Maintenance cost calculation support system and maintenance cost calculation support program

Abstract

The objective of the present invention is to solve the problems of the prior art, namely, to provide a maintenance cost calculation support system and maintenance cost calculation support program that can compare the "items" in the construction breakdown statement with the "item names" in the BELCA record without relying on the judgment of the worker or solely on the "replacement dictionary table". [Solution] The maintenance cost calculation support system is a system that supports the calculation of maintenance costs for buildings, and comprises a BELCA storage means, a construction breakdown file input means, a BELCA record selection means, and a maintenance cost calculation means. Of these, the BELCA record selection means selects a BELCA identifier based on the similarity between construction cost text data and actual text data.

Description

Technical Field

[0002] , ,

[0004] , , , , ,

[0003]

[0001] The invention of the present application relates to a technology for calculating maintenance costs for buildings. More specifically, it relates to a maintenance cost calculation support system and a maintenance cost calculation support program that can calculate the costs related to maintenance for each item using, as input information, an "itemized construction cost statement" that shows the details of construction costs.

Background Art

[0002] When constructing a new building, the construction cost is determined between the client and the contractor, and usually an itemized construction cost statement showing the details of the construction cost is created. This itemized construction cost statement shows, for each "item" such as construction materials, equipment, installation costs, labor costs, consumables and sundries, expenses, etc., the quantity and unit, the amount per unit quantity (so-called unit price), and the amount for each item obtained by multiplying the quantity by the unit price. The sum of the amounts for all items is the total construction cost. Conventionally, the itemized construction cost statement has mainly been used at the start of construction, at the time of a design change contract, and at the time of final settlement after delivery.

[0003] In 2001, the Basic Law for the Promotion of a Recycling-based Society was enacted, and manufacturers and others are to bear certain responsibilities (extended producer responsibility) not only for the environmental impact during the manufacturing stage (process) but also for reducing the environmental impact after use. As a result, the life cycle of all kinds of things such as various structures and products has come to be emphasized. Buildings are no exception, and efforts are being made to consider reducing the environmental impact throughout the entire life cycle from planning, design, construction to maintenance, operation, and demolition, and to calculate the costs (life cycle cost: LCC) throughout the entire life cycle by including the budgets for each stage.

[0004] Against this backdrop, "long-term repair plans" for buildings have recently come into common practice. These long-term repair plans include both renewal plans and repair plans for buildings. The renewal plan outlines the renewal cycle and renewal costs, while the repair plan outlines the repair cycle and repair costs. Generally, renewal and repair plans are planned item by item in the construction breakdown, and renewal and repair costs are generally planned based on the amount (quantity x unit price) related to each item. In other words, it is believed that long-term repair plans can be efficiently created by essentially reusing the construction breakdown.

[0005] For example, if the construction breakdown is created using spreadsheet software such as Microsoft Excel or Google Sheets, it is conceivable that the appropriate renewal cycle and cost coefficient (the multiplier to the original cost required for renewal) for each item could be automatically selected from a designated database. In other words, a computer would be used to compare the "item name" representing the item in the construction breakdown with the "item name" entered in the database, and then retrieve the renewal cycle and renewal cost rate related to that item.

[0006] The Building and Equipment Long-life Cycle Association (BELCA®), a public interest incorporated association, provides a database and codes for "Data Collection for Building Lifecycle Management" (hereinafter referred to as the "BELCA Database" for convenience), which provides information necessary for maintaining buildings, namely information on "renewal cycles" and "renewal costs," and information on "repair cycles" and "repair costs." Here, information on renewal costs is information necessary to calculate renewal costs, and is directly a coefficient multiplied by the amount related to the item in the construction breakdown (hereinafter simply referred to as the "renewal coefficient"). Similarly, information on repair costs is a coefficient related to the unit price (hereinafter simply referred to as the "repair unit price coefficient") and a coefficient related to the quantity (hereinafter simply referred to as the "repair quantity coefficient") multiplied by the amount related to the item in the construction breakdown. In other words, renewal costs are obtained by multiplying the amount in the construction breakdown by the renewal coefficient, and repair costs are obtained by multiplying the amount in the construction breakdown by the repair unit price coefficient and the repair quantity coefficient.

[0007] By applying the renewal cycle and renewal coefficient, repair cycle, repair unit price coefficient and repair quantity coefficient from the BELCA database to the construction breakdown sheet created using spreadsheet software, the renewal costs required for each renewal cycle and the renewal costs required for each repair cycle can be automatically obtained, making it possible to create a long-term repair plan extremely efficiently. At this time, it is necessary to select the BELCA record that is most suitable for the items in the construction breakdown sheet from among the records that make up the BELCA database (hereinafter simply referred to as "BELCA record"), and in order to select it, there is no choice but to rely on the "item name". This item name is text information that represents the content of the item in the construction breakdown sheet and is text information that represents the content of the BELCA record.

[0008] Incidentally, it is known that the item names entered in the construction breakdown sheet reflect the individuality of the person who prepared it, and it is not uncommon for the same item to have different names depending on the person who prepared it. For example, when "concrete" is the item, the item name may be written as "ready-mix concrete," or as "fresh concrete" or "ready-mix con." Similarly, when "formwork" is the item, the item name may be written as "concrete panel," or as "concrete panel" or "plywood for concrete formwork."

[0009] On the other hand, BELCA records contain only one type of item name. This means that although the "item names" in the construction breakdown and the "item names" in the BELCA database may appear different on the surface, they often refer to the same item. Anyone involved in construction can understand that "concrete" and "ready-mix concrete" refer to the same item, and that "concrete formwork panels" and "plywood for concrete formwork" refer to the same item, but it is not easy for someone not involved to determine that these are the same thing. Furthermore, since construction work generally consists of many types of work and uses a large number of items, even someone involved may not be able to determine whether different item names refer to the same item.

[0010] Thus, the task of matching the "item names" in the construction breakdown statement with the "item names" in the BELCA database is extremely difficult and prone to errors. Moreover, the number of items that make up the construction breakdown statement for building work is enormous, and the BELCA database consists of a vast number of records. Therefore, this task is time-consuming and there is even a risk of frequent errors.

[0011] Therefore, it is conceivable to convert (translate, so to speak) the item names, which are written with various names, into a predetermined item name. In the example above, the item names "ready-mix concrete," "fresh concrete," and "ready-mix con" are all converted to "concrete," and the item names "concrete panel," "concrete panel," and "plywood for concrete formwork" are all converted to "formwork." Techniques for unifying the item names in construction breakdowns to a predetermined term have been proposed before. For example, Patent Document 1 proposes a technique for comparing information in order documents with the cost estimation results from a cost estimation system by replacing terms with the same meaning used in the documents with a single term for comparison. [Prior art documents] [Patent Documents]

[0012] [Patent Document 1] Japanese Patent Publication No. 2017-10501 [Overview of the project] [Problems that the invention aims to solve]

[0013] The invention disclosed in Patent Document 1 is a technique for replacing terms by creating a so-called "replacement dictionary table" in which "terms to be replaced (multiple candidates)" and "terms to be replaced (one term)" correspond. However, as mentioned above, the number of items related to construction work is enormous, and the BELCA database contains a huge number of records, so it is extremely difficult and not at all practical to create a "replacement dictionary table" that can convert every item without omission.

[0014] The objective of the present invention is to solve the problems of the prior art, namely, to provide a maintenance cost calculation support system and maintenance cost calculation support program that can compare the "items" in the construction breakdown statement with the "item names" in the BELCA record without relying on the judgment of the worker or solely on the "replacement dictionary table". [Means for solving the problem]

[0015] The present invention is based on a novel idea: creating "actual data" by compiling examples of how item names in construction breakdown statements correspond to item names in BELCA records, and then extracting information necessary for calculating building maintenance costs from the BELCA database by comparing the items in construction breakdown statements with this actual data.

[0016] The maintenance cost calculation support system of the present invention is a system that supports the calculation of maintenance costs for buildings, and comprises a BELCA storage means, a construction breakdown file input means, a BELCA record selection means, and a maintenance cost calculation means. Of these, the BELCA storage means is a means for storing multiple BELCA records, and the construction breakdown file input means is a means for inputting a "construction breakdown file" consisting of multiple construction cost records. The BELCA record selection means is a means for selecting BELCA records using the construction breakdown file and actual data (data consisting of multiple actual records), and the maintenance cost calculation means is a means for calculating "renewal costs" and "repair costs" related to the maintenance of buildings. A BELCA record is composed of a "BELCA identifier (identifier of the BELCA record)", "renewal cycle", "renewal coefficient", "repair cycle", "repair unit price coefficient", and "repair quantity coefficient". A construction cost record is composed of "construction cost text data" and "amount by item", and an actual record is composed of "actual text data" and a "BELCA identifier". The BELCA record selection means selects actual records for each construction cost record included in the construction breakdown file based on the similarity between the construction cost text data and the actual text data, and also selects BELCA records from the BELCA storage means that correspond to the same BELCA identifier as the BELCA identifier included in the selected actual record. The maintenance cost calculation means reads the renewal cycle and renewal coefficient included in the BELCA record selected by the BELCA record selection means for each construction cost record included in the construction breakdown file, and calculates the renewal cost as the product of the item-specific amount related to the construction cost record and the renewal coefficient. Similarly, the maintenance cost calculation means reads the repair cycle, repair unit price coefficient and repair quantity coefficient included in the BELCA record selected by the BELCA record selection means for each construction cost record included in the construction breakdown file, and calculates the repair cost as the product of the item-specific amount related to the construction cost record, the repair unit price coefficient and the repair quantity coefficient. This makes it possible to grasp the renewal costs required for each renewal cycle and the repair costs required for each repair cycle for each construction cost record.

[0017] The maintenance cost calculation support system of the present invention can also calculate "similarity" based on construction cost text data and actual text data, and select actual records based on this similarity.

[0018] The maintenance cost calculation support system of the present invention can also be configured to select a BELCA record from among multiple candidates. In this case, the BELCA record selection means first calculates the similarity based on the construction cost text data and the actual text data, and selects k (k is a natural number) actual records based on the similarity. Next, the BELCA record selection means classifies the k actual records into groups according to the BELCA identifier, and then calculates an "individual index value" by summing the similarities of the actual records belonging to the group. Finally, the BELCA record selection means selects the BELCA record corresponding to the BELCA identifier of the group with the highest individual index value.

[0019] The maintenance cost calculation support system of the present invention can also be configured to select BELCA records based on "maximum similarity." In this case, the BELCA record selection means extracts the maximum similarity that represents the highest value among the similarities related to the group, and calculates the product of the number of actual records belonging to each group and the maximum similarity as an individual index value. The BELCA record selection means then selects the BELCA record related to the BELCA identifier of the group with the highest individual index value.

[0020] The maintenance cost calculation support system of the present invention can also be configured to select actual text data whose similarity exceeds a predetermined threshold, and to select up to k actual records that include such actual text data.

[0021] The maintenance cost calculation support system of the present invention can also calculate cosine similarity based on construction cost text data and actual text data, and select BELCA records based on this cosine similarity.

[0022] The maintenance cost calculation support system of the present invention can also be configured to select BELCA records using machine learning. In this case, the BELCA record selection means outputs a BELCA identifier from the construction cost text data by applying machine learning to the training data, and selects the BELCA record corresponding to that BELCA identifier. This training data is generated based on the "actual text data" and "BELCA identifier" included in the actual records.

[0023] The maintenance cost calculation support system of the present invention may also be configured such that the construction cost records include construction cost units and the actual records include actual units. In this case, the BELCA record selection means selects from the actual data storage means an actual record that includes an actual unit which is the same unit as the construction cost units included in the construction cost records.

[0024] The maintenance cost calculation support program of the present invention is a program that causes a computer to execute a function of supporting the calculation of the maintenance cost of a building, and has a function of causing the computer to execute a work breakdown sheet file import process, a BELCA record selection process, and a maintenance cost calculation process. Among these, in the work breakdown sheet file import process, the work breakdown sheet file input by the operator is imported. In the BELCA record selection process, for each construction cost record included in the work breakdown sheet file, an actual performance record is selected based on the similarity between the construction cost text data and the actual performance text data, and a BELCA record related to the same BELCA identifier as the BELCA identifier included in the actual performance record is selected from the BELCA storage means. In the maintenance cost calculation process, for each construction cost record included in the work breakdown sheet file, the update cycle and update coefficient included in the BELCA record selected in the BELCA record selection process are read, and the product of the item-by-item amount related to the construction cost record and the update coefficient is calculated as the update cost. Similarly, in the maintenance cost calculation process, for each construction cost record included in the work breakdown sheet file, the repair cycle, repair unit price coefficient, and repair quantity coefficient included in the BELCA record selected in the BELCA record selection process are read, and the product of the item-by-item amount related to the construction cost record, the repair unit price coefficient, and the repair quantity coefficient is calculated as the repair cost. Thereby, the update cost required for each update cycle and the repair cost required for each repair cycle can be grasped for each construction cost record.

Effect of the Invention

[0025] The maintenance cost calculation support system and the maintenance cost calculation support program have the following effects. (1) Conventionally, since the "item" of the work breakdown sheet was compared with the "item name" of the BELCA record by the judgment of the operator, it took a significant amount of time to calculate the maintenance cost related to the building. Specifically, it was a task that required at least 10 man-hours of skilled workers for calculating the maintenance cost of one property. On the other hand, according to the present invention, anyone can operate it without securing skilled workers, and moreover, the result can be obtained in about several minutes for calculating the maintenance cost of one property. (2) Also, by comparing the "items" in the construction breakdown document with the "items" in the BELCA record without relying on the judgment of the operator, human error can be suppressed, that is, the maintenance cost related to the building can be calculated with high accuracy. (3) Since the maintenance cost related to the building can be calculated quickly and with high accuracy, an efficient and high-level long-term repair plan can be formulated.

Brief Description of Drawings

[0026] [Figure 1] Block diagram showing the main components of the maintenance cost calculation support system of the present invention. [Figure 2] Model diagram schematically showing a construction breakdown document file composed of a plurality of construction cost records. [Figure 3] Model diagram schematically showing a BELCA database composed of a plurality of BELCA records. [Figure 4] Model diagram schematically showing performance data composed of a plurality of performance records. [Figure 5] Block diagram showing the main components of the BELCA record selection means using the weighted k-nearest neighbor method. [Figure 6] Step diagram explaining the weighted k-nearest neighbor method. [Figure 7] Flow diagram showing an example of the main processing flow until the maintenance cost calculation support system of the present invention calculates the total construction cost. [Figure 8] Flow diagram showing an example of the processing flow that the maintenance cost calculation support program of the present invention causes a computer to execute.

Modes for Carrying Out the Invention

[0027] An example of an embodiment of the maintenance cost calculation support system and the maintenance cost calculation support program of the present invention will be described based on the drawings.

[0028] 1. Maintenance Cost Calculation Support System First, the maintenance cost calculation support system of the present invention will be explained in detail with reference to the diagram. The maintenance cost calculation support program of the present invention is the program used when executing the maintenance cost calculation support system of the present invention. Therefore, the maintenance cost calculation support system of the present invention will be explained first, followed by the maintenance cost calculation support program of the present invention.

[0029] Figure 1 is a block diagram showing the main components of the maintenance cost calculation support system 100 of the present invention. As shown in this figure, the maintenance cost calculation support system 100 of the present invention is composed of a construction breakdown file input means 110, a BELCA record selection means 120, a maintenance cost calculation means 130, and a BELCA storage means 140, and may also be composed of actual data storage means 150, etc.

[0030] Of the main components constituting the maintenance cost calculation support system 100, the construction breakdown file input means 110, the BELCA record selection means 120, and the maintenance cost calculation means 130 can be manufactured as dedicated components, or a general-purpose computer device can be used. In other words, by having the computer device perform calculation processing using the maintenance cost calculation support program of the present invention, which will be described later, each means performs processing specific to that means. This computer device is equipped with a processor such as a CPU, memory such as ROM or RAM, input means such as a mouse or keyboard, and a display, and performs calculation processing according to a predetermined program, and can be composed of a personal computer (PC), a server, a tablet PC such as an iPad (registered trademark), a mobile terminal including a smartphone, etc.

[0031] The BELCA storage means 140 and the performance data storage means 150 can utilize the storage devices of a general-purpose computer device (e.g., a personal computer), or they can be built on a database server. When built on a database server, it can be located on a local network (LAN: Local Area Network), or it can be a cloud server that stores data via the internet.

[0032] The following will provide a detailed explanation of each of the main elements that constitute the maintenance cost calculation support system 100 of the present invention.

[0033] (Method for inputting construction breakdown file) The construction breakdown file input means 110 is a means for inputting a construction breakdown file, and can be configured to allow an operator to import the construction breakdown file by operating a pointing device (such as a mouse, touch panel, pen tablet, touchpad, trackpad, or trackball) or keyboard. Here, the construction breakdown file is a "construction breakdown" showing the construction costs related to a building, created using so-called spreadsheet software such as "Microsoft Excel" or "Google Sheets". However, this construction breakdown is a so-called "construction breakdown with amounts" in which the amounts for each item (hereinafter simply referred to as "item-specific amounts") are entered.

[0034] Since the construction cost breakdown file corresponds to the construction cost breakdown, it is composed of multiple items related to construction work (hereinafter referred to as "construction cost records"). Figure 2 is a schematic model diagram showing a construction cost breakdown file composed of multiple construction cost records. As shown in this figure, each construction cost record is composed of text data that represents the content of the construction cost record (in this figure, name, quality, shape and dimensions) and item-specific amounts. Furthermore, it can also be composed of information such as the ID (IDentification) of the construction cost record, unit, quantity, and unit price, or it can be composed of classifications such as major and minor categories. For convenience, here we will refer to the ID of the construction cost record as the "construction cost identifier," the text data that represents the content of the construction cost record as the "construction cost text data," the unit of the construction cost record as the "construction cost unit," the quantity of the construction cost record as the "construction cost quantity," and the unit price related to the construction cost record as the "item-specific unit price." In other words, the item-specific amount is the value obtained by multiplying the construction cost quantity by the item-specific unit price. The itemized amounts include not only the purchase costs of equipment, but also labor costs such as transportation and installation. As shown in this diagram, when a construction cost record includes multiple pieces of information to represent its contents, such as "name" and "quality / shape / dimensions," these multiple pieces of information can be integrated into the construction cost text data, or either one of them can be used as the construction cost text data.

[0035] (BELCA memory device) The BELCA storage means 140 is a means for storing a "BELCA database" consisting of multiple BELCA records. Figure 3 is a schematic model diagram showing a BELCA database composed of multiple BELCA records. As shown in this figure, each BELCA record is composed of the BELCA record ID, update cycle, update coefficient, repair cycle, repair unit price coefficient, and repair quantity coefficient, and can also be composed of information such as text data and units that represent the contents of the BELCA record. For convenience, the BELCA record ID will be referred to as the "BELCA identifier," the text data that represents the contents of the BELCA record as the "BELCA text data," and the units related to the BELCA record as the "BELCA unit."

[0036] Here, the renewal cycle is the period (time) during which the equipment related to that item should be renewed, and the renewal coefficient is a coefficient multiplied by the cost of newly installing the equipment related to that item (i.e., the item-specific amount). Similarly, the repair cycle is the period (time) during which the equipment related to that item should be repaired, the repair unit cost coefficient is a coefficient multiplied by the item-specific amount of the equipment related to that item, and the repair quantity coefficient is the quantity of equipment related to that item that needs to be repaired. By multiplying the item-specific amount by the renewal coefficient, the renewal cost required for each renewal cycle is obtained, and by multiplying the item-specific amount by the repair unit cost coefficient and the repair quantity coefficient, the repair cost required for each repair cycle is obtained. Note that the BELCA database can also be specified so as not to include the repair quantity coefficient in the BELCA record, in which case the repair quantity coefficient is set to a predetermined default value (for example, 1.0) and the repair cost is calculated.

[0037] (Means for storing performance data) The performance data storage means 150 is a means for storing "performance data" consisting of multiple performance records. Figure 4 is a schematic model diagram showing performance data composed of multiple performance records. Performance data does not necessarily have to be in a database structure (although it may be), and can also be composed as a simple collection of performance records. As shown in this figure, each performance record includes at least text data representing the content of the performance record and a BELCA identifier, and can also include information such as the performance record's ID, unit, and BELCA text data. For convenience, here we will refer to the performance record's ID as the "performance identifier," the text data representing the content of the performance record as the "performance text data," and the unit of the performance record as the "performance unit." In addition to performance records created to constitute performance data, performance text data associated with BELCA identifiers etc. in other operations can be used as performance records, or performance text data associated with BELCA identifiers when generating the learning data described later can be used as performance records.

[0038] As described above, the performance data is a compilation of records from when construction cost records and BELCA records were linked in the past, and is literally a database based on "actual results" of linked records. Records can be linked by workers comparing construction cost text data and BELCA text data, extracted from the results of work performed manually for long-term repair plans, or linked using the maintenance cost calculation support system 100 of the present invention. For example, in Figure 4, a construction cost record where the construction cost text data is "ready-mix concrete" and a BELCA record where the BELCA text data is "concrete" are linked, and necessary information (construction cost text data, BELCA identifier, etc.) is taken from the construction cost record and BELCA record respectively, and an actual record with the actual identifier "AC_00002" is generated. Note that the construction cost text data taken from the construction cost record is named "actual text data" when stored as an actual record, and similarly, the construction cost unit taken from the construction cost record is named "actual unit" when stored as an actual record. In other words, the origin of the performance text data and performance units is said to be the construction cost text data and construction cost units.

[0039] As described above, performance data is generated based on actuals associated with construction cost records and BELCA records. Therefore, while the performance identifier in the performance data is unique information that does not overlap with other performance records, the BELCA identifier in the performance data may overlap with other performance records. For example, in the diagram, the performance records with performance identifiers "AC_00006" to "AC_00008" all have the BELCA identifier "ID_100003".

[0040] (BELCA Record Selection Method) The BELCA record selection means 120 is a means of selecting an appropriate BELCA record (hereinafter referred to as the "suitable BELCA record") for the construction cost record from the "BELCA database" stored in the BELCA storage means 140. Of course, a suitable BELCA record is selected for all (or some) of the construction cost records that make up the construction breakdown file. The general procedure by which the BELCA record selection means 120 selects a suitable BELCA record is described below.

[0041] First, the BELCA record selection means 120 queries the actual data storage means 150 for the construction cost text data included in the construction cost record and detects actual text data similar to that construction cost text data. Then, it selects actual records containing the detected actual text data (hereinafter referred to as "similar actual records") from the actual data storage means 150. At this time, it is also possible to select similar actual records on the condition that they contain the same actual unit as the construction cost unit related to the construction cost record. Furthermore, it is possible to select one actual record as a similar actual record, or to select two or more actual records as similar actual records. In addition, the BELCA record selection means 120 queries the BELCA storage means 140 (BELCA database) for the BELCA identifier included in the selected similar actual records and selects BELCA records related to the same BELCA identifier as the similar actual record as "compatible BELCA records". The compatible BELCA records selected by the BELCA record selection means 120 can also be modified as appropriate by operator operation.

[0042] The BELCA record selection means 120 can utilize various conventional techniques, including machine learning, to select suitable BELCA records. For example, the degree of similarity between construction cost text data and actual performance text data can be expressed as "similarity," and suitable BELCA records can be selected according to this similarity. Alternatively, the BELCA record that receives the most votes (in this case, the number of similar actual performance records) using the "k-nearest neighbors" method based on similarity can be selected as the suitable BELCA record. Various conventional methods can be used to measure similarity. For example, there are methods that determine similarity based on a small "distance" value calculated based on construction cost text data and actual performance text data (hereinafter referred to as the "distance-based method"), and methods that determine similarity based on a value that approaches 1 when calculated based on construction cost text data and actual performance text data (hereinafter referred to as the "similarity-based method"). Typical examples of distance-based methods include methods using Euclidean distance, Manhattan distance, and Chebyshev distance, while typical examples of similarity-based methods include methods using cosine similarity, Jacquard coefficients, and Dice coefficients.

[0043] Alternatively, a model can be trained using machine learning (e.g., Deep Learning), and this model can be used to select BELCA records. This training data consists of construction cost text data with BELCA identifiers attached. This data can be generated specifically for machine learning, or it can be generated using multiple actual records that make up the actual data. In this case, training data can be generated by attaching actual text data to construction cost text data. Then, the model generated by machine learning and the "construction cost text data" are used to select a predetermined "actual text data." In other words, "actual text data" with high similarity to the selected "construction cost text data" is selected, and the BELCA record (i.e., the matching BELCA record) of the actual record containing the selected actual text data is determined.

[0044] Furthermore, the BELCA record selection means 120 can also detect actual performance text data similar to construction cost text data using a new method (hereinafter referred to as "weighted k-nearest neighbors") that applies the conventionally used "k-nearest neighbors" method. The procedure for detecting similar actual performance text data using the weighted k-nearest neighbors method will be explained in detail below.

[0045] Figure 5 is a block diagram showing the main components of the BELCA record selection means 120 using the weighted k-nearest neighbor method, and Figure 6 is a step diagram explaining the weighted k-nearest neighbor method. As shown in Figure 5, the BELCA record selection means 120 in this case can be configured to include a vector transformation means 121, a similarity calculation means 122, a candidate record selection means 123, a group classification means 124, an individual index value calculation means 125, a k-nearest neighbor selection means 126, and so on.

[0046] As shown in Figure 5, the vector conversion means 121 first generates a "construction cost text vector" by converting the construction cost text data into a vector, and also generates a "actual text vector" by converting the actual text data into a vector. At this time, actual text vectors are generated for multiple (for example, all) actual text data included in the actual data. Various conventional techniques can be used for this vector conversion. Furthermore, when performing vector conversion, it is possible to refer to a word dictionary table to avoid inappropriate delimiters, perform normalization to unify similar terms, and weight (or remove) delimited terms according to their frequency of appearance.

[0047] Once the construction cost text vector and the actual data text vector are obtained, the similarity calculation means 122 calculates the "cosine similarity" based on these vectors. In this case, multiple actual data sets are considered for one construction cost text vector, meaning that multiple cosine similarities are calculated for each kist vector included in the actual data. Various conventional techniques can also be used in calculating this cosine similarity.

[0048] When multiple cosine similarity scores are calculated for each performance text vector included in the performance data, the candidate record selection means 123 selects k (k is a natural number) performance records in descending order of cosine similarity. It is advisable to predetermine the value of "k". Alternatively, it is possible to select all performance records whose cosine similarity exceeds a predetermined threshold (hereinafter referred to as the "similarity threshold") without limiting the number to be selected (e.g., k). Furthermore, it is possible to select up to k performance records whose cosine similarity exceeds the similarity threshold. Therefore, in this case, k performance records may be selected, or fewer than k performance records may be selected. In this case, it is also possible to set a selection condition that the construction cost unit of the construction cost record related to the construction cost text vector and the performance unit of the performance record are the same. For convenience, the k performance records selected here will be called "provisional similar performance records". Then, the group classification means 124 classifies the k provisional similar performance records into groups (in the figure, Group A, Group B, and Group C) as shown in Figure 6. Specifically, provisional similar performance records containing a common BELCA identifier are grouped together to form a group. For example, in Figure 6, Group A consists of provisional similar performance records containing the BELCA identifier "ID_100001", Group B consists of provisional similar performance records containing the BELCA identifier "ID_100002", and Group C consists of provisional similar performance records containing the BELCA identifier "ID_100003".

[0049] When provisional similar performance records are classified into groups, the individual indicator value calculation means 125 calculates the "individual indicator value". In calculating the individual indicator value, first, the maximum value among the cosine similarities of the provisional similar performance records belonging to each group is set as the "maximum cosine similarity". Then, as shown in Figure 6, the individual indicator value is calculated for each group by multiplying the "maximum cosine similarity" by "the number of provisional similar performance records belonging to that group". Alternatively, the individual indicator value calculation means 125 can be configured to calculate the individual indicator value by summing the individual cosine similarities (cosine similarity for each provisional similar performance record belonging to the group) for each group. In any case, the individual indicator value is obtained for each group, that is, as many as there are groups.

[0050] Once individual index values ​​are obtained for each group, the k-nearest neighbor selection means 126 selects the group with the highest individual index value, outputs a BELCA identifier corresponding to that group, and further selects the BELCA record associated with that BELCA identifier as the suitable BELCA record. Thus, in the conventional "k-nearest neighbor" method, even if there are many actual records within a group but their similarity is low, a BELCA record related to that group will be selected. On the other hand, with the "weighted k-nearest neighbor" method, the BELCA identifier can be determined not by simple majority vote, but by considering the degree of similarity (in this case, cosine similarity), and thus suitable BELCA records can be selected. Note that although this explanation uses cosine similarity, other similarity-based methods or distance-based methods can of course be adopted.

[0051] (Method for calculating maintenance costs) The maintenance cost calculation means 130 is a means for calculating the replacement costs and repair costs among the maintenance costs. The procedure by which the maintenance cost calculation means 130 calculates the replacement costs and repair costs will be described below.

[0052] First, the maintenance cost calculation means 130 reads the item-specific amount for each construction cost record (i.e., item-specific) contained in the construction breakdown file. It also reads the suitable BELCA record selected by the BELCA record selection means 120, and reads the renewal cycle and renewal coefficient, repair cycle and repair unit price coefficient, and repair quantity coefficient contained in that suitable BELCA record. At this time, each construction cost record already has a suitable BELCA record associated with it, meaning that the renewal coefficient, repair unit price coefficient, etc. related to the suitable BELCA record are associated with that construction cost record and can be read. Of course, the item-specific amount, renewal coefficient, repair unit price coefficient, repair quantity coefficient, etc. are read for all construction cost records that make up the construction breakdown file, or for some of the construction cost records. Of course, construction cost records for which no actual records exceeding the similarity threshold are selected are considered to be excluded from repair or renewal, and since they are not assigned a BELCA identifier (i.e., a BELCA record), they are not read here.

[0053] When the maintenance cost calculation means 130 reads out item-specific amounts, renewal coefficients, repair unit cost coefficients, repair quantity coefficients, etc., it calculates the renewal costs required for each renewal cycle by multiplying the item-specific amounts by the renewal coefficients, and calculates the repair costs required for each repair cycle by multiplying the item-specific amounts by the repair unit cost coefficients and repair quantity coefficients. However, in cases where the BELCA record is specified not to include the repair quantity coefficient, the repair cost is calculated with the repair quantity coefficient set to a default value (for example, 1.0). Of course, renewal costs and repair costs are calculated for all construction cost records that make up the construction breakdown file, or for some of the construction cost records, respectively.

[0054] (Process flow) The main processes of the maintenance cost calculation support system 100 of the present invention will be explained in detail below with reference to Figure 7. Figure 7 is a flowchart showing an example of the main processes of the maintenance cost calculation support system 100 of the present invention until the total construction cost is calculated. The center column shows the processes to be executed, the left column shows the information necessary for those processes, and the right column shows the information generated from those processes.

[0055] To calculate the total construction cost, the operator first inputs the construction breakdown file using the construction breakdown file input means 110 (Step 201 in Figure 7). The BELCA record selection means 120 then selects similar performance records (Step 202 in Figure 7) and a suitable BELCA record (Step 203 in Figure 7). The maintenance cost calculation means 130 then calculates the renewal cost for each renewal cycle by multiplying the itemized amount by the renewal coefficient, and also calculates the repair cost for each repair cycle by multiplying the itemized amount by the repair unit price coefficient and the repair quantity coefficient (Step 204 in Figure 7).

[0056] 2. Maintenance Cost Calculation Support Program Next, the maintenance cost calculation support program will be explained in detail with reference to Figure 8. Figure 8 is a flowchart showing an example of the processing flow that the maintenance cost calculation support program instructs the computer to execute. The maintenance cost calculation support program is the program used when executing the maintenance cost calculation support system 100 described so far. Therefore, explanations that overlap with those explained in the maintenance cost calculation support system 100 will be avoided, and the explanation will mainly focus on content specific to the maintenance cost calculation support program. In other words, content not described here is the same as that explained in "1. Maintenance Cost Calculation Support System".

[0057] First, in the construction breakdown file import process, the construction breakdown file entered by the operator using the construction breakdown file input means 110 is imported (Step 301 in Figure 8). This allows similar actual records to be selected for each construction cost record that makes up the construction breakdown file (Step 302 in Figure 8), and a suitable BELCA record to be selected (Step 303 in Figure 8). Next, the renewal cost for each renewal cycle is calculated by multiplying the item-specific amount by the renewal coefficient, and the repair cost for each repair cycle is calculated by multiplying the item-specific amount by the repair unit price coefficient and the repair quantity coefficient (Step 304 in Figure 8). If the renewal costs and repair costs for all (or some) of the construction cost records that make up the construction breakdown file have been calculated (Yes in Step 305 in Figure 8), the process is terminated. Otherwise (No in Step 305 in Figure 8), the calculation of repair costs (Step 304 in Figure 8) is repeatedly performed while changing the construction cost records. [Industrial applicability]

[0058] The maintenance cost calculation support system and maintenance cost calculation support program can be effectively implemented not only for buildings but also for civil engineering structures, electrical equipment, and the like. Considering that the present invention enables the efficient formulation of long-term repair plans, and ultimately leads to the extended lifespan of buildings under appropriate maintenance, the present invention is not only industrially applicable but also has the potential to make a significant social contribution. [Explanation of Symbols]

[0059] 100 Maintenance cost calculation support system for the present invention 110 (Method for inputting construction breakdown file for maintenance cost calculation support system) 120 (Method for selecting BELCA records in the maintenance cost calculation support system) 121 Vector transformation means (of the BELCA record selection means) 122 Similarity calculation means (for BELCA record selection means) 123 Candidate record selection method (for BELCA record selection method) 124 Group classification means (of the BELCA record selection means) 125 (Method for calculating individual indicator values ​​of BELCA record selection means) 126 k-nearest neighbor selection method (for BELCA record selection method) 130 (Maintenance cost calculation means of the maintenance cost calculation support system) 140 (BELCA storage means for the maintenance cost calculation support system) 150 (Means for storing actual data of the maintenance cost calculation support system)

Claims

1. A system that assists in calculating the maintenance costs of buildings, A BELCA storage means for storing multiple BELCA records, A means for inputting a construction breakdown file consisting of multiple construction cost records, A BELCA record selection means that selects the BELCA record using the aforementioned construction breakdown file and performance data consisting of multiple performance records, The building comprises a means for calculating maintenance costs, which calculates renewal costs and repair costs related to the maintenance of the building. The BELCA record is composed of a BELCA identifier, which is the identifier of the BELCA record, an update cycle, an update coefficient, a repair cycle, a repair unit price coefficient, and a repair quantity coefficient. The aforementioned construction cost record consists of construction cost text data and itemized amounts, The aforementioned performance record comprises performance text data and the aforementioned BELCA identifier. The BELCA record selection means selects the actual record for each construction cost record included in the construction breakdown file based on the similarity between the construction cost text data and the actual text data. Furthermore, the BELCA record selection means selects from the BELCA storage means a BELCA record relating to the same BELCA identifier as the BELCA identifier included in the selected performance record. The maintenance cost calculation means reads the renewal cycle and renewal coefficient included in the BELCA record selected by the BELCA record selection means for each construction cost record included in the construction breakdown file, and calculates the product of the itemized amount related to the construction cost record and the renewal coefficient as the renewal cost. Furthermore, the maintenance cost calculation means reads out the repair cycle, the repair unit price coefficient, and the repair quantity coefficient included in the BELCA record selected by the BELCA record selection means for each construction cost record included in the construction breakdown file, and calculates the repair cost as the product of the itemized amount related to the construction cost record, the repair unit price coefficient, and the repair quantity coefficient. For each construction cost record, the renewal costs incurred for each renewal cycle and the repair costs incurred for each repair cycle can be determined. A maintenance cost calculation support system characterized by the following features.

2. The BELCA record selection means calculates similarity based on the construction cost text data and the actual performance text data, and selects the actual performance record based on the similarity. Maintenance cost calculation support system as described in feature 1.

3. The BELCA record selection means selects k (where k is a natural number) of the aforementioned records based on the similarity, Furthermore, the BELCA record selection means classifies the k performance records into groups according to the BELCA identifier, and then calculates individual index values ​​by summing the similarity scores of the performance records belonging to the group. Furthermore, the BELCA record selection means selects the BELCA record relating to the BELCA identifier of the group whose individual index value is the largest. The maintenance cost calculation support system according to feature 2.

4. The BELCA record selection means selects k (where k is a natural number) of the aforementioned records based on the similarity, Furthermore, the BELCA record selection means classifies the k performance records into groups according to the BELCA identifier, extracts the maximum similarity value that represents the highest value among the similarities related to the group, and calculates the product of the number of performance records belonging to the group and the maximum similarity value as an individual index value. Furthermore, the BELCA record selection means selects the BELCA record relating to the BELCA identifier of the group whose individual index value is the largest. The maintenance cost calculation support system according to feature 2.

5. The BELCA record selection means selects up to k (where k is a natural number) of the performance records that include the performance text data whose similarity exceeds a predetermined threshold. A maintenance cost calculation support system according to claim 3 or 4.

6. The BELCA record selection means calculates the cosine similarity based on the construction cost text data and the actual text data, Furthermore, the BELCA record selection means selects the BELCA record based on the cosine similarity. A maintenance cost calculation support system according to claim 3 or 4.

7. The BELCA record selection means selects the BELCA identifier from the construction cost text data by machine learning the performance text data included in the performance record and the learning data based on the BELCA identifier. Maintenance cost calculation support system as described in feature 1.

8. The aforementioned construction cost record is composed of construction cost units, The aforementioned performance record is composed of performance units, The BELCA record selection means selects the actual record which includes the actual unit which is the same unit as the construction cost unit included in the construction cost record. Maintenance cost calculation support system as described in feature 1.

9. A program that causes a computer to perform a function to assist in calculating the maintenance costs of a building, The BELCA storage means stores multiple BELCA records. The process involves importing the entered construction breakdown file, and A BELCA record selection process is performed to select the BELCA record using the aforementioned construction breakdown file and performance data consisting of multiple performance records. The system includes a function to cause the computer to perform a maintenance cost calculation process that calculates the renewal costs and repair costs related to the maintenance of the aforementioned building. The BELCA record is composed of a BELCA identifier, which is the identifier of the BELCA record, an update cycle, an update coefficient, a repair cycle, a repair unit price coefficient, and a repair quantity coefficient. The aforementioned construction breakdown file consists of multiple construction cost records, The aforementioned construction cost record consists of construction cost text data and itemized amounts, The aforementioned performance record comprises performance text data and the aforementioned BELCA identifier. In the BELCA record selection process, for each construction cost record included in the construction breakdown file, the actual record is selected based on the similarity between the construction cost text data and the actual text data. Furthermore, in the BELCA record selection process, a BELCA record relating to the same BELCA identifier as the BELCA identifier included in the selected performance record is selected from the BELCA storage means. In the maintenance cost calculation process, for each construction cost record included in the construction breakdown file, the renewal cycle and renewal coefficient included in the BELCA record selected in the BELCA record selection process are read, and the product of the itemized amount related to the construction cost record and the renewal coefficient is calculated as the renewal cost. Furthermore, in the maintenance cost calculation process, for each construction cost record included in the construction breakdown file, the repair cycle, the repair unit price coefficient, and the repair quantity coefficient included in the BELCA record selected in the BELCA record selection process are read, and the product of the itemized amount related to the construction cost record, the repair unit price coefficient, and the repair quantity coefficient is calculated as the repair cost. For each construction cost record, the renewal costs incurred for each renewal cycle and the repair costs incurred for each repair cycle can be determined. A maintenance cost calculation support program characterized by the following features.

Images