory 202 as a computer program and executed by the CPU 201. In addition, the management data 206 used by the policy effectiveness estimation system 101, as well as power data and business data 207 may be stored in the secondary storage device 203. The power data and business data 207 correspond to the actual data management unit 106 in Figure 1. The management data 206 corresponds to the management database 107 in Figure 1. The management data 206 and the power data and business data 207 are stored not within the server 200, but in an external storage device connected to the server 200 via a network. Note that Server 200 will be implemented as a virtual machine, not a physical machine.

[0021] Figure 3 is a flowchart showing an example of the overall processing of the policy effect estimation system 101. This flow starts when a request is received from client 102 to estimate the policy effect for a target store. This request includes the names of one or more policies whose effects are to be estimated, attribute information of the target store, actual data such as electricity data for the target store before the policy is implemented, or a destination for accessing such actual data.

[0022] The model creation unit 104 creates a model for performing the analysis necessary for the measures at the target facility, based on information regarding the effectiveness indicators of the target facility during the second period included in the first period for estimating the effectiveness of the measures, operational information related to the target facility, and policy management information related to the measures handled by the policy effectiveness estimation system 101.

[0023] Here, the first period for estimating the effectiveness of the measures may be the period for estimating the effectiveness of the measures as specified in the requirements, for example, one year. The target facility may be the facility for which the effectiveness of the measures is to be estimated, and in the embodiment, it is described as the target store. The effectiveness indicator is an element that serves as an indicator for judging the effectiveness of the measures, and in the embodiment, it is described as power consumption. Information regarding the effectiveness indicator is described in the embodiment as power data actually measured. The target facility may be the facility for which the effectiveness of the measures is to be estimated, and in the embodiment, it is described as the target store. The business information may be data related to operations at the facility, data related to the operation of equipment, data related to external factors, etc., which are necessary for estimating the effectiveness of the measures, and in the embodiment, it is described as business data. The second period may be the period for obtaining information regarding the effectiveness indicator of the target facility. The measure management information regarding the measures may be information about the measures necessary for estimating the effectiveness of the measures, and in the embodiment, it is described as the measure management table 400.

[0024] In this flowchart, for the measures requested by client 102, a model is created that performs the analysis necessary for the measures at the target store, for example, an operational state estimation model corresponding to the measures, by referring to the measure management table 400 and using actual measurement data from the target store. The measure management table 400 may be one that pre-stores the information necessary to estimate the effect of each measure for which the measure effect estimation system 101 is the target. A specific example of the measure management table 400 will be described later.

[0025] In the flowchart shown in Figure 3, power data and operational data are read from actual measurement data at the target store specified in the request (S301), and the analysis logic 403 for the measure is executed to create a model necessary to evaluate the specified measure (S302). For example, if the measure is "review of on / off operation of cooking equipment," a cooking equipment operation state estimation model and a cooking equipment operation optimization model are created. For example, in the cooking equipment operation state estimation model, cooking time and number of items cooked are used as operational data from cooking records to estimate the operating state of the cooking equipment, such as cooking or keeping warm, at each point in time in the power data. In the cooking equipment operation optimization model, data such as the number of items cooked per product, the cooking temperature of the cooking equipment, the time and power consumption required for preheating, and the power consumption during keeping warm are used to output the number of items cooked per batch, cooking timing, and power on and off timing that minimize the power consumption of the cooking equipment. For the model that performs the analysis necessary for the measure at the target store, the basic model may be a pre-prepared program, for example, stored in the model management unit 105. The basic model may be prepared according to various measures. The basic model for this can be one in which the operation of the object to be evaluated or analyzed in a particular measure is virtually realized by referring to the rated power of the object to be evaluated or analyzed in that measure, for example, for each type of equipment, and then the operation of that object is analyzed or evaluated.

[0026] Here, the operating state estimation model may be a program that estimates the operating state of equipment used in the target store using power data, which is the actual measured value of the target store, and business data corresponding to that data. For example, the basic model for the cooking equipment operating state estimation model described above may be a program that virtually realizes the operation of cooking equipment according to given conditions and calculates its power consumption. In step S302, a basic model corresponding to the specified measure is selected, and the power data and business data read in step S301 are set in that basic model to create a model that performs the analysis necessary for the measure at the target store. For example, when creating a cooking equipment operating state estimation model, a basic model corresponding to the cooking equipment operating state estimation model is selected from various basic models. The cooking equipment operating state estimation model is created by reading data related to the cooking equipment used in the target store, such as the rated power when the cooking equipment is on and the rated power when it is kept warm, from the aforementioned power data and business data and setting it in this basic model. The processing in steps S301 and S302 may be performed in the model creation unit 104.

[0027] The policy analysis data generation unit 108 creates information on effectiveness indicators and operational information for the target facility corresponding to the first period, based on information on effectiveness indicators and operational information for the existing facility, information on effectiveness indicators and operational information for the target facility during the second period, and information on policies. Here, the information on effectiveness indicators and operational information for the existing facility may be information on effectiveness indicators and operational information for the existing facility when policies are implemented and when they are not.

[0028] The policy analysis data generation unit 108 calculates the correlation between the target facility's effectiveness indicators and the target facility's operational information, which is the correlation between the target facility's effectiveness indicators and the target facility's operational information. It also calculates the correlation between the existing facility's effectiveness indicators and the existing facility's operational information, and uses the correlation between the target facility's effectiveness indicators and the existing facility's effectiveness indicators to create information about the target facility's effectiveness indicators and operational information corresponding to the first period.

[0029] The effectiveness indicator can be electricity consumption, and the operational information may include time-series data. Within the operational information, information regarding equipment whose effectiveness is to be estimated in a policy can be time-series data collected from the equipment.

[0030] The policy analysis data generation unit 108 calculates the degree of similarity between information on evaluation indicators for existing facilities and information on evaluation indicators for target facilities. For existing facilities where the degree of similarity is higher than or equal to a predetermined threshold, it determines the correlation between the information on evaluation indicators and operational information, and creates information on effect indicators and operational information for the target facilities corresponding to the first period. These processes are described in steps S803, S804, S304, and S305.

[0031] The policy analysis data generation unit 108 supplements the business information of the target facility for the period from the first period excluding the second period with business information of existing facilities for which the absolute value of the difference between the correlation between the information on evaluation indicators for the target facility and the business information, and the correlation between the information on evaluation indicators for existing facilities and the business information, is smaller than or equal to a predetermined threshold. These processes are described in steps S801 to S809.

[0032] Next, the created model is saved to the model management unit 105 (S303).

[0033] Next, based on the actual measurement data of the target stores that were loaded, power data corresponding to the period for which the effect estimation is to be performed is created (S304). An example of the power data creation process (S304) will be described later with reference to Figure 5.

[0034] Next, business data corresponding to the period for which the effects of the measures are estimated is created (S305). An example of this business data creation process (S305) will be described later with reference to Figure 8. Here, the logic and business data used to analyze the measures may be managed in the measure management table 400. The measure management table 400 may be stored in the management database 107.

[0035] Figure 4 shows an example of the configuration of the policy management table 400. The policy management table 400 has a policy 401 field which manages the policy name that identifies the policy, a device type 402 field which manages the type of device targeted by the policy, and an analysis logic 403 field which defines the logic for performing the analysis. Furthermore, it has a business data type 404 field which manages the type of business data necessary to execute the analysis logic 403, an analysis parameter 405 field which manages the parameters used for analyzing the policy, and an effect estimation parameter 406 field which manages the parameters used when estimating the amount of the policy's effect as an economic effect.

[0036] The aforementioned analysis logic may be the analysis method for a policy as defined in analysis logic 403 in policy management table 400, and this analysis method may be a program. For example, if the policy is "review of power on / off operation of cooking equipment," the analysis logic may be a "cooking equipment operating state estimation" logic that processes power data to estimate the operating state of the target equipment, such as the preheating state or operating state of the equipment. Furthermore, the analysis logic may also include various optimization logics, such as cooking equipment operation optimization logic, to review operations, such as turning off the power if the preheating time is long for the estimated operating state, and these logics may be managed. The analysis logic should include both elements that can be common across various policies and elements that are individually tailored to each policy. For example, the "cooking equipment operating state estimation" logic can be common across various cooking equipment policies.

[0037] The operational data used in these measures is time-series data, created from sensor recordings, operational records, POS data, etc. Examples include the number of units cooked and sold of product A for each product prepared using cooking equipment, and outside temperature data related to air conditioning can also be used as operational data.

[0038] Figure 5 shows an example flowchart of step S304, which involves creating power data for the target stores during the period for which the effect estimation is to be performed, in order to estimate the effect of the measure. The process in this flowchart may involve creating power data for the equipment targeted by the measure for the period during which the target stores do not have the actual measurement data necessary for the effect estimation. Here, the period for which the effect estimation of the measure is to be performed may be defined by default in the system, or the client 102 may input an arbitrary period in the request. For example, the period for which the effect estimation of the measure is to be performed may be one year. In the process in this example flowchart, the data for the period during which the target stores do not have the actual measurement data necessary for the effect estimation may be supplemented by using the performance data management unit 106 that has already acquired performance data of existing stores managed by the performance data management unit 106. For the performance data of existing stores, data before and after the introduction of the measure may be collected at the existing stores and managed by the performance data management unit 106.

[0039] Figure 6 shows an example configuration of a performance data management table 600 for managing performance data at existing stores. The performance data management table 600 manages information about all existing stores that have already been acquired. The performance data management table 600 is stored in the performance data management unit 106. Furthermore, this performance data may be managed for each existing store. Figure 6(a) shows an example of a management table for electricity data at an existing store, and Figure 6(b) shows an example of a management table for business data at an existing store. The information stored and managed in the performance data management table 600 is various information and data related to the target stores for which the policy effectiveness estimation system 101 handles policies and estimates the effects of policies. In other words, it may be information and data from when the same policy as the policy effectiveness estimation system 101 is expected to handle is implemented at an existing store. Alternatively, it may be information and data from when the same policy is implemented at an existing store that is similar to the target stores that the policy effectiveness estimation system 101 is expected to handle. Furthermore, information and data such as the type of equipment, model of equipment, sampling interval of acquired data, and type of business data that are expected to be used in the target stores that the policy effect estimation system 101 is expected to handle may also be used.

[0040] The example power data management table shown in Figure 6(a) includes a field for store 601, which indicates the name that identifies the store; a field for start date and time 602, which sets the start time of data acquisition; and a field for latest update date and time 603, which sets the latest acquisition time. Furthermore, it includes a field for equipment type 604, which identifies the equipment targeted by the measure; a field for model 605, which indicates the model of the target equipment; a field for data 606, which indicates the name that identifies the data to be read when reading power data; and a field for sampling interval 607, which sets the sampling interval for power data. If there are multiple target equipment of the same type and power data is acquired separately for each equipment, they may be registered individually. However, if power data is not acquired separately for each equipment, they may be grouped by equipment type and registered as a single power data entry. In that case, the model of each equipment may be registered in the model field.

[0041] The example business data management table shown in Figure 6(b) includes a store field 601 which indicates the name that identifies the store, a business data type field 608 which sets the type of business data, a start date and time field 609 which sets the start time of data acquisition, a latest update date and time field 610 which sets the latest acquisition time, and a data field 611 which identifies the data when it is read. As shown in Figure 6(b), business data type 608 may include time-series data, as exemplified by the latest update date and time field 603.

[0042] In the flowchart shown in Figure 5, step S501 executes a process to extract candidate existing stores in order to supplement the power data of the target store using actual data from existing stores for periods when actual measurement data for the target store is unavailable. That is, stores with similar attributes to the target store from among existing stores from which actual data has already been collected are extracted as candidates for the existing stores to be used for the supplementation. The extraction of the aforementioned candidates may be performed using the information in the store attribute management table 700 exemplified in Figure 7. Furthermore, the attribute information of operations for each store may be the business procedures of the company to which the store belongs. In addition, the attributes of the equipment used in each store may be the model number of the equipment, etc., as attribute information related to the equipment. Information on the building conditions and equipment configuration of existing stores may be read from the store attribute management table 700 exemplified in Figure 7, and clustering may be performed between the target store and existing stores based on the distance between the data related to store attribute information, using the information of the target store obtained in response to the request from client 102. Here, the information on building conditions and equipment configurations of existing stores can be read from all existing stores, or from dependent stores selected as appropriate. As a result of the clustering process, existing stores that belong to the same cluster as the target store are identified, and the distance to the target store is calculated as a degree of similarity, for example, similarity score. Here, the similarity score can be a positive value, with a value closer to 0 indicating a higher degree of similarity. That is, stores with a similarity score within a predetermined threshold may be identified as candidates. Alternatively, n stores with the smallest similarity scores may be identified as candidates. Here, the integer n may be arbitrarily determined by client 102. In the above explanation, we described similarity, where the degree of similarity is represented by a positive value, with a value closer to 0 indicating a higher degree of similarity. Generally speaking, the higher the degree of similarity between the two pieces of information being compared, in other words, the greater the value, the more similar the two pieces of information can be treated.

[0043] Alternatively, as a method for searching for stores with similar attributes, different from the method described above, one may calculate the similarity of building conditions, the similarity of equipment configuration and its attributes, and the similarity of business attributes separately using clustering processing, and then add up these similarities to identify stores that are similar, i.e., stores with a small sum of similarity.

[0044] Figure 7 shows an example of a store attribute management table 700. The store attribute management table 700 stores information about the attributes of each existing store. The store attribute management table 700 is preferably stored in the management database 107. The store attribute management table 700 shown in Figure 7(a) manages store 701, which is a field indicating the name that identifies the store, company 702, which is a field that identifies the company to which the store belongs, and building conditions 703, which is a field that manages conditions such as the location and area of ​​the store, as attribute information of an existing store. The store attribute management table 700 shown in Figure 7(a) may further manage equipment configuration 704, which is a field that manages the equipment configuration such as air conditioning and cooking equipment in the store. The store attribute management table 700 shown in Figure 7(b) shows an example of attributes related to business operations, and for company 702, which is a field that identifies the company to which the store belongs, it manages, for example, cooking equipment operating standards 705 that specify the operating standards for cooking equipment and air conditioning operating standards 706 that specify the operating standards for air conditioning equipment, as equipment-specific operating standards corresponding to the company. The store attribute management table 700 shown in Figure 7(c) illustrates an example of equipment attributes. For example, the table may manage the model number 707, which is a field indicating the model number that identifies the equipment, and the equipment type 708, which is a field indicating the function of the equipment, as well as the rated power 709, which indicates the rated power of the equipment. Furthermore, for existing stores, the store attribute management table 700 may manage the implemented measures, their implementation dates, and the settings of the analytical parameters used in those measures.

[0045] In the flowchart shown in Figure 5, step S502, following step S501, determines whether a candidate existing store has power data after the implementation of a measure that matches the equipment type and sampling interval of the measure specified in the request. If it does not, the store is removed from the candidates. In this determination, the equipment type 604 and sampling interval 607 of the existing store may be referred to. Next, it is determined whether a candidate existing store has data from before the implementation of the measure specified in the request at that existing store, by referring to the start date and time of data acquisition in the actual data management table 600. If there is no data from before the implementation of the measure, the store may be removed from the candidates (S503).

[0046] Next, for the target store that was accepted in the request, data for the same period in the past is extracted from the power data of candidate existing stores managed by the performance data management unit 106 for the period during which measured power data exists (S504). Here, the acquisition year of the same period in the past does not matter, only the date and time need to be the same. For example, the period for estimating the effect of the measure in the request from client 102 is one year, from March of one year to February of the following year, and the specified period for acquiring new power data at the target store is, for example, one month, and the power data acquired at the target store accepted in the request is, for example, data from March. In this case, the data for the same period in the past extracted in S504 can be the data from March, before the measure was introduced at that existing store. Here, the acquisition year does not matter.

[0047] Next, the degree of similarity, or similarity score, is calculated between the extracted candidate existing store power data and the target store power data (S505). For calculating this similarity score, methods such as using the Euclidean distance, which is calculated based on the distance between corresponding points at time for both waveforms, or dynamic time warping, which calculates the shortest path by exhaustively determining the distance to each point in both waveforms, may be used. The calculated similarity score is a positive value, and the closer the value is to 0, the higher the degree of similarity. In step S505 described above, the similarity score is calculated between the extracted candidate existing store power data, which is information on the evaluation indicators of existing facilities, and the target store power data, which is information on the evaluation indicators of the target facility. In other words, step S505 is an example of using similarity as an indicator to represent the degree of similarity. When similarity is used as an indicator to represent the degree of similarity, that is, when the Euclidean distance method or dynamic time warping described above is used to calculate the degree of similarity, the value is positive, and the closer the value is to 0, the higher the degree of similarity. However, generally speaking, the greater the degree of similarity between the two pieces of information being compared, in other words, the more similar the two pieces of information can be treated. This is also true in describing embodiments for carrying out the present invention.

[0048] Next, based on the similarity calculated in step S505, "existing stores with highly similar data are identified," that is, if the degree of similarity is equal to or higher than a predetermined threshold, or in other words, if the similarity is less than or equal to a pre-set threshold, existing stores are identified in order of the highest degree of similarity (S506). It should be noted here that this is based on the similarity calculated in step S505, that is, in this case, the similarity is determined by the method using the Euclidean distance or the dynamic time stretching method described above. In other words, generally, in step S506, it is sufficient to use the degree of similarity calculated in step S505, and to identify existing stores in order of the highest degree of similarity.

[0049] Next, from the performance data of the identified existing stores managed by the Performance Data Management Unit 106, electricity data is obtained for the period during which the effectiveness of the measures will be estimated, for example, one year, excluding the month in which electricity data exists for the target store (March in the example above), specifically for the period from April to February of the following year (S507). The acquisition of electricity data here may be done at predetermined intervals, for example, every month. Alternatively, this data may be obtained from the Performance Data Management Unit 106 for each period. That is, for example, if electricity consumption is recorded in minute increments, for example, data for April may be obtained from "4 / 1 00:00:00 to 4 / 30 23:59:00". In some cases, the performance data of the initially targeted existing stores may not contain enough data to cover a full year. Therefore, it is determined whether or not monthly data for the period during which the effectiveness will be estimated is available, including the month in which actual measurement data is available (S508). If the data is not available, the process returns to step S506 and selects another store from among the existing stores with similar data to obtain the power data for the missing period (S507). For example, if the first existing store has power data for April to September, then combined with the actual measurement data from the target store, power data for March to September will be available, but power data for October to February of the following year may be missing. In that case, the process of determining whether the next existing store has power data for October to February of the following year as actual data and obtaining the missing data if available (S507) is repeated.

[0050] Once data for the period to be estimated is available, the monthly data is combined to create power data for the estimation period, for example, one year (S509). After that, the process of creating power data for the target equipment for the estimation period, for example, one year, is terminated.

[0051] Figure 8 is a flowchart of an example of the process for creating business data in step S305. In step S305, processing may be carried out using the power data created in step S304 and the business data necessary to estimate the effect of the measures. The business data necessary to estimate the effect of the measures is necessary to supplement the period during which actual measurement data for the target store is insufficient by using the performance data of existing stores.

[0052] In step S305, the first step is to identify the types of business data necessary to estimate the effectiveness of the measures from the measure management table 400 (S801).

[0053] Next, using the actual electricity measurement data for the target store received as a request, a correlation coefficient between the electricity data and the business data for the target store is calculated (S802) as one of the indicators showing the degree of influence of each business data on the electricity data. Here, the correlation coefficient between the electricity data and the business data for the target store may be the same as the business correlation index for the target facility mentioned above. The business data used to calculate the correlation may be all the target business data, or it may be arbitrarily selected by client 102. For example, the correlation coefficient between electricity data and "number of units of product a cooked", and the correlation coefficient between electricity data and "number of units of product b cooked" may be calculated. The business data used to calculate this correlation coefficient may be arbitrarily selected, or it may be all the business data. Here, when calculating the correlation coefficient with the business data using the loaded electricity data, the electricity data may be converted into energy consumption data. Since the electricity data is an instantaneous value measured at each sampling interval, the average of the electricity values ​​over a certain period (e.g., 1 hour) may be calculated, and the calculated electricity may be converted into energy consumption data by assuming that it was used continuously for 1 hour. Here, if the data interval for business data is one hour, the correlation coefficient is calculated by matching the intervals by using one-hour intervals for the electricity consumption data.

[0054] Next, for existing stores that have power data that is highly similar to the power data of the target store identified in the power data creation process (S304), the business data for each store is read from the performance data management unit 106 (S803). Here, business data for the same period as when the power data was extracted in step S504 may be read.

[0055] Next, for the business data read in step S803, the correlation coefficient between the power data and business data of the existing store is calculated (S804) as one of the indicators showing the degree of influence of each business data on the power data. Here, the correlation coefficient between the power data and business data of the existing store corresponds to the business correlation index of the existing facility's effectiveness indicator mentioned above. Furthermore, the business data used to calculate the correlation coefficient may be all the target business data, and may be arbitrarily selected by the client 102. In addition, the correlation coefficient between the power data, which is the actual data for each existing store, and each business data may be calculated monthly at the time of registering both data in the actual data management unit 106, and the store, calculation month, business data, and correlation coefficient may be managed in a management table.

[0056] Next, the business data type is selected (S805). This simply involves selecting the business data type necessary to supplement the power data of the target store. For example, in the actual data management table 600, an appropriate business data type is selected from business data types 608 corresponding to the store in store 601.

[0057] Next, the absolute value of the difference between the correlation coefficient of the target store calculated in step S802 and the correlation coefficient of the candidate existing store calculated in step S804 is calculated, and the existing store with the smallest calculated value, i.e., the closest degree of correlation, is identified (S806). Here, judging by the absolute value of the difference in correlation coefficients means that whether the difference is positive or negative is treated the same. Alternatively, existing stores whose absolute value of the difference is smaller than or equal to a predetermined threshold may be selected, and further identification may be made from these selected existing stores.

[0058] Next, for the period over which the effect estimation is to be performed, for example one year, business data is obtained from the performance data of the identified existing stores for the period other than the month in which the relevant business data for the target store is available (March in the example above) (April to February of the following year in the example above), i.e., the period during which the data is missing (S807). Here, the acquisition of business data may be performed at predetermined intervals, for example, every month. Monthly data may also be obtained from the performance data management unit 106.

[0059] Next, we will explain step S808. Initially, the performance data of the existing stores used for supplementation may not have all the business data corresponding to the period for which the effect estimation is to be performed, for example, one year. Therefore, for the period for which the effect estimation is to be performed, it is determined whether all the relevant business data is available for all months, including months for which actual measurement data exists at the target store (S808). If it is not available, return to step S806, select the existing store with the next smallest absolute value of the difference in the calculated correlation coefficient, and repeat the acquisition of the relevant business data for the missing period until all the data for the period for which the effect estimation is to be performed is available. Furthermore, the processing from S806 to S808 may be repeated for all the business data necessary for estimating the effect of the measure.

[0060] Next, the process is terminated when all the business data corresponding to the period for which the effect estimation is to be performed is available (S809).

[0061] The processing in steps S304 and S305 is performed by the policy analysis data generation unit 108.

[0062] The policy effect prediction unit 109 uses information on effect indicators and operational information corresponding to the first period for the target facility, as well as an operational state estimation model, to calculate information on effect indicators for the target facility when the policy is implemented and when it is not (i.e., before and after the introduction of the policy) during the first period, and estimates the effect of the policy based on the calculated information on effect indicators for the target facility when the policy is implemented and when it is not during the first period.

[0063] In the flowchart shown in Figure 3, following the creation of business data in step S305, the power consumption of the target store before the implementation of the measures is estimated (S306) using the power data and business data for the target store over a period of effect estimation, for example, one year, created in steps S304 and S305. In this estimation of power consumption before the implementation of the measures, the operating state estimation model created from actual measurement data of the target store, which was created in step S302 and managed by the model management unit 105, may be used. In step S306, for example, in the case of a measure to review the on / off status of cooking equipment, the operating state may be estimated using the operating state estimation model for cooking equipment and the necessary business data for that month, for example, monthly power data created in S304, and the power consumption of that cooking equipment may be estimated. The operating state estimation of equipment may be estimated using data to which the operating state has been assigned to the data for each time point. The duration of the operating state of the equipment may be calculated from the power consumption at each time point in the power data and the data on the operating state of the equipment corresponding to each time point. In step S306, the power consumption of the equipment can be calculated by multiplying the calculated duration by the power for each operating state based on the rated power corresponding to the equipment type and adding them together. The rated power corresponding to the equipment type can be defined in advance as a parameter, or defined in the attributes related to the equipment as exemplified in Figure 7(c) in the store attribute management table 700. For example, for cooking equipment, the power consumption / month can be set to "(total preheating time / month × preheating power + total warming time / month × warming power + total cooking time / month × cooking power)". The power consumption can be calculated similarly for all months during the period for which the effect estimation is performed, for example, one year. Alternatively, the power consumption before the introduction of the measure can be calculated by summing the power consumption for each month based on the created annual power consumption data. In step S306, the same processing as described above can be performed on the equipment necessary to estimate the effect of the measure, and the power consumption before the introduction of the measure can be calculated at the target store by adding up the results.

[0064] Next, the power consumption for a period during which the effects of the measure will be estimated, for example, one year, is estimated (S307). In this power consumption estimation, the model and analytical parameters 405 necessary for the measure, such as thresholds, saved in step S303, are used. For the annual power data created in step S304, the operating state is estimated using, for example, an operating state estimation model managed on a monthly basis. Since the processing in step S307 is the same as the operating state estimation in step S306, the data estimated in step S306 is used. Next, using an optimization model, for example, a cooking equipment optimization model and the necessary business data, the on / off timing to optimize the operating state of the equipment is output, and power consumption data is created when the operation is performed based on the output result. For example, the cooking equipment optimization model estimates that if the warming state continues above a threshold, the power should be turned off, and the power consumption for that time is set to 0. Power consumption data according to the optimized operating state for each time is output as a result. Next, similar to the processing in S306, the duration of the operating state is calculated from the operating state assigned to each time. The calculated duration is multiplied by the power consumption for each operating state based on the rated power corresponding to the equipment type exemplified in Figure 7(c) for the equipment attributes exemplified in the store attribute management table 700, either by a predefined value or by the power consumption for each operating state based on the equipment type exemplified in Figure 7(c), and the results are added together to calculate the power consumption. If the calculation is performed on a monthly basis, the power consumption is calculated similarly for all months. Alternatively, the power consumption when a measure is introduced may be calculated by summing the power consumption each month based on the power consumption data output as an estimation result by applying an optimization model. In step S307, the same processing as described above may be performed on the equipment necessary to estimate the effect of the measure, and the power consumption after the introduction of the measure in the target store may be calculated by adding up the results.

[0065] Next, the effect of the measure is estimated from the amount of electricity consumed before the measure was implemented, estimated in S306, and the amount of electricity consumed after the measure was implemented, estimated in S307 (S308). In step S308, for example, the difference between the amount of electricity consumed before and after the measure was implemented each month can be calculated. By multiplying this difference by a predefined unit price of electricity, the economic effect of the measure each month is calculated, and the process is terminated by summing up the values ​​for the period of effect estimation.

[0066] Figure 9 shows an example of the input / output screen of the policy effect estimation system 101. The upper part of the screen 901 is the request setting screen, where the attribute information of the target store, measured power data and business data at the target store, and their types are entered. In this example, the location, store area, etc. can be entered in the "Target Store Attributes" field. In the "Data" field, the access location where the power data and business data are located is entered. For store attribute information, the information may also be created in file format and that file may be entered. In the "Policy" field, the policy to estimate the effect of can be selected. All policies may be selected.

[0067] The bottom section 902 of the screen outputs the estimated effects for each selected measure. The "Effect" column on screen 903 outputs the estimated effect results in monetary terms. Since the economic effect is calculated by estimating the amount of electricity consumed each month, the amount can differ from month to month. To make the monthly fluctuations easy to see at a glance, it is recommended to output the fluctuation ranges of the amount for the month with the maximum effect and the month with the minimum effect. In this example on screen 903, the bar shape shown in the "Range" column consists of a left side with light hatching and a right side with dark hatching. The part where the left and right shapes meet corresponds to the average, with the left side representing the minimum fluctuation range of "-300,000 yen" and the right side representing the maximum fluctuation range of "+1,000,000 yen".

[0068] Furthermore, since data is supplemented with actual data from existing stores for months other than those with actual measurements, the range of variation may differ depending on which existing store's data is used. Also, changing the settings of the analysis parameters and effect estimation parameters used to calculate power consumption when introducing the measures may affect the effect. The flowchart shown in Figure 3 explains the case where the process of estimating the effect of the measures is performed once. However, in the process of creating power data and business data for a period in which the effect estimation is performed, for example, one year, the impact of data and parameters can be visualized by using actual data from a different candidate existing store than the first estimation process, or by performing multiple estimation processes with changed parameters. The economic effect calculated as a result of performing multiple estimation processes can be summarized and output as the range of variation between the maximum value and the initial value. Screen 904 shows the results when multiple effect estimations are performed by changing the actual data and parameters used in the estimation process. For example, the range of variation for power data may show the results of the economic effect by store using the actual data used when creating the annual power data, and the details may be shown on a separate screen 905. On separate screen 905, clicking the rectangular icon with light and dark hatching at the intersection of the "Variability" column and the "Power Data Variability" column in the "Item" column on screen 904 will cause separate screen 905 to pop up. If data from multiple stores is used, it is good to show the combination of stores on the horizontal axis. In the case of parameters, the horizontal axis should show the value of the set parameter. Furthermore, only if the factor is usage data, it is good to show a graph of the store data using actual data on screen 906. On screen 906, clicking the lightly hatched square icon at the intersection of the "Details" column and the "Power Data Variability" column will cause it to pop up. It is good to output power data for the same month as the target store's power data, and time-series graphs of the relevant business data for each type of business data.

[0069] As described above, according to the present invention, even if only short-term measurement data is available for the target store where the effectiveness of the measures is to be estimated, it becomes possible to estimate the long-term effectiveness of the measures by using data collected and accumulated at other stores.

[0070] [Example 2] Embodiment 2 of the present invention will now be described. Embodiment 2 differs from Embodiment 1 in that it uses a different management method, specifically the policy management table 400 and the performance data management table 600, and uses a graph database for the management database 107.

[0071] Figure 10 shows an example of management information using a graph database. The nodes may consist of Policy 1001, Store 1002, Company 1003, Analysis Logic 1004, Equipment 1005, Operations 1006, Power Data 1007, and Operations Data 1008. Properties 1009 are set for each node. For example, a Policy node may have properties such as Policy name and target, and may define "Policy Name: Review of Cooking Equipment On / Off Operation, Target: Cooking Equipment" as a single node. Edges 1010 may also be set to represent the relationships between each node. In a Policy node, in order to manage the relationship between the Analysis Logic node and the Equipment node, the node "Policy Name: Review of Cooking Equipment On / Off Operation" may have the edge "USE_LOGICS" set for the Analysis Logic nodes "Type: Estimation of Cooking Equipment Operation State, Target: Cooking Equipment" and "Type: Optimization of Cooking Equipment Operation, Target: Cooking Equipment", and "TARGET" set for the Equipment node "Type: Cooking Equipment, Model: C1". The analysis logic node 1004 should manage the node that defines the power data for each device used in the analysis and the node that defines the business data by associating them with "USE_POWER_DATA" and "USE_BUSINESS_DATA," respectively.

[0072] The flowchart in Example 2 may be the same as the flowchart examples shown in Figures 3, 5, and 8 in Example 1. It is recommended to use the management information in the graph database to identify and load the power data, business data, etc., necessary for analyzing the measures. In step S801 of the flowchart example shown in Figure 8, it is recommended to identify the type of business data necessary for the measures for which the effect estimation will be performed. By specifying the measures from the database, the analysis logic nodes "Type: Estimation of Cooking Equipment Operation Status" and "Type: Optimization of Cooking Equipment Operation", which have the edge "USE_LOGICS", can be identified from the node "Measure Name: Review of Cooking Equipment On / Off Operation". From each analysis logic node, it is recommended to identify the power data node "Target Equipment: Cooking Equipment, ..." and the business data node "Data Name: Number of Products A Cooked, ...", which have the edges "USE_POWE_DATA" and "USE_BUSINESS_DATA" set. It is recommended to obtain the business data using the attribute "Data Link Destination" of the identified nodes. In Example 1, when data is managed using tables, table joins are required during retrieval, whereas the configuration in Example 2 may allow for faster data retrieval.

[0073] Furthermore, in step S804 of the flowchart in Figure 8, when the correlation coefficient is calculated, a bidirectional edge "HAS_CORRELATION" can be set between the power data node and the business data node. The attribute of the edge is set to the correlation coefficient as the key and the calculated correlation coefficient as the value. The relationship between power data and business data may be managed in a database. By doing so, the calculation process in S804 can be reduced for data for which the correlation coefficient has already been calculated, and the processing based on the correlation coefficient of existing stores in S806 can be sped up.

[0074] [Example 3] Embodiment 3 of the present invention will now be described. Embodiment 3 differs from Embodiment 1 in that it uses a different method for creating electricity consumption data during the period for estimating the effectiveness of the measures. The difference between Embodiment 3 and Embodiment 1 is that in the process of S304 in the flowchart of Figure 3, the data obtained from the actual data management is used as the main power data in the electricity consumption processing. In addition, the main power data may be input to the measure effectiveness estimation system 101 as the actual measurement data for the target store. In the actual data management, the main power data may be managed for each store.

[0075] In the policy effect estimation system 101 of this embodiment, the initial input main power data is used to create a separation model for the target store using a logic that separates the power waveform for each piece of equipment, and the power consumption for each piece of equipment is estimated. The created separation model can be used for any policy and can be managed in model management. In the flowchart shown in Figure 3, a model for the policy is created (S302) and saved (S303) using the power data separated for each piece of equipment.

[0076] In creating power data for the effect estimation period in step S304 of Example 3, existing stores with similar attributes are extracted in the flowchart of Figure 5 (S501), and the actual measured data of the target store and the main power data of existing stores for the same period are extracted from the performance data management. The degree of similarity between the actual measured main power data and the main power data of existing stores, for example, the similarity score, may be calculated to identify existing stores with data of a high degree of similarity, or the main power data for the missing period may be obtained from the identified stores to supplement the data.

[0077] In the process of estimating the amount of power consumed before the implementation of the measures in Example 3 (S306), since the data used is the annual main power data, it is possible to first use a management separation model to separate the main power data into power data for each device and create power data for each device that is the target of the measures. Next, the subsequent processing to estimate the operating state using an operating state estimation model is carried out as in Example 1. The subsequent processing to estimate the effect of the measures may be the same as in Example 1.

[0078] As described above, in Example 3, even if the target store does not create individual power data for each piece of equipment, it is possible to supplement the data using the main power data and estimate the effectiveness of the measures.

[0079] In the above embodiments, the facility described was a store, but the present invention can be implemented in any facility to which an effect indicator such as electricity consumption is applicable, such as a factory or a residence, not limited to a store. Furthermore, the evaluation indicator is not limited to electricity consumption, but can be various other things such as water volume, hot water volume, or oil volume. In these cases as well, the same effects as those described in the embodiments will be achieved.

[0080] Although the present invention has been described in detail with reference to the accompanying drawings, the present invention is not limited to such specific configurations and includes various modifications and equivalent configurations within the spirit of the attached claims.

[0081] It should be noted that the present invention is not limited to the embodiments described above, and the components can be modified and implemented in practice without departing from the spirit of the invention.

[0082] It should be noted that the present invention is not limited to the embodiments described above, but includes various modifications and equivalent configurations within the spirit of the attached claims. For example, the embodiments described above are described in detail for the purpose of clearly illustrating the present invention, and the present invention is not necessarily limited to having all the described configurations. Furthermore, some of the configurations of one embodiment may be replaced with those of another embodiment. Furthermore, configurations of other embodiments may be added to the configuration of one embodiment. Furthermore, some of the configurations of each embodiment may be added, deleted, or replaced with those of other embodiments.

[0083] Furthermore, each of the aforementioned configurations, functions, processing units, and processing means may be implemented in hardware, for example, by designing them as integrated circuits, or they may be implemented in software by having a processor interpret and execute programs that realize each function.

[0084] Information such as programs, tables, and files that implement each function can be stored in memory, hard disks, SSDs (Solid State Drives), or other storage media such as IC cards, SD cards, and DVDs.

[0085] Furthermore, the control lines and information lines shown are those deemed necessary for explanation purposes and do not necessarily represent all control lines and information lines required for implementation. In reality, it can be assumed that almost all components are interconnected. [Explanation of Symbols]

[0086] 101...Policy Effectiveness Estimation System 102...Client 103... Request Reception Department 104...Model Learning Department 105...Model Management Department 106...Performance Data Management Department 107...Management Database 108...Data generation department for policy analysis 109...Policy Effectiveness Prediction Department 110...Economic Impact Calculation Department 112...Target Store Actual Measurement Data Management Department

Claims

1. A policy effectiveness estimation system for estimating the effectiveness of policies related to effectiveness indicators in a facility, The computer is comprised of an arithmetic unit that performs calculations and implements the following functional units, and a storage unit that the arithmetic unit can access. A request receiving unit that receives requests for estimation of the effects of the measures at the target facilities where the measures are implemented, A model creation unit creates a model for performing the analysis necessary for the measures at the target facility based on information regarding the effectiveness indicators of the target facility during the second period included in the first period for estimating the effects of the measures, operational information of the target facility, and policy management information related to the measures. A data generation unit for policy analysis generates information on the effectiveness indicators and business information for the target facility during the first period, based on information on the effectiveness indicators and business information for the existing facility, information on the effectiveness indicators and business information for the target facility during the second period, and information on the policy; A policy effectiveness prediction unit calculates information on the effectiveness indicators and operational information of the target facility during the first period, and the model, for the case in which the policy is implemented and the case in which it is not implemented during the first period, and estimates the effect of the policy by supplementing the information on the effectiveness indicators and operational information of the target facility during the second period with information on the effectiveness indicators and operational information of existing facilities based on the calculated information on the effectiveness indicators of the target facility during the first period, and the model, A policy effectiveness estimation system characterized by having the following features.

2. A policy effect estimation system according to claim 1, A policy effectiveness estimation system characterized in that the aforementioned effect indicator is the amount of electricity.

3. A policy effect estimation system according to claim 1, The aforementioned business information is characterized by including time-series data, and is used to estimate the effectiveness of a policy.

4. A policy effect estimation system according to claim 2, The aforementioned business information includes time-series data, The aforementioned data generation unit for policy analysis is: The correlation between the information on the effectiveness indicators of the target facility and the business information of the target facility is calculated to show the correlation between the information on the effectiveness indicators of the target facility and the business information of the target facility. The correlation between the information on the effectiveness indicators of the existing facility and the business information of the existing facility is calculated to show the correlation between the two. A policy effectiveness estimation system characterized by generating information on the effectiveness indicators and business information for the first period of the target facility using the correlation ratio of the effectiveness indicators of the target facility and the correlation ratio of the effectiveness indicators of existing facilities.

5. A policy effect estimation system according to claim 1, The aforementioned data generation unit for policy analysis is: The degree of similarity between the information regarding the evaluation indicators of the existing facility and the information regarding the evaluation indicators of the target facility is calculated. The calculated degree of similarity is compared with a predetermined threshold, As a result of the above comparison, for the existing facilities that were determined to have a high degree of similarity, the correlation between information on evaluation indicators and operational information was calculated. A policy effectiveness estimation system characterized by generating information regarding the effectiveness indicators and operational information for the target facility during the first period using the correlation coefficient.

6. A policy effect estimation system according to claim 5, The aforementioned data generation unit for policy analysis is: The absolute difference between the correlation between the information on the evaluation indicators for the target facility and the operational information, and the correlation between the information on the evaluation indicators for the existing facility and the operational information, is compared with a predetermined threshold. A policy effectiveness estimation system characterized by supplementing the business information of the target facility during the first period other than the second period with the business information of the existing facility determined to have a small absolute value difference in the correlation as a result of the comparison.

7. A method for estimating the effectiveness of a policy, which is performed by a policy effectiveness estimation system that estimates the effectiveness of a policy regarding effectiveness indicators in a facility, The policy effect estimation system is comprised of a computer having a calculation unit that performs calculation processing and a storage unit accessible by the calculation unit. The aforementioned method for estimating the effectiveness of the measures is: The calculation unit receives a request in a request receiving step for estimating the effect of the measures at the target facility where the measures are implemented, The calculation unit creates a model for performing the analysis necessary for the measures at the target facility based on information regarding the effectiveness indicators of the target facility during a second period included in the first period for estimating the effectiveness of the measures, operational information regarding the target facility, and policy management information regarding the measures. The calculation unit generates data for policy analysis, which generates information on the effectiveness indicators and business information for the target facility during the first period by supplementing the information on the effectiveness indicators and business information for the target facility during the second period with the information on the effectiveness indicators and business information for the target facility during the second period, based on the information on the effectiveness indicators and business information for the existing facility, the information on the effectiveness indicators and business information for the target facility during the second period, and the information on the policy, A method for estimating the effectiveness of a policy, characterized in that the calculation unit calculates information regarding the effectiveness indicators for the target facility when the policy is implemented and when it is not implemented during the first period, using information regarding the effectiveness indicators and business information corresponding to the first period for the target facility, and the model, and estimates the effectiveness of the policy based on the calculated information regarding the effectiveness indicators for the target facility when the policy is implemented and when it is not implemented during the first period.

Images