JP6904914B2 - Decision table generator and decision table generation method - Google Patents

Description

The present invention relates to a decision table generator and a decision table generation method.

Patent Document 1 describes "development of software specifications" in order to "develop technology for creating and verifying tabular software specifications" and "to achieve reduction of manpower for creating and decoding decision tables and reduction of storage capacity while ensuring readability". A device that supports, a proposition specification receiving means that accepts a plurality of propositional specifications consisting of a condition part having a plurality of conditions and a result part having a plurality of results, and each line is one of the above-mentioned plurality of conditions or the above-mentioned plurality of conditions. A decision table creating means for creating a decision table in a table format in which each column indicates each of the above-mentioned multiple proposition formats from the specifications of the above-mentioned multiple proposition formats, and a above-mentioned plurality of proposition formats. The means for creating a decision table is provided with a consistency verification means for determining whether or not each of the specifications is consistent with each other. When the above-mentioned propositional specification receiving means obtains the above-mentioned plurality of propositional format specifications, the above-mentioned plurality of propositional formats are obtained. For each of the specifications of the above propositional form, the condition part and the result part of the specification of the propositional form are taken out to create a row of the above decision table, and the column corresponding to the specification of the propositional form is marked in the column of the row, and the above consistency The sex verification means extracts any two propositional specifications out of the plurality of propositional specifications from the above determination table, determines whether or not the logical product of the condition part of the above two propositional specifications can be satisfied, and determines whether or not the logical product of the condition part of the above two propositional specifications can be satisfied. Provided is a device that outputs when each of the above-mentioned specifications of the plurality of propositional forms contradicts each other if it can be satisfied. "

Japanese Unexamined Patent Publication No. 2012-192203

Generally, an execution sequence of instructions in a source code (hereinafter referred to as an "execution path") is a branch instruction (hereinafter referred to as an "If statement") or an iterative instruction (hereinafter referred to as a "loop statement") in the source code. It depends on the number of) and increases exponentially in some cases. Therefore, when the number of If statements and loop statements increases and the scale of the source code becomes enormous, the path of the source code also increases exponentially, making it difficult for a person to check the source code.

Therefore, paying attention to a certain variable, the set of the value assignment to the variable (hereinafter referred to as "result") and the condition in which the assignment occurs (hereinafter referred to as "input condition") is regarded as a rule and the source. There is a method to improve the efficiency of checking the source code specifications by visualizing the rules extracted from the code in the form of a decision table.

Symbol execution is known as a method of extracting rules from source code. When extracting a rule from the source code, one execution path simply corresponds to one rule. Therefore, the number of extracted rules increases exponentially with the number of If statements and loop statements.

Patent Document 1 describes a technique for generating a decision table from a propositional specification (corresponding to a rule). However, when a huge number of rules are converted into a decision table, the generated decision table becomes huge and it becomes difficult for a person to confirm the decision table. Further, Patent Document 1 does not disclose anything regarding a technique for supporting confirmation of a huge decision table.

The present invention has been made in view of this background, and an object of the present invention is to make it possible to generate a partial decision table for a rule and to reduce the burden on a person who confirms the rule.

One of the present inventions for solving the above problems is a decision table generator, in which the software specifications are based on one or more rules including an input condition and a result which is information indicating the value of an output variable. An information storage unit that stores the rule specifications that are the described information and the generation range specification information that is the information that specifies the generation range of the decision table, and the range of the rule specifications that is specified by the generation range specification information. A decision table generator for generating a decision table is provided, and the component of the input condition is a term that is at least one of a comparison expression, a truth type variable, and a truth value, or a denial of the above term. The decision table generation unit is a logical expression obtained by combining a logical expression generated based on the input condition or the rule specification with a logical sum of clauses which are logically equivalent logical products of the constituent elements. Convert to the selection standard form, determine whether or not each component affects the input condition, exclude the component that does not affect the input condition from the input condition, and remove the decision table. Generate.

In addition, the problems disclosed in the present application and the solutions thereof will be clarified by the column of the form for carrying out the invention and the drawings.

According to the present invention, it is possible to generate a partial decision table for a rule, and it is possible to reduce the burden on the person who confirms the rule.

It is a figure which shows the hardware configuration of the decision table generator. It is a figure which shows the function which the decision table generator has. This is an example of a rule specification. This is an example of the generation result. This is an example of the setting screen. It is a flowchart explaining the decision table generation process. It is a flowchart explaining the decision table generation process of a difference. It is a flowchart explaining the whole decision table generation process. It is a flowchart explaining the logical expression equivalence conversion process.

Hereinafter, embodiments will be described with reference to the drawings. In the following description, the same or similar configurations may be designated by the same reference numerals and duplicate description may be omitted.

FIG. 1 shows the hardware configuration of the decision table generation device 100 described as one embodiment, and FIG. 2 shows the functions provided by the decision table generation device 100, respectively. The decision table generation device 100 is an information processing device (computer) used for developing and maintaining software (for example, those related to business such as finance, insurance, manufacturing, distribution, communication, education, medical care, etc.). The decision table generation device 100 may be realized by using virtual information processing resources such as a cloud server provided by a cloud system.

Generally, the specifications of software are determined in accordance with legal regulations regarding products and services, business rules and standards, business rules, system rules, check rules, and the like. Hereinafter, in the above specifications, an input condition and a set of results when the input condition is satisfied will be referred to as a "rule". The rules are, for example, a description in a natural sentence such as "If A = 1 And B = 2 Then R = 1" or a description according to a predetermined rule that can be interpreted by an information processing device such as "If A = 1 And B = 2 Then R = 1". It is expressed in various forms such as a description in a decision table that defines the correspondence between input conditions and results in a tabular format.

Each of the above input conditions and results is composed of one or more components. The component is, for example, any of a comparison expression, a true / false type variable, and a false value (hereinafter referred to as a "term"), or a negation of a term. In the following description, a term or a negation of a term is referred to as a "literal". The above input condition is represented by, for example, a logical expression in which literals are combined using logical operators (logical product (And), logical sum (Or), etc.). In addition, the above result is, for example, "
It is represented by a logical expression consisting of literals in the form of "variable = value", "variable = variable", or "variable = calculation formula containing 0 or more variables".

In the following description, the variables that appear in the result of the rule are referred to as "output variables", and the variables other than that of the rule are referred to as "input variables". A set of one or more rules having a common output variable is called a "rule set", and a set of rule sets is called a "rule specification". The rule specifications can also be automatically generated from the source code using an information processing device by, for example, a source code analysis technique such as symbol execution.

The decision table generation device 100 includes rule specifications and information for specifying a range (hereinafter, referred to as "generation range") for generating a decision table in the rule specifications (hereinafter, referred to as "generation range specification information"). Based on, a set of rules expressing the input conditions is generated by a logical expression in which the literals are combined by a logical product, and a decision table is generated based on the generated set.

The above generation range specific information is, for example, a rule specification (first rule specification) to be compared with respect to a rule specification (first rule specification) (hereinafter, referred to as "decision table generation target") which is a generation target of a decision table. 2 rule specifications) (hereinafter referred to as "comparison target") is information for specifying. The rule specifications to be compared above are, for example, the rule specifications (source code) before the modification (before the change) when the rule specification (source code) for which the decision table is generated is changed (modified).
Is.

Further, the above-mentioned generation range specifying information is, for example, information for specifying the generation range by evaluating the metric value acquired for the above-mentioned rule specification and the above-mentioned metric value among the rule specifications to be generated in the decision table ( For example, it is information including a threshold value and an evaluation formula (hereinafter, referred to as “metrics evaluation information”). The decision table generator 100 acquires, for example, a metric value from a source code obtained by converting a rule specification.

Further, the above-mentioned generation range specifying information is information including, for example, a logical expression (hereinafter, referred to as "filter condition") for extracting the rule specification of the generation range from the rule specification to be generated in the decision table. The above filter condition is described by a logical expression such as "70 <x <80 And y = True".

The above-mentioned metric values are, for example, "the number of rare literal combination appearances", "rule length", "distance", "number of included Elses", and the like. Of these, the "number of appearances of rare literal combinations" is the number of literal combinations that appear only in the number of rules below the threshold value among the literal combinations included in the rule specifications. The combination of literals is not limited to the combination of two literals, and may be a combination of three or more literals. The "rule length" is the number of literals included in the rule input condition. The "distance" is, for example, between two literals when there is a source code corresponding to the rule specification and which instruction in the above source code corresponds to the literal included in the input condition of the rule. This is the difference between the line numbers of the above instructions (description interval). When three or more literals are included in the input condition of the rule, for example, the difference between the line numbers of the corresponding instructions is obtained for all combinations of literals, and the average, maximum, minimum, etc. are set as distances. In addition to the difference in line numbers, for example, the distance between each instruction may be obtained based on whether or not each instruction corresponding to each literal is a different function or a different file. In addition, the "number of Else included" measures which instruction in the above source code corresponds to the literal included in the rule when there is a source code corresponding to the rule specification, and "If Ten" is added to the corresponding instruction. It is the number of Else instructions included in the "Else" type instruction, counted without allowing duplication.

Further, the above-mentioned metric evaluation information includes, for example, (1) a rule including a combination of literals having a low appearance frequency (the appearance frequency is equal to or less than a preset threshold value) among the rules included in the rule specification as a decision table generation target. , (2) A rule with a large number of literals included in the input condition (the input condition contains more than a predetermined number of literals) is targeted for decision table generation, and (3) the distance is long (more than a preset threshold value or more). ) The rule is the decision table generation target, (4) the rule containing a large number of Else (greater than or equal to the preset threshold) is the decision table generation target, and the like. Here, the reason why a rule including a combination of literals with a low frequency of appearance as in (1) is set as a decision table generation target is, for example, to visualize a part of the rule specification that is likely to be overlooked as a decision table. .. Also, as in (2), a rule with many input conditions is targeted for decision table generation. For example, since errors are likely to occur in complicated parts with many input conditions, this is visualized as a decision table and attention is paid to the user. This is to encourage. Also, when a rule with a long distance is targeted for decision table generation as in (3), errors are likely to occur if the distance is long (for example, long source code or source code divided into multiple files fits on one screen. This is because the decision table is generated and visualized to draw the user's attention. Also, as in (4), a rule with a large number of Else included in the rule is targeted for decision table generation when the user describes the source code in the form of "If condition Ten process 1 Else process 2". Since the corresponding condition is not explicitly displayed on the source code and the user may not be aware of the condition corresponding to Else, if the number of Else included in the rule is large, an error is likely to occur and a decision table is generated. This is to visualize and call attention to the user.

As shown in FIG. 1, the decision table generation device 100 includes a processor 110, a main storage device 120, an auxiliary storage device 130, an input device 140, an output device 150, and a communication device 160. These are connected to each other so as to be able to communicate with each other via a communication means such as a bus (not shown).

The processor 110 is configured by using, for example, a CPU (Central Processing Unit) or an MPU (Micro Processing Unit). When the processor 110 reads and executes the program stored in the main storage device 120, various functions of the decision table generation device 100 are realized.

The main storage device 120 is a device that stores programs and data, and is, for example, a ROM (Read Only Memory), a RAM (Random Access Memory), a non-volatile semiconductor memory (NVRAM (Non Volatile RAM)), and the like.

The auxiliary storage device 130 includes, for example, a hard disk drive, an SSD (Solid State Drive), an optical storage device (CD (Compact Disc), DVD (Digital Versatile Disc), etc.).
, A storage system, an IC card, a reading / writing device for a recording medium such as an SD memory card or an optical recording medium, a storage area of a cloud server, and the like. Programs and data stored in the auxiliary storage device 130 are read into the main storage device 120 at any time.

The input device 140 is, for example, a keyboard, a mouse, a touch panel, a card reader, a voice input device, and the like.

The output device 150 is a user interface that provides the user with various information such as processing progress and processing results. For example, an image display device (liquid crystal monitor, LCD (Liquid Crystal Display), graphic card, etc.), an audio output device (speaker). Etc.), printing equipment, etc. In addition, for example, the decision table generation device 100 may be configured to input and output information to and from other devices via the communication device 160.

The communication device 160 is a wired or wireless communication interface that realizes communication with other devices via a communication means such as a LAN (Local Area Network) or the Internet, and is, for example, a NIC (Network Interface Card). , Wireless communication module, USB (Universa)
l Serial Interface) module, serial communication module, etc.

As shown in FIG. 2, the decision table generation device 100 includes the functions of the information storage unit 205, the rule specification generation unit 210, the decision table generation unit 220, the decision table inverse conversion unit 230, and the metric value acquisition unit 240.

The information storage unit 205 stores the source code 260, the rule specification 270, the generation range specifying information 280, and the generation result 290. The source code 260 is, for example, a source code input by the user to the decision table generator 100. The rule specification 270 is, for example, a rule specification input to the decision table generation device 100 by the user, a rule specification automatically generated (extracted) by the rule specification generation unit 210 from the source code stored in the source code 260, and a decision table inverse conversion unit. 230 is a rule specification or the like automatically generated by reverse conversion from the decision table. It should be noted that how to give the rule specifications to the decision table generator 100, that is, the input method of the rule specifications is not necessarily limited. As a rule specification input method, for example, a method of directly inputting a rule specification, a method of generating a decision table from the above source code by inputting a source code written in a predetermined language, and a method of inputting a decision table. There is a method in which the decision table generator 100 reversely converts the decision table into the rule specifications. The generation range specifying information 280 is, for example, information acquired via the setting screen 500 described later. The generation result 290 is information such as a decision table generated by the decision table generation unit 220 based on the rule specification 270 and the generation range specifying information 280.

Among the above functions, the rule specification generation unit 210 extracts a rule specification from the source code 260 by a source code analysis technique such as symbol execution, and stores the extracted rule specification as the rule specification 270. The decision table generation unit 220 generates a decision table based on the rule specification 270 and the generation range specifying information 280, and stores the generated decision table as the generation result 290. The decision table inverse conversion unit 230 reversely converts the decision table into the rule specifications. One row of the decision table corresponds to one rule. The decision table inverse conversion unit 230 reversely converts the decision table into the rule specifications by converting one row of the decision table into the format of "If input condition Ten result". The metric value acquisition unit 240 acquires the metric value from the source code 260.

FIG. 3 shows an example of the rule specification 270. As shown in the figure, the rule specification 270 includes one or more rule sets having the same result variable. An identifier (hereinafter referred to as "rule set ID") is assigned to each rule set. Each rule set contains one or more rules described in the "If-Then" format. An identifier (hereinafter referred to as "rule ID") is assigned to each rule.

The rule specification 270 shown in the figure includes a rule set 310 having a rule set ID of “RS1” and a rule set 320 having a rule set ID of “RS2”. The rule set 310 and the rule set 320 each include four rules having rule IDs "1" to "4".

When the specific value of the input variable is defined in each rule set, it is uniquely determined whether the input condition of each rule is satisfied (true). When the input condition of the rule is satisfied, the value of the output variable is determined so that the result of the rule is satisfied. For example, in the rule set 310 illustrated in FIG. 3, when the specific values of the input variables "x", "y", and "z" are "100", "-1", and "-1", respectively. The input condition of the rule whose rule ID is "4" is satisfied, and the specific value of the output variable "out" is set to "0".

FIG. 4 is an example of the generation result 290 generated by the decision table generation unit 220 and stored by the information storage unit 205. The generated result 290 contains one or more decision tables. An identifier (hereinafter referred to as "decision table ID") is assigned to each decision table. The generation result 290 shown in the figure includes four decision tables whose decision table IDs are “DT1” to “DT4”, respectively. The decision table contains one or more rows, and each row is given an identifier (hereinafter referred to as "row number"). Each row of the decision table includes an input condition "If" and a result "Then", and one row corresponds to one rule. For example, the second row of the decision table 410 corresponds to the rule "If x <= 50 And y> = 0 Then out = 1", and the third row of the decision table 410 corresponds to "If 0 <x And x <= 50". Corresponds to the rule "Then out = 2". The cells in each row correspond to zero or more literals with the same variable. For example, cell 411 of the decision table 410 corresponds to the literal "x <= 50", cell 412 of the decision table 410 corresponds to the literal "y> = 0", and the decision table 41
Cell 413 of 0 corresponds to literals "0 <x" and "x <= 50". The "-" in cell 414 of the decision table 410 means that the value of the variable "y" is irrelevant.

The decision table 410 is a decision table generated based on the rule set 310 shown in FIG. Here, if the rule ID of the rule set 310 is "x>50" from the rules of "1" and "2", "out" becomes "1". That is, even if "y>0","x>0","x>100", and "Not (x>100)" are removed from these rules, it affects the input condition that "out" becomes "1". There is no. In this way, the decision table generation unit 220 removes literals (hereinafter, referred to as "don't care") that do not affect the input conditions even if they are removed when the decision table is generated. In the case of the first row of the decision table 410, all literals other than "x>50" are removed as don't care, so the decision table row number "1"
"Is the content corresponding to" If x> 50 Then out = 1 ".

In the decision table 420, the rule set 310 shown in FIG. 3 is used as the decision table generation target, and the rule set 320 is used as the comparison target. This is an example in which the combination of the "difference input condition") and the result of each rule set in the difference input condition (hereinafter, the combination of the difference input condition and the above result is referred to as "difference") is used as a decision table. .. Comparing the ruleset 310 with the ruleset 320, the literals "x>100","x<=100" (Not (x> 100)), "x>50","x<=" contained in the ruleset 310 50 "(Not (x> 50)), respectively, in rule set 320,"x> = 100 "," x <100 "(Not (x> = 100)),"x> = 50 "," x <50 " (Not (x> = 50)), but only when "50 <= x And x <= 50", the rule set 310 has "out" as "2", while the rule set 320 has. “Out” becomes “1”, and both have different results. The first row of the decision table 420 shows the difference between the two as a decision table, and "If 50 <= x And x <= 50 Then out (RS1 (rule set 310)) = 2 And out (RS2 (rule)) The content corresponds to the rule "Set 320)) = 1".

The decision table 430 is an example in which the rule set 310 shown in FIG. 3 is used as the decision table generation target and the decision table is generated for the range satisfying the filter condition “out = 2 And y <0”. Decision table generator 22
0 generates a decision table for the range satisfying the filtering condition when the filtering condition is specified in the generation range specific information 280. The decision table 430 is an example in which the decision table is generated only for the third row satisfying the filtering condition “out = 2 And y <0” in the decision table 410 which is the entire decision table of the rule set 310.

The decision table 440 targets the rule set 310 shown in FIG. 3 as the decision table generation target, and the range in which the length of the rule, which is one of the metric values acquired by the metric value acquisition unit 240 in FIG. 2, satisfies the threshold value “2” or more. This is an example of generating a decision table for only. The decision table generation unit 220 generates a decision table for a range of the rule set 310 in which the metric value is equal to or more than the threshold value (or less than or equal to the threshold value), for example, when the generation range specific information 280 is set to the metric value and the threshold value. .. The decision table 430 is an example in which the decision table is generated only for the second row, the third row, and the fourth row in which the rule length is 2 or more, out of 410 which is the entire decision table of the rule set 310. ..

FIG. 5 is an example of a screen (hereinafter, referred to as “setting screen 500”) that the decision table generation unit 220 outputs to the output device 150 when the decision table is generated. As shown in the figure, the setting screen 500
Is provided with a decision table generation target input field 510, a comparison target input field 520, a rule metric input field 530, a filter condition input field 540, and a decision table generation button 550.

Information for designating the decision table generation target is set in the decision table generation target input field 510. As shown in the figure, the generation target input field 510 of the decision table includes a text form 511, a radio button 512, and a text form 513. In the text form 511, the file name of the file including the data (source code, specification rule, or decision table) for which the decision table is generated is set. The radio button 512 accepts the specification of the input file type (either the source code, the specification rule, or the decision table). In the text form 513, the output variable name is set when "source code" is selected as the input file type by the radio button 512, and when other than that is selected, the rule specification of the decision table generation target is set. The rule set ID is set. In the example of the figure, "rule.cbl" is set in the text form 511, the source code is selected by the radio button 512, and the output variable "out" is set in the text form 513. The text form 511 and the text form 513 may be used as a dialog box or the like. Further, the radio button 512 may be used as a list box or the like.

In the comparison target input field 520, information necessary when the comparison target is used to specify the generation range of the decision table is set. As shown in the figure, the comparison target input field 520 includes a check box 521, a text form 522, a radio button 523, and a text form 524. The check box 521 is set to indicate the user's intention as to whether or not to use the comparison target to specify the generation range of the decision table, and when the comparison target is used, the user checks the check box 521. In the text form 522, the file name of the file including the data to be compared (source code, specification rule, or decision table) is set. Radio button 523 accepts the specification of the input file type (either source code, specification rule, or decision table). In the text form 524, the output variable name is set when "source code" is selected as the input file type by the radio button 523, and the rule set ID of the rule specification to be compared is set in other cases. Will be done. In the example of the figure, the check box 521 is not checked, "old_rule.smt2" is set in the text form 522, the specification rule is selected by the radio button 523, and the rule set ID "RS2" is set in the text form 524. Is set. Text form 5
22 and the text form 524 may be a dialog box or the like. Further, the radio button 523 may be used as a list box or the like.

In the rule metric input field 530, information necessary for specifying the generation range of the decision table by the metric is set. As shown in the figure, the rule metric input field 530 includes a check box 531, a metric input table 532, and a one-row addition button 533. The check box 531 is set to indicate the user's intention whether or not to use the metric to specify the generation range of the decision table, and when the metric is used, the user checks the check box 531. In the metric input table 532, a metric name and a threshold value are set as metric evaluation information. The one-line addition button 533 is operated by the user when adding the number of lines in the metric input table 532. In the example of the figure, the check box 531 is not checked, and the metric name “rule length” and the threshold value “2” are input to the metric input table 532. The list box for specifying the length of the metric may be a text form or the like, and the text form for inputting the threshold value may be a list box or the like.

In the filter condition input field 540, information necessary for using the filter condition as information for specifying the generation range of the decision table is set. As shown in the figure, the filter condition input field 540 includes a check box 541 and a text form 542. The check box 541 is set to indicate the user's intention whether or not to use the filter condition to specify the generation range of the decision table, and when the filter condition is used, the user checks the check box 541. The filter condition is input to the text form 542. In the example of the figure, the check box 541 is checked and the filter condition "out = 2 And y <0" is set in the text form 542.

After setting the necessary information on the setting screen 500, when the user operates the decision table generation button 550, the decision table generation device 100 starts the process for generating the decision table. In the above process, first, the decision table generation unit 220 sets the contents of the setting screen 500 (decision table generation target, decision table, comparison target, metric name, threshold value, filter condition, etc.) as the generation range specific information 280, and the information storage unit 205. To memorize. Subsequently, the decision table generation unit 220 starts the generation of the decision table for the rule set for which the decision table is generated, using the generation range specifying information 280. When the decision table generation target or the comparison target is specified in the source code on the setting screen 500, the decision table generation unit 220 uses the rule specification generation unit 210 from the source code before the decision table generation process. The rule specifications are extracted, and the extracted rule specifications are stored in the information storage unit 205 as the rule specifications 270. Further, for example, when a generation target or a comparison target is specified in the decision table, the decision table generation unit 220 reverse-converts the decision table into a rule specification using the decision table inverse conversion unit 230, and the reverse-converted rule specification is ruled. It is stored in the information storage unit 205 as the specification 270. The extraction of the specification rule and the inverse transformation of the decision table may be performed before the information storage unit 205 stores the setting screen 500 as the generation range specifying information 280.

FIG. 6 is a flowchart illustrating a process (hereinafter, referred to as “decision table generation process S600”) performed by the decision table generation unit 220 when generating the decision table. Hereinafter, the decision table generation process S600 will be described with reference to the figure. The decision table generation unit 220 starts the decision table generation process S600, for example, triggered by a user's instruction operation. Further, the decision table generation unit 220 automatically executes the decision table generation process S600 by batch processing or the like at a scheduled timing, for example. In the following, the letter "S" attached before the code means a processing step.

First, the decision table generation unit 220 reads the rule specification 270 (rule set) for which the decision table is generated (S601) and also reads the generation range specifying information 280 (S602).

Subsequently, the decision table generation unit 220 determines whether or not the comparison target is set as the generation range specifying information 280 (S603). If the comparison target is set (S603: YES), the process proceeds to S604, and if the comparison target is not set (S603: NO), the process proceeds to S605.

In S604, the decision table generation unit 220 performs a process of generating a decision table based on the above-mentioned difference (hereinafter, referred to as “difference decision table generation process S604”). The details of the difference determination table generation process S604 will be described later. After the completion of the difference determination table generation process S604, the process proceeds to S606.

In S605, the decision table generation unit 220 performs a process of generating a decision table based on the entire decision table generation target (hereinafter, referred to as “whole decision table generation process S605”). Details of the entire decision table generation process S605 will be described later. After the completion of the entire decision table generation process S605, the process proceeds to S607.

In S606, the decision table generation unit 220 determines whether or not the result (return value) of the difference decision table generation process S604 is "no difference". If the result of the difference decision table generation process S604 is "no difference", the decision table generation unit 220 ends the decision table generation process S600. If the result of the difference determination table generation process S604 is not "no difference", the process proceeds to S607.

In S607, the decision table generation unit 220 determines whether or not the filter condition is set (whether or not the generation range specifying information 280 includes the setting of the filter condition). When the filter condition is set in the generation range specific information 280 (S607: YES), the process proceeds to S608, and when the filter condition is not set (S607: NO), the process proceeds to S609.

In S608, the decision table generation unit 220 excludes a portion (row) that does not satisfy the filter condition from the decision table generated in S604 or S605. More specifically, the decision table generator 220 determines whether the "filter condition And input condition And result" can be true for each row of the decision table (hereinafter, the fact that it can be true is referred to as "satisfiable" and is true. (It is called "unsatisfiable"), and if it is unsatisfactory, the row is excluded from the decision table. There is an SMT solver (Satisfiability Modulo Theories Solver) as a technique for determining the satisfiability of a logical expression. The decision table generation unit 220 determines the satisfiability by using, for example, an SMT solver or the like.

Subsequently, the decision table generation unit 220 determines whether or not the information related to the metric is set as the generation range specifying information 280 (S609). If the information about the metric is set (S609: YES), the process proceeds to S610, and if the information about the metric is not set (S609: NO), the process proceeds to S612.

In S610, the metric value acquisition unit 240 acquires the metric value for each row of the decision table. Subsequently, the decision table generation unit 220 excludes rows having a metric value equal to or less than a preset threshold value (or more than the above threshold value) from the decision table (S611). After that, the process proceeds to S612.

In S612, the decision table generation unit 220 stores the decision table generated by the above processing as the generation result 290. The execution order of the processes S603 to S606 and the processes S607 to S611 in FIG. 6 may be reversed.

FIG. 7 is a flowchart illustrating the “difference determination table generation process S604” in FIG. Hereinafter, the difference determination table generation process S604 will be described with reference to the figure.

First, the decision table generation unit 220 determines whether or not the output variable of the decision table generation target and the output variable of the comparison target match (variable names match) (S701). If they match (S701: YES), the process proceeds to S702, and if they do not match (S702: NO), the process proceeds to S703.

In S702, the decision table generation unit 220 makes the output variables different between the decision table generation target and the comparison target. For example, the decision table generation unit 220 replaces the output variable in the result of each rule of the decision table generation target and the comparison target with "output variable (rule set ID of the corresponding rule set)". For example, when the decision table generation target is the rule set 310 of FIG. 3 and the comparison target is the rule set 320 of FIG. 3, the decision table generation unit 220 sets the output variables to "out (RS1)" and "out (out (RS1)", respectively. Replace with RS2) ”. The method of replacing the output variable is not necessarily limited to the above method.

Subsequently, the decision table generation unit 220 expresses the difference between the decision table generation target and the comparison target by a logical expression (S703). Specifically, the decision table generator 220 first joins the input condition and the result by a logical product for each rule included in the rule set, and further joins them by a logical sum to generate the decision table and the decision table. Convert the comparison target to a logical expression. For example, the rule set 310 in FIG. 3 is "(((y> 0 Or x> 0) And x> 100) And (out = 1)) Or (((y> 0 Or x> 0) And Not (x> 100) And x> 50) And (out = 1)) Or (((y> 0 Or x> 0) And Not (x> 100) And Not (x> 50)) And (out = 2)) Or ( (Not (y> 0
Or x> 0)) And (out = 0)) ”is converted to the formula. It should be noted that this logical expression means that the input condition and the result are satisfied in one or more of the rules whose rule IDs are "1" to "4" included in the rule set 310.

Subsequently, the decision table generation unit 220 generates a logical expression representing the difference using the logical expression of the rule set. Here, the difference is a combination of an input condition that leads to a result in which the rule set of the decision table generation target and the comparison target are different and the result in the input condition.
It can be expressed by the logical expression "And Not (output variable for decision table generation = output variable for comparison)". This formula represents a case where the value of the output variable of the decision table generation target and the value of the output variable of the comparison target are different among the input conditions and results that satisfy the decision table generation target and the comparison target.

Subsequently, the decision table generation unit 220 joins the logical expressions representing the differences by the logical product of the literals (hereinafter, the logical expression in which the literals are connected by the logical product is referred to as a "section"), and the clauses are ORed. It is converted into an equivalent logical expression of the combined form (hereinafter, referred to as "disjunctive normal form") (S704). The details of this conversion process (hereinafter, referred to as “logical formula equivalent conversion process S704”) will be described later.

Subsequently, the decision table generation unit 220 determines whether or not the result (return value) of the logical expression equivalent conversion process S704 is "the disjunctive normal form is false (no difference)" (S705). When the result of the logical expression equivalent conversion process S704 is "the disjunctive normal form is false (no difference)" (S705: YES), the decision table generation unit 220 is the result of the process (difference decision table generation process S604). The process is terminated with (return value) set to "no difference", and then the process proceeds to S606 of the decision table generation process S600 of FIG. On the other hand, when the result of the logical expression equivalent conversion process S704 is "the disjunctive normal form is not false (with a difference)" (S705: NO), the process proceeds to S706.

In S706, the decision table generation unit 220 repeatedly executes the processes of S707 to S710 for all the sections of the disjunctive normal form generated in S704. When the processing of S707 to S710 is completed for all the sections of the disjunctive normal form, the decision table generation unit 220 sets the result (return value) of the processing (difference decision table generation processing S604) as "with difference" and performs the processing. finish. After that, the process proceeds to S606 of the decision table generation process S600 of FIG.

In S707, the decision table generation unit 220 acquires a combination of the difference input condition and the result in the difference input condition (hereinafter, referred to as “result pair”). The decision table generator 220 uses a logical expression in which all the literals included in the clause that do not include the output variable are combined by a logical product as a difference input condition, and sets all the literals included in the clause that include the output variable as the difference input condition. Make a result pair. Each section of the disjunctive normal form generated in S704 includes one literal including an output variable to be generated as a decision table and one literal including an output variable to be compared.

In S708, the decision table generation unit 220 determines whether or not the discrepancy in the values of the output variables is obvious in the literal included in the result pair generated in S707. Specifically, for example, the decision table generation unit 220 determines the satisfiability of "Not (difference input condition-> a literal in which the right side of each literal included in the result pair is connected by an inequality sign)". When the above formula cannot be satisfied, the formula "(difference input condition-> literal in which the right side of each literal included in the result pair is combined by an inequality number" excluding negation from the above formula is always true. That is, the difference. If the input conditions are satisfied, the values assigned to the output variables to be generated and the output variables to be compared are always different. Therefore, the decision table generation unit 220 assumes that the discrepancy between the values of the output variables is obvious. When the logical expression can be satisfied, the determination table generation unit 220 determines that the mismatch of the values of the output variables is not obvious. When the determination table generation unit 220 determines that the mismatch of the values of the output variables is obvious (S708: YES), the process proceeds to S710. When the determination table generation unit 220 determines that the discrepancy between the values of the output variables is not obvious (S708: NO), the process proceeds to S709.

In S709, the decision table generation unit 220 extracts the right side from each literal included in the result pair, generates a literal in which each of the right sides is connected by an inequality sign (hereinafter, referred to as "result mismatch literal"), and the above result. Combine unmatched literals with differential input conditions as a logical product.

In S710, the decision table generation unit 220 sets the difference input condition and the result pair as the difference in the decision table.

FIG. 8 is a flowchart illustrating “whole decision table generation process S605” in FIG. Hereinafter, the entire decision table generation process S605 will be described together with the figure.

First, the decision table generation unit 220 groups rules having the same result with respect to the rules included in the decision table generation target (S801). Hereinafter, a set of rules having the same result will be referred to as a "rule group".

Subsequently, in S802, the decision table generation unit 220 repeatedly executes the processes of S803 to S809 for all the rule groups generated in S801.

In S803, the decision table generation unit 220 generates a logical expression (hereinafter, referred to as “integrated input condition”) in which the input conditions of all the rules in the rule group are ORed together.

Subsequently, the decision table generation unit 220 converts the generated integrated input condition into the disjunctive normal form (S804). The details of this conversion process (hereinafter, referred to as “logical formula equivalent conversion process S804”) will be described later.

Subsequently, the decision table generation unit 220 determines whether or not the result of the logical expression equivalent conversion process S804 is "the disjunctive normal form is false" (S805). When the result of the logical expression equivalent conversion process S804 is "the disjunctive normal form is false" (S805: YES), the decision table generation unit 220 skips the processes of S806 to S809. If the result of the logical expression equivalent conversion process S804 is not "the disjunctive normal form is false" (S805: NO), the process proceeds to S806.

In S806, the decision table generation unit 220 extracts the clauses of the disjunctive normal form as input conditions, respectively. At this time, one or more input conditions are extracted.

Subsequently, the decision table generation unit 220 extracts the result from the rules in the rule group (S807). At this time, all the rules in the rule group have the same result, and only one result is extracted.

Subsequently, the decision table generation unit 220 generates a rule (hereinafter, referred to as "decision table rule") using the input condition extracted in S806 as the input condition and the result extracted in S806 as the result (S808). One decision table rule is generated for each input condition extracted in S806.

Subsequently, the decision table generation unit 220 sets the generated decision table rule in the decision table (S809).

FIG. 9 shows the above-mentioned logical formula equivalent conversion process (“logical formula equivalent conversion process S704” in FIG. 7 or “logical formula equivalent conversion process S804” in FIG. 8; hereinafter, referred to as “logical formula equivalent conversion process S900”). It is a flowchart explaining. Hereinafter, the logical expression equivalent conversion process S900 will be described with reference to the figure.

First, the decision table generation unit 220 extracts terms from the logical expression to be converted (S901).

Subsequently, the decision table generation unit 220 generates negations of all the terms extracted in S901. Hereinafter, the set of negations of the terms extracted in S901 and the terms generated in S902 will be referred to as a "literal set" (S902). The decision table generation unit 220 does not generate a negation for a term in which a negation exists at the time of extraction in S901.

Subsequently, the decision table generation unit 220 determines whether or not the negation of the logical expression to be converted can be satisfied (S903). If the negation of the logical expression to be converted is unsatisfactory (S903: NO), the process proceeds to S904, and if the negation of the logical expression to be converted is satisfiable (S903: YES), the process proceeds to S905.

In S904, the decision table generation unit 220 ends the logical expression equivalent conversion process S900 with "True" as the disjunctive normal form (return value) after conversion.

In S905, the decision table generation unit 220 determines whether or not the logical expression to be converted can be satisfied. If the logical expression to be converted cannot be satisfied (S905: NO), the process proceeds to S906, and if it can be satisfied (S905: YES), the process proceeds to S907.

In S906, the decision table generation unit 220 ends the logical expression equivalent conversion process S900 with "False" as the disjunctive normal form (return value) after conversion.

In S907, the decision table generation unit 220 acquires the assignment of the value to the variable to which the logical expression to be converted satisfies (hereinafter, referred to as “satisfaction solution”). The decision table generation unit 220 acquires a satisfied solution by using, for example, an SMT solver having a function of acquiring a satisfied solution of a logical expression that can be satisfied.

Subsequently, the decision table generation unit 220 acquires all the literals to be satisfied in the satisfying solution acquired in S907 (S908). Hereinafter, in the satisfaction solution acquired in S907, the set of literals to be satisfied is referred to as a "satisfaction clause".

Subsequently, the decision table generator 220 excludes don't care from the sufficiency node (S909). The determination table generation unit 220 excludes one literal from the satisfied clause, and can satisfy a logical expression in which all the literals included in the excluded satisfied clause and the negation of the logical expression to be converted are combined by a logical product. Judge whether or not. When it is determined that the satisfiability cannot be satisfied, the decision table generation unit 220 determines that the satisfied clause after exclusion satisfies the logical expression to be converted, and determines that the excluded literal is don't care. The decision table generation unit 220 determines in order whether or not the literals included in the sufficiency clause are don't care, and if it is don't care, excludes the don't care from the sufficiency clause. At this time, the inclusion relationship between the literals may be examined, and the order of determining whether or not the literal care may be performed may be determined based on the inclusion relationship.

Subsequently, the decision table generation unit 220 adds the sufficiency clause excluding the don't care in S906 to the set of sufficiency clauses (hereinafter, referred to as "sufficiency clause set") (S910).

Subsequently, the decision table generation unit 220 joins the duplication prohibition formula with the logical expression to be converted by a logical product (S911). That is, in order to prevent the generation of similar satisfied clauses, the decision table generation unit 220 joins all the literals included in the satisfied clauses by a logical product, and converts the logical expression that negates the whole as a duplicate prohibition expression. It is combined with the logical expression of.

Subsequently, the decision table generation unit 220 determines whether or not the logical expression to be converted can be satisfied (S9).
12). When the logical expression to be converted can be satisfied (S912: YES), the process returns to S907, and when the logical expression to be converted cannot be satisfied (S912: NO), the process proceeds to S913.

In S913, the decision table generation unit 220 generates a logical expression in which all literals are combined by a logical product for all the satisfied clauses included in the satisfied set clause, and all of the above logical expressions are joined by a logical sum. Generate an expression and use this as the return value. After that, the process proceeds to the process following the process of the caller. That is, in the case of the logical expression equivalent conversion process S704 of FIG. 7, the process proceeds to S705, and in the case of the logical expression equivalent conversion process S804 of FIG. 8, the process proceeds to S805.

As described above, the decision table generation device 100 generates a decision table for the range specified by the generation range specifying information 280 among the rule specifications for which the decision table is generated, and therefore is included in the rule specifications, for example. When the number of rules is huge, it is possible to reduce the burden on the person who checks the rule specifications.

Further, the decision table generator 100 is, for example, among the rule specifications for which the decision table is generated, the same input condition that produces different results, and the result and comparison target of the rule specification for which the decision table is generated under the input condition. Generate a decision table for the combination (difference) with the result of the rule specification to be performed. Therefore, the decision table can be generated only for the part that the user wants to check (for example, the part related to before and after the change (correction) of the rule specification (source code)), and the user can check the rule specification efficiently. Can be done.

Further, the decision table generation device 100 generates a decision table for the generation range specified based on, for example, the metric value and the metric evaluation information. Therefore, the user can generate a decision table for the part to be confirmed by appropriately setting the metric value and the metric evaluation information, and the user can efficiently confirm the rule specifications.

Further, the decision table generation device 100 generates a decision table for a generation range specified by applying a filter condition to a rule specification, for example. Therefore, the user can generate a decision table for the part to be confirmed by setting the filter conditions appropriately, and the user can efficiently confirm the rule specifications.

Although the present invention has been specifically described above based on the embodiments, it goes without saying that the present invention is not limited to the above embodiments and can be variously modified without departing from the gist thereof. do not have. For example, the above-described embodiment has been described in detail in order to explain the present invention in an easy-to-understand manner, and is not necessarily limited to the one including all the described configurations. Further, it is possible to add / delete / replace a part of the configuration of the above embodiment with another configuration.

For example, in the above, the case where the number of literals included in the result of the rule is one has been described as an example, but "variable = value", "variable = variable", "variable = calculation formula including 0 or more variables", etc. By combining the literals of the form by a logical product, it is possible to express a rule containing two or more literals in the result by the same logical expression as described above. If the result has a rule containing two or more literals, the result is divided by the logical product, each literal after the division is used as the result, and the input condition of the rule containing two or more literals in the above result is input. By generating the rule of, it is possible to convert a rule containing two or more literals in the above result into a rule containing one literal in the result.

Further, each of the above configurations, functional units, processing units, processing means and the like may be realized by hardware by designing a part or all of them by, for example, an integrated circuit. Further, each of the above configurations, functions, and the like may be realized by software by the processor interpreting and executing a program that realizes each function. Information such as programs, tables, and files that realize each function can be placed in a memory, a hard disk, a recording device such as an SSD (Solid State Drive), or a recording medium such as an IC card, an SD card, or a DVD.

Further, in each of the above figures, the control lines and information lines are shown as necessary for explanation, and not all the control lines and information lines in the implementation are necessarily shown. For example, in practice almost all configurations may be considered interconnected.

Further, the arrangement form of the various functional units, the various processing units, and the various databases of the decision table generation device 100 described above is only an example. The arrangement form of the various functional units, the various processing units, and the various databases can be changed to the optimum arrangement form from the viewpoints of the performance, processing efficiency, communication efficiency, and the like of the hardware and software included in the decision table generator 100.

Further, the above-mentioned configurations of various databases (schema, etc.) can be flexibly changed from the viewpoints of efficient use of resources, improvement of processing efficiency, improvement of access efficiency, improvement of search efficiency, and the like.

100 Decision table generator, 205 Information storage unit, 210 Rule specification generator, 220 Decision table generator, 230 Decision table inverse conversion unit, 240 Metrics value acquisition unit, 260 Source code, 270 Rule specifications, 280 Generation range identification information, 290 generation result, 500 setting screen, S600 decision table generation process, S604 difference decision table generation process, S605 entire decision table generation process, S900 logical expression equivalent conversion process

Claims (12)

The rule specifications, which are information that describes the software specifications by one or more rules, including the input conditions and the results, which are the information indicating the values of the output variables, and the rule specifications.
Generation range identification information, which is information that specifies the generation range of the decision table
Information storage unit that stores
Of the rule specifications, a decision table generation unit that generates a decision table for the range specified by the generation range specific information, and
With
The components of the input condition are a term that is at least one of a comparison expression, a true / false variable, and a false value, or a negation of the term.
The decision table generation unit
A logical expression generated based on the input condition or the rule specification is converted into a disjunctive normal form which is a logical expression in which clauses which are logically equivalent logical products of the constituent elements are combined by a logical sum.
It is determined for each component whether or not it affects the input condition, and the component that does not affect the input condition is excluded from the input condition to generate the decision table.
Decision table generator. The decision table generator according to claim 1.
The generation range specifying information includes information for specifying a second rule specification to be compared with respect to the first rule specification to be generated in the decision table.
The decision table generation unit has the same input condition that results in different results among the first rule specification and the second rule specification, the result of the first rule specification under the input condition, and the second rule specification. Generate a decision table for the combination with the result of the rule specification,
Decision table generator. The decision table generator according to claim 1.
The generation range specifying information includes a metric value acquired for the rule specification that is a generation target of a decision table, and metric evaluation information that is information for evaluating the metric value and specifying the generation range.
The decision table generation unit specifies the generation range based on the metric value and the metric evaluation information, and generates a decision table for the specified generation range.
Decision table generator. The decision table generator according to claim 3.
The metric evaluation information includes a threshold value to be compared with the metric value.
The decision table generation unit generates a decision table for the generation range specified by comparing the metric value with the threshold value.
Decision table generator. The rule specifications, which are information that describes the software specifications by one or more rules, including the input conditions and the results, which are the information indicating the values of the output variables, and the rule specifications.
Generation range identification information, which is information that specifies the generation range of the decision table
Information storage unit that stores
Of the rule specifications, a decision table generation unit that generates a decision table for the range specified by the generation range specific information, and
With
The components of the input condition are a term that is at least one of a comparison expression, a true / false variable, and a false value, or a negation of the term.
The generation range specifying information includes a metric value acquired for the rule specification that is a generation target of a decision table, and metric evaluation information that is information for evaluating the metric value and specifying the generation range.
The decision table generation unit specifies the generation range based on the metric value and the metric evaluation information, and generates a decision table for the specified generation range.
The metric value is the number of combinations of the components of the input condition.
The metric evaluation information includes information that specifies the rule including the combination that appears only in a number of rules equal to or less than a preset threshold value as the generation range.
Decision table generator. The rule specifications, which are information that describes the software specifications by one or more rules, including the input conditions and the results, which are the information indicating the values of the output variables, and the rule specifications.
Generation range identification information, which is information that specifies the generation range of the decision table
Information storage unit that stores
Of the rule specifications, a decision table generation unit that generates a decision table for the range specified by the generation range specific information, and
With
The components of the input condition are a term that is at least one of a comparison expression, a true / false variable, and a false value, or a negation of the term.
The generation range specifying information includes a metric value acquired for the rule specification that is a generation target of a decision table, and metric evaluation information that is information for evaluating the metric value and specifying the generation range.
The decision table generation unit specifies the generation range based on the metric value and the metric evaluation information, and generates a decision table for the specified generation range.
The metric value is the number of the components of the input condition.
The metric evaluation information includes information that specifies the rule whose number of components is equal to or greater than a preset threshold value as the generation range.
Decision table generator. The rule specifications, which are information that describes the software specifications by one or more rules, including the input conditions and the results, which are the information indicating the values of the output variables, and the rule specifications.
Generation range identification information, which is information that specifies the generation range of the decision table
Information storage unit that stores
Of the rule specifications, a decision table generation unit that generates a decision table for the range specified by the generation range specific information, and
With
The components of the input condition are a term that is at least one of a comparison expression, a true / false variable, and a false value, or a negation of the term.
The generation range specifying information includes a metric value acquired for the rule specification that is a generation target of a decision table, and metric evaluation information that is information for evaluating the metric value and specifying the generation range.
The decision table generation unit specifies the generation range based on the metric value and the metric evaluation information, and generates a decision table for the specified generation range.
The rule specifications are generated based on the source code.
The metric value is a distance in the source code of an instruction corresponding to the component of the input condition of the rule.
The metric evaluation information includes information that specifies the rule whose distance is equal to or greater than a preset threshold value as the generation range.
Decision table generator. The rule specifications, which are information that describes the software specifications by one or more rules, including the input conditions and the results, which are the information indicating the values of the output variables, and the rule specifications.
Generation range identification information, which is information that specifies the generation range of the decision table
Information storage unit that stores
Of the rule specifications, a decision table generation unit that generates a decision table for the range specified by the generation range specific information, and
With
The components of the input condition are a term that is at least one of a comparison expression, a true / false variable, and a false value, or a negation of the term.
The generation range specifying information includes a metric value acquired for the rule specification that is a generation target of a decision table, and metric evaluation information that is information for evaluating the metric value and specifying the generation range.
The decision table generation unit specifies the generation range based on the metric value and the metric evaluation information, and generates a decision table for the specified generation range.
The rule specifications are generated based on the source code.
The metric value is the number of Else instructions in the "If-Then-Else" type instruction included in the instruction of the source code corresponding to the component of the input condition of the rule.
The metric evaluation information includes information that specifies the rule in which the number of Else instructions is equal to or greater than a preset threshold value as the generation range.
Decision table generator. The decision table generator according to claim 1.
The generation range specific information includes a filter condition described by a logical expression.
The decision table generator generates a decision table for the generation range specified by applying the filter condition to the rule specification.
Decision table generator. The decision table generator according to claim 1.
The information storage unit stores the source code and stores the source code.
It further includes a user interface that accepts the output variable, and a rule specification generator that automatically generates the rule specification based on the source code and the output variable.
Decision table generator. The decision table generator according to claim 1.
The information storage unit stores the decision table and
A decision table inverse conversion unit that automatically generates the rule specifications based on the decision table is further provided.
Decision table generator. Information processing device
The rule specifications, which are information that describes the software specifications by one or more rules, including the input conditions and the results, which are the information indicating the values of the output variables, and the rule specifications.
Generation range identification information, which is information that specifies the generation range of the decision table
Steps to remember,
A step of generating a decision table for a range specified by the generation range specifying information in the rule specifications.
And
The components of the input condition are a term that is at least one of a comparison expression, a true / false variable, and a false value, or a negation of the term.
The information processing device
A step of converting a logical expression generated based on the input condition or the rule specification into a disjunctive normal form which is a logical expression in which clauses which are logical products of the constituent elements are logically equivalent and ORed.
A step of determining whether or not each component affects the input condition, excluding the component that does not affect the input condition from the input condition, and generating the decision table.
How to generate a decision table.

Images