JP6221177B2 - Rule verification device, rule verification method, and program - Google Patents

Description

  The present invention relates to a rule verification device, a rule verification method, and a program.

When a large amount of data is classified or searched by a combination of numerical ranges of conditions in each attribute of the data, classification or search rules may be defined by a plurality of IF-THEN rules.
Patent Document 1 describes a technique for adjusting a rule specified as an adjustment target so as to be a rule evaluation standard specified by a user based on data to be analyzed as a related technique.

JP 2001-312408 A

By the way, in a plurality of IF to THEN format rules including conditions and results, the classification or search condition indicated by each IF is indicated by a numerical range, and is associated with the attribute indicated by THEN by a combination of the numerical ranges. In this case, each time new data is added, a numerical value range and an attribute indicated by the IF-THEN rule are added. For example, an accounting department of a company may create and manage a list showing expenses to be borne by each organization for each product using spreadsheet software or the like. In the list showing the expenses that each organization bears for each product, the expenses that each organization bears are data and rules in the IF to THEN format over time from the start of product development to the completion of mass production of the product. Sometimes added. In such a case, for the same product, the result indicated by THEN even though the conditions indicated by IF are the same between the added IF-THEN rule and the previously added IF-THEN rule. There is a possibility that a contradictory relationship in which the two are different and a redundant relationship in which the IF to THEN rules are the same or a part of the rules overlap.
For this reason, when rules in the IF-THEN format are added sequentially, there is a need for a technique for identifying a rule that may have a contradiction or redundancy relationship between a plurality of rules in the IF-THEN format. .

  Accordingly, an object of the present invention is to provide a rule verification device, a rule verification method, and a program that can solve the above-described problems.

In order to achieve the above object, according to one aspect of the present invention, a rule verification device includes a convex hull generation unit, a duplication determination unit, a duplication rule specification unit, and a rule division unit. The convex hull generating unit generates a convex hull based on each rule of a plurality of rules in which a combination of record value ranges is associated with an attribute of the record. The overlap determination unit determines whether there is an overlap between the plurality of convex hulls generated by the convex hull generation unit. The duplication rule specifying unit specifies a combination of rules indicated by a convex hull that the duplication judgment unit judges to have duplication. The rule division unit excludes any one of the ranges overlapping with other rules from the combination of the value ranges of the records constituting one rule of the rule combinations specified by the duplication rule specifying unit. Thus, a new division rule for eliminating the duplication is generated.

  According to another aspect of the present invention, the rule verification device includes a relationship determination unit. The relationship determination unit determines whether the two rules are inconsistent or redundant based on attributes indicated by two rules of the combination of rules indicated by the convex hull specified by the duplication rule specifying unit. judge.

  Moreover, according to another aspect of the present invention, the relation determination unit provided in the rule verification device has different attributes indicated by two rules of a combination of rules indicated by the convex hull specified by the duplication rule specifying unit , It is determined that the two rules are in a contradictory relationship.

According to another aspect of the present invention, the rule division portion in which the rule verification device comprises divides a rule the convex hull is shown. When the relation determination unit determines that two rules out of a combination of rules indicated by the convex hull specified by the duplication rule specification unit are in a contradiction, the rule division unit uses the two rules as a user instruction. Split based on.

  According to another aspect of the present invention, the relationship determination unit included in the rule verification device has the same attribute indicated by two rules in the combination of rules indicated by the convex hull specified by the duplication rule specifying unit. If there is, it is determined that the two rules are in a redundant relationship.

  According to another aspect of the present invention, the rule verification device includes an inclusive convex hull calculation unit. The inclusion convex hull calculation unit corresponds to the two rules when the relation determination unit determines that two rules of the combination of rules indicated by the convex hull specified by the duplication rule specification unit are in a redundant relationship. An inclusive convex hull including two convex hulls is calculated.

  According to another aspect of the present invention, the rule verification device includes a rule presentation unit. The rule presentation unit displays a rule indicated by the convex hull or the inclusion convex hull on a display unit. The rule presenting unit determines whether or not there is an overlap between the inclusion convex hull calculated by the inclusion convex hull calculation unit and another convex hull, and the duplication determination unit determines whether the overlapping determination hull is the inclusion convex hull. And the rule indicated by the convex hull when it is determined that there is no overlap between the convex hull and the rule indicated by the convex hull divided by the rule dividing unit are displayed on the display unit.

According to another aspect of the present invention, rule verification apparatus comprises a hull含凸follicle calculator, and a rule presentation unit. Before SL inclusion convex hull calculation unit calculates the inclusion convex hull encompasses convex hull generated based on rules according to the same attribute of the record. The rule presentation unit displays a rule indicated by the convex hull or the inclusion convex hull on a display unit.

According to another aspect of the present invention, the rule verification method generates a convex hull based on each rule of a plurality of rules in which a combination of record value ranges and attributes of the record are associated with each other. It is determined whether or not there is overlap between hulls, the combination of rules indicated by the convex hull that is determined to be overlapping is specified, and the value of the record that constitutes one rule of the specified rule combinations By excluding any one of the ranges overlapping with other rules from the combination of ranges, a new rule for division that eliminates the overlap is generated.

According to another aspect of the present invention, a program generates a convex hull based on each rule of a plurality of rules that associates a combination of a record value range and a record attribute. A convex hull generation unit, a duplication determination unit for determining whether there is an overlap between the plurality of convex hulls generated by the convex hull generation unit, and a convex hull for which the duplication determination unit determines that there is an overlap. Duplicate rule specifying means for specifying a combination of rules to be displayed , any of ranges that overlap with other rules from a combination of value ranges of the records constituting one rule among the combinations of rules specified by the duplicate rule specifying means By excluding one of these, it is made to function as a rule dividing means for generating a new rule for dividing to eliminate the duplication.

  According to the rule verification device according to the embodiment of the present invention, when the IF-THEN format rule is added sequentially, the added IF-THEN format rule and the previously added IF-THEN format rule It is possible to identify rules that may have contradictory relationships or redundant relationships between them.

It is a figure which shows an example of a structure of the rule verification apparatus 10 by 1st embodiment of this invention. It is a figure which shows an example of the IF-THEN rule which the memory | storage part 110 by 1st embodiment of this invention has memorize | stored. An example of the IF-THEN rule used when specifying that the management source code (responsibility section) indicating the organization responsible for product management (here, section) in the first embodiment of the present invention is the identifier RN0 is shown. FIG. It is a figure which shows an example of the processing flow of the rule verification apparatus 10 by 1st embodiment of this invention. It is a figure which shows an example of the convex hull which the convex hull production | generation part 101 produces | generates by the process of step S2. It is a figure which shows the example of the judgment performed when the relationship determination part 104 calculates the inclusion convex hull which includes the part in a redundant relationship. It is a figure which shows the example of the judgment performed when the relationship determination part 104 calculates the inclusion convex hull which includes the part in a redundant relationship. It is a figure which shows the example of the judgment performed when the relationship determination part 104 calculates the inclusion convex hull which includes the part in a redundant relationship. FIG. 9 is a diagram illustrating an example of the contradiction in the IF to THEN format rules. It is a figure which shows an example of determination of whether there is a contradiction which the rule verification apparatus 10 performs when there are two conditions which IF shows and the management source codes (responsibility section) of two rules differ. It is a figure which shows an example of the IF-THEN rule which the memory | storage part 110 by 2nd embodiment of this invention has memorize | stored. It is a figure which shows an example of IF-THEN rule used when specifying that the management origin code (responsibility section) which shows the organization of the responsibility of management of a product is identifier RN0 in 2nd embodiment of this invention.

<First embodiment>
FIG. 1 is a diagram showing an example of the configuration of the rule verification device 10 according to the first embodiment of the present invention.
As shown in FIG. 1, the rule verification device 10 according to the first embodiment includes a convex hull generation unit 101, an overlap determination unit 102, an overlap rule specification unit 103, a relationship determination unit 104, and a rule division unit 105. , An inclusive convex hull calculation unit 106, a rule presentation unit 107, a storage unit 110, and a display unit 120.

  The convex hull generation unit 101 included in the rule verification device 10 generates a convex hull based on each rule of a plurality of rules in which combinations of record value ranges stored in the storage unit 110 are associated with attributes of the record. To do. The combination of the record value ranges is a condition indicated by IF of a rule (hereinafter referred to as IF-THEN rule) indicated in IF-THEN format. The record attribute is a result indicated by THEN of the IF to THEN rule.

The overlap determination unit 102 determines whether there is an overlap between the plurality of convex hulls generated by the convex hull generation unit 101.
The duplication rule specifying unit 103 specifies a combination of IF to THEN rules indicated by the convex hulls that the duplication determination unit 102 has determined to have duplication.

  The relationship determination unit 104 determines whether the two IF to THEN rules are inconsistent based on the attributes indicated by the two IF to THEN rules among the combinations of IF to THEN rules indicated by the convex hull specified by the duplication rule specifying unit 103. And the redundant relationship is determined. When the attributes indicated by the two IF-THEN rules in the combination of IF-THEN rules indicated by the convex hull specified by the duplication rule specifying unit 103 are different, the relationship determination unit 104 determines that the two IF-THEN rules are inconsistent. It is determined that Further, when the attributes indicated by the two IF to THEN rules in the combination of IF to THEN rules indicated by the convex hull specified by the duplication rule specifying unit 103 are the same, the relationship determination unit 104 determines that the two IF to THEN It is determined that the rule is in a redundant relationship.

When the relationship determination unit 104 determines that two IF to THEN rules of the combination of IF to THEN rules indicated by the convex hull specified by the duplication rule specifying unit 103 are in a contradictory relationship, the rule dividing unit 105 One IF-THEN rule is divided based on a user instruction.
The inclusive convex hull calculating unit 106, when the relationship determining unit 104 determines that two IF to THEN rules of the combination of IF to THEN indicated by the convex hull specified by the duplication rule specifying unit 103 are in a redundant relationship, An inclusive convex hull including two convex hulls corresponding to the two IF to THEN rules is calculated.

The rule presenting unit 107 displays information to be notified to the user, such as displaying a division candidate for canceling the contradictory relationship of the IF to THEN rules and a candidate for canceling the redundant relationship.
The storage unit 110 stores various information necessary for processing of the rule verification device 10 such as a plurality of IF to THEN rules in which combinations of record value ranges are associated with attributes of the record.
The display unit 120 indicates the IF to THEN rule indicated by the inclusion convex hull when the overlap determination unit 102 determines that there is no overlap between the inclusion convex hull and other convex hulls, or the convex hull divided by the rule division unit 105 Various information notified to the user, such as IF to THEN rules, is displayed.

FIG. 2 is a diagram illustrating an example of IF to THEN rules stored in the storage unit 110 according to the first embodiment of the present invention.
FIG. 3 is a diagram illustrating an example of IF to be used when the management source code (responsibility section) indicating the product management responsibility organization (section in this case) is the identifier RN0 in the first embodiment of the present invention. It is a figure which shows an example of a THEN rule.

FIG. 2 shows a plurality of IF to THEN rules that associate rule #, serial #, product #, item code, cost element, and management source code (management section). For example, rule # 1 is a product specified by an identifier 172 of serial # and an identifier 1 of product #. The item code “pn to pn + 1” and the cost element “an to an + 1” are the numerical range of the condition indicated by the IF in the IF to THEN rule expressed in the format of “minimum value to maximum value”. For example, in rule # 1, the minimum value of the item code is pn, and the maximum value is pn + 1. In rule # 1, the minimum value of the cost element is an, and the maximum value is an + 1.
Further, in the rule # 1, for the product specified by the serial number identifier 172 and the product # identifier 1, the condition indicated by the IF is that the item code is between pn and pn + 1, and the cost element is an In the case of an + 1, the organization (responsibility section) responsible for managing the product indicated by THEN is RU1. Serial # is recorded from # 1.

Further, the example of the IF to THEN rule shown in FIG. 3 shows the rule for rule # 2 in FIG. This FIG. 3 shows the result indicated by THEN when the value of the item code record is between cn and cn + 1 and the value of the cost element record is between dn and dn + 1 as the condition indicated by the IF. It means that the value of the record of the management source code (responsibility section) is RN0.
Note that the rule verification device 10 according to the first embodiment of the present invention detects rule duplication and contradiction in a plurality of such IF to THEN rules, and simplifies the rules. And processing to correct the contradiction of rules.

FIG. 4 is a diagram illustrating an example of a processing flow of the rule verification device 10 according to the first embodiment of the present invention.
Next, the processing flow of the rule verification device 10 according to the first embodiment will be described with reference to FIGS.
The processing performed by the rule verification device 10 according to the first embodiment shown in FIG. 4 includes a combination of record value ranges (conditions indicated by IF of the IF to THEN rule) and an attribute of the record (THEN of the IF to THEN rule). A convex hull is generated based on each rule of the plurality of rules associated with each other, and it is determined whether or not there is an overlap between the plurality of convex hulls. It includes a process for specifying a rule combination, that is, a rule that may have a contradiction or redundancy.

First, the convex hull generation unit 101 included in the rule verification device 10 reads out a plurality of rules, that is, a plurality of IF to THEN rules, associated with a combination of record value ranges and attributes of the record from the storage unit 110 (step S100). S1). For example, the plurality of IF to THEN rules that the convex hull generation unit 101 reads from the storage unit 110 include rule #, serial #, product #, item code, cost element, management, and the like as shown in FIG. This is a rule that associates the original code (management section).
When the convex hull generating unit 101 reads a plurality of IF to THEN rules from the storage unit 110, the minimum value of each condition (in this case, the item code and the cost element) indicated by the IF of each of the plurality of IF to THEN rules to be read A convex hull is obtained based on the maximum value (range) (step S2).

FIG. 5 is a diagram illustrating an example of the convex hull generated by the convex hull generating unit 101 in the process of step S2.
As shown in FIG. 5, the convex hull generation unit 101 has two points, the minimum value and the maximum value of the item code in each IF to THEN rule read from the storage unit 110 as shown in FIG. 2, and each IF to THEN. A convex hull is obtained for a point set of four coordinates on a two-dimensional plane indicated by a combination of the minimum value and the maximum value of the cost element in the rule. For example, the convex hull generation unit 101 may obtain a convex hull from a point set using a known algorithm such as a divide-and-conquer method. Then, the convex hull generation unit 101 outputs the information indicated by the obtained convex hull to the duplication determination unit 102.

When the duplication determination unit 102 receives the information indicated by the convex hull generated by the convex hull generation unit 101, all of the plurality of convex hulls including the four coordinates on the input two-dimensional plane are selected. It is determined whether or not there is an overlap in the convex hull for the combination (step S3). For example, the overlap determination unit 102 may determine whether or not there is an overlap in the convex hull using a known algorithm such as a GJK (Gilbert-Johnson-Keerthi) algorithm that determines the overlap of the convex hull polyhedron. If the overlap determining unit 102 determines that there is no overlap in the convex hulls for all combinations that select two convex hulls out of a plurality of convex hulls including four coordinates on the two-dimensional plane, the overlapping hull overlaps It is determined that there is no (step S3, NO).
In addition, when the overlap determining unit 102 determines that there is an overlap in the convex hull for all combinations that select two convex hulls out of a plurality of convex hulls including four coordinates on the two-dimensional plane, Are determined to be duplicated (step S3, YES). If the overlap determination unit 102 determines that there is an overlap in the convex hull, the overlap determination unit 102 outputs information indicated by the convex hull determined to have overlap to the overlap rule specifying unit 103.

  When the duplication rule specifying unit 103 inputs information indicated by the convex hull that is determined to have duplication, a plurality of IFs associating the combination of record value ranges and the attributes of the record as shown in FIG. Read the ~ THEN rule from the storage unit 110. For example, the duplication rule specifying unit 103 compares the range of each condition in the plurality of IF to THEN rules read from the storage unit 110 with the range of each condition in the rule corresponding to the convex hull determined to have duplication. . Then, the duplication rule specifying unit 103 matches the range of each condition in the plurality of IF to THEN rules read from the storage unit 110 with the range of each condition in the rule corresponding to the convex hull determined to have duplication. The rule combination is identified as the rule combination indicated by the convex hull that is determined to have an overlap (step S4). The duplication rule specifying unit 103 outputs a combination of rules indicated by the convex hull determined to have duplication to the relationship judgment unit 104.

When the relationship determination unit 104 inputs a combination of rules indicated by the convex hull determined to have overlap from the overlap rule specifying unit 103, the relationship determination unit 104 determines whether the attributes of the two rules are the same (step S5). .
When the relationship determination unit 104 determines that the attributes of the two rules indicated by the convex hulls that are determined to have overlap are the same (YES in step S5), the relationship determination unit 104 determines that the two rules are in a redundant relationship ( Step S6). Then, the relationship determination unit 104 outputs the two rules determined to be in a redundant relationship to the inclusion convex hull calculation unit 106 (step S7).

6 to 8 are diagrams illustrating examples of determinations performed when the relationship determination unit 104 calculates an inclusive convex hull including a portion having a redundant relationship.
First, with reference to FIG. 6, an inclusion relationship that is determined by the relationship determination unit 104 when a plurality of rules have an inclusion relationship will be described.
An example of the inclusion relationship determined by the relationship determination unit 104 illustrated in FIG. 6 is the inclusion relationship for the IF to THEN rule # 1 and the rule # 3 illustrated in FIG.
Assume that the relationship determination unit 104 reads out rule # 1 and rule # 3 from the storage unit 110, as indicated by (1) in FIG. When the range of cost elements in rule # 1 is an to an + 1, the range of cost elements in rule # 3 is bn to bn + 1, and the relationship bn <an <an + 1 <bn + 1 holds, rule # 3 is rule # 1 Is included. In this case, the relationship determination unit 104 generates bn to bn + 1 as new cost element rules.
However, as indicated by (2) in FIG. 6, for example, a rule (in this case, rule #) in which the relationship of bn <an <an + 1 <sn <sn + 1 <bn + 1 holds and becomes the management source code (in this case, the identifier USZ). When at least one of 5) exists in bn to bn + 1, the relationship determination unit 104 does not generate a new cost element rule.

Next, with reference to FIG. 7, the division relationship determined by the relationship determination unit 104 when a plurality of rules are in a division relationship will be described.
An example of the division relationship determined by the relationship determination unit 104 illustrated in FIG. 7 is the division relationship for the IF to THEN rule # 1, rule # 3, and rule # 6 illustrated in FIG.
Assume that the relationship determining unit 104 reads out rule # 1, rule # 3, and rule # 6 from the storage unit 110, as indicated by (1) in FIG. The range of cost elements in rule # 1 is an to an + 1, the range of cost elements in rule # 3 is bn to bn + 1, the range of cost elements in rule # 6 is cn to cn + 1, and an <an + 1 = bn When the relationship <bn + 1 <cn <cn + 1 is satisfied, rule # 1, rule # 3, and rule # 6 are divided. In this case, the relationship determination unit 104 generates an to cn + 1 as a new cost element rule using the maximum convex hull.
However, as indicated by (2) in FIG. 7, for example, the relationship of an <an + 1 = bn <bn + 1 <sn <sn + 1 <cn <cn + 1 holds, and the rule that becomes the management source code (in this case, the identifier USZ) is bn + 1. If it is present in ~ cn, the relationship determination unit 104 generates an ~ bn + 1 as a new cost element rule.

Next, the overlapping relationship determined by the relationship determining unit 104 when a plurality of rules are in an overlapping relationship will be described with reference to FIG.
An example of the overlapping relationship determined by the relationship determining unit 104 illustrated in FIG. 8 is the overlapping relationship for the rules # 1 and # 3 in the IF to THEN format illustrated in FIG.
As shown in FIG. 8, it is assumed that the relationship determination unit 104 reads out rule # 1 and rule # 3 from the storage unit 110. When the range of cost elements in rule # 1 is an to an + 1, the range of cost elements in rule # 3 is bn to bn + 1, and the relationship of an <bn <an + 1 <bn + 1 holds, rule # 1 and rule # 3 Are overlapping. In this case, the relationship determination unit 104 generates an to bn + 1 as a new cost element rule.

Note that the convex hull generated by the relationship determination unit 104 described here uses, for example, the minimum and maximum IF conditions for each rule of a plurality of rules as a point set, and is an algorithm such as Divide and Conquer Method (Divide and Command Method). ) Or the like.
This is the description of the example in which the convex hull is obtained based on the minimum value to the maximum value (range) of the IF condition for each rule of the plurality of rules read by the relationship determination unit 104 in step S7. By the processing in step S7, the redundant relationship can be eliminated by degenerating a plurality of rules indicating the redundant relationship such as the inclusion relationship, the division relationship, and the overlapping relationship.
Note that the process of canceling the redundant relationship by the relationship determining unit 104 degenerating the plurality of rules in step S7 is performed for all combinations of the plurality of rules in each condition.

  In addition, when the relationship determination unit 104 determines that the attributes of the two rules indicated by the convex hulls that are determined to have overlap are not the same (NO in step S5), the relationship determination unit 104 determines that the two rules are in a contradictory relationship. (Step S8). Then, the relationship determining unit 104 outputs the two rules determined to be in a contradictory relationship to the rule dividing unit 105 (step S9).

  When the rule determination unit 105 inputs two rules determined to be inconsistent by the relationship determination unit 104, the rule division unit 105 includes four coordinates on the two-dimensional plane indicated by the convex hull generated for each of the two input rules. How to divide the overlapping parts, that is, the parts in contradiction, and generate candidates for eliminating the contradiction (step S10).

FIG. 9 is a diagram illustrating an example of the contradiction in the IF to THEN format rules.
An example of the contradiction relationship described with reference to FIG. 9 is the conflict relationship between rule # 1 and rule # 6 in the IF to THEN format shown in FIG.
As shown in FIG. 9, it is assumed that the convex hull generation unit 101 reads out rule # 1 and rule # 6 from the storage unit 110. The range of cost elements in rule # 1 is an to an + 1, the range of cost elements in rule # 6 is sn to sn + 1, and the relationship of an <sn <an + 1 <sn + 1 holds, and the management source code (in this case, identifier When a rule indicating USZ) exists in an to an + 1, rule # 1 and rule # 6 are contradictory. However, if there is one condition indicating IF (in this case, a cost element), it is determined whether the range of sn to an + 1 is included in the new rule # 1 or the new rule # 5, and a new cost element rule An to a′n + 1 are generated as # 1, and a′n + 1 to bn + 1 are generated as rule # 5 of a new cost element. However, if the IF indicates one condition, there is no logical problem.

FIG. 10 is a diagram illustrating an example of determining whether there is a contradiction performed by the rule verification device 10 when there are two conditions indicated by the IF and the management source codes (responsibility section) of the two rules are different.
The two conditions indicated by the IF in this case are a cost element and an item code.
When there is no overlap in any of the two conditions indicated by the IF, the rule verification device 10 determines that there is no complete contradiction. When only one of the two conditions indicated by the IF is duplicated, the rule verification device 10 determines that there is no problem in theory. When there is an overlap in both of the two conditions indicated by the IF, the rule verification device 10 determines that there is a contradiction. For example, a candidate for division that eliminates the contradiction generated by the rule division unit 105 is division that excludes an overlapping range from one of two rules. The rule division unit 105 outputs the generated division candidates for eliminating the generated contradiction to the rule instruction unit 107 (step S11).
When the rule dividing unit 105 inputs a division candidate for eliminating the contradiction, the rule presenting unit 107 outputs a signal for displaying the division candidate for generating the contradiction generated by the rule dividing unit 105 to the display unit 120 ( Step S12).
When the display unit 120 receives a signal for displaying the candidate for division for eliminating the contradiction from the rule presenting unit 107, the display unit 120 displays the candidate for division for eliminating the contradiction based on the input signal (step S13). Note that the display unit 120 displays a part in a contradictory relationship as a rectangle that is easily recognized by the user, for example, painted in red.
Here, the user performs a user operation for designating division based on the division candidates for eliminating the contradiction. Then, the rule dividing unit 105 divides the rule based on the user operation (step S14).
Next, the control unit of the rule verification device 10 determines whether or not to end the process based on the user operation (step S15). When the user performs a user operation to end the process (step S15, YES), the control unit of the rule verification apparatus 10 ends the process of the rule verification apparatus 10. Moreover, the control part of the rule verification apparatus 10 returns to the process of step S3, when a user performs user operation which does not complete | finish a process (step S15, NO).

In addition, the two rules that the relationship determining unit 104 determines to have a redundant relationship in step S <b> 7 are output to the inclusive convex hull calculating unit 106. Then, the inclusion convex hull calculation unit 106 inputs two rules that the relationship determination unit 104 determines to have a redundant relationship. The inclusive convex hull calculating unit 106 includes an overlapping convex part including a redundant part of a range including four coordinates on a two-dimensional plane indicating the convex hull generated for each of the two input rules. A hull is calculated (step S16), and a candidate for eliminating the redundant relationship is generated. For example, the candidate for eliminating the redundant relationship generated by the inclusive convex hull calculating unit 106 is an inclusive convex hull obtained using a known algorithm including two convex hulls corresponding to two rules. The inclusion convex hull calculation unit 106 outputs candidates for canceling the generated redundancy relationship to the rule instruction unit 107 (step S17).
When the rule division unit 105 outputs a division candidate for eliminating the contradiction relationship to the rule instruction unit 107 in step S11, the rule presentation unit 107 displays a signal for displaying the division candidate for eliminating the contradiction generated by the rule division unit 105. Is output to the display unit 120 (step S18).
In addition, when the inclusion convex hull calculation unit 106 outputs a candidate for canceling the redundancy relationship to the rule instruction unit 107 in step S13, the rule presentation unit 107 displays the candidate for canceling the redundancy relationship generated by the inclusion convex hull calculation unit 106. The signal to be output is output to the display unit 120 (step S19).
Moreover, if the display part 120 inputs the signal which displays the candidate which cancels a redundant relationship from the rule presentation part 107, it will display the candidate which cancels a redundant relation based on the input signal (step S20).
Next, the control unit of the rule verification device 10 determines whether or not to end the process based on the user operation (step S21). When the user performs a user operation to end the process (step S21, YES), the control unit of the rule verification apparatus 10 ends the process of the rule verification apparatus 10. Moreover, the control part of the rule verification apparatus 10 returns to the process of step S3, when a user performs user operation which does not complete | finish a process (step S21, NO).

The rule verification device 10 according to the first embodiment of the present invention has been described above.
The rule verification device 10 according to the embodiment of the present invention includes a convex hull generation unit 101, an overlap determination unit 102, and an overlap rule specification unit 103. The convex hull generating unit 101 generates a convex hull based on each rule of a plurality of rules in which combinations of record value ranges are associated with attributes of the record. The overlap determination unit 102 determines whether there is an overlap between the plurality of convex hulls generated by the convex hull generation unit 101. The duplication rule specifying unit 103 specifies a combination of rules indicated by the convex hulls that the duplication determination unit 102 has determined to have duplication.
In this way, when IF-THEN format rules are added sequentially, there is a contradiction or redundancy relationship between the added IF-THEN format rules and the previously added IF-THEN format rules. It is possible to identify rules that may have occurred.
Further, the rule verification device 10 further includes a relationship determination unit 104, a rule division unit 105, an inclusive convex hull calculation unit 106, a rule presentation unit 107, and a display unit 120, so that the contradiction relationship specified by the duplication rule specification unit 103 It is possible to visually notify a user of a rule that may have a redundant relationship and correct it to a rule having no contradiction or redundant relationship.

<Second Embodiment>
Next, the rule verification device 10 according to the second embodiment of the present invention will be described.
The configuration of the management system 1 according to the second embodiment is the same as that of the management system 1 according to the first embodiment. However, the IF condition of the IF to THEN rule that the convex hull generation unit 101 according to the second embodiment reads from the storage unit 110 includes a request source and a recording source in addition to the item code and the cost element in the first embodiment. ing.

FIG. 11 is a diagram illustrating an example of IF to THEN rules stored in the storage unit 110 according to the second embodiment of the present invention.
FIG. 12 shows an example of IF to THEN rules used when the management source code (responsibility section) indicating the organization responsible for product management is specified as the identifier RN0 in the second embodiment of the present invention. FIG.

FIG. 11 shows a plurality of IF to THEN rules that associate rule #, serial #, product #, request source, accounting source, item code, cost element, and management source code (management section). Is shown. For example, rule # 1 is a product specified by an identifier 172 of serial # and an identifier 1 of product #. The request source “zn to zn + 1”, the recording source “dn to dn + 1”, the item code “pn to pn + 1”, and the cost element “an to an + 1” are IF to THEN expressed in the format of “minimum value to maximum value”. It is a numerical range of the condition indicated by IF in the rule. For example, in rule # 1, the minimum value of the request source is zn, and the maximum value is zn + 1. The minimum value of the recording source is dn, and the maximum value is dn + 1. The minimum value of the item code is pn, and the maximum value is pn + 1. In rule # 1, the minimum value of the cost element is an, and the maximum value is an + 1.
Further, in the rule # 1, for the product specified by the serial # identifier 172 and the product # identifier 1, the condition indicated by the IF is that the request source is between zn and zn + 1, and the accounting source is dn If it is between dn + 1, the item code is between pn and pn + 1, and the cost element is an to an + 1, the organization (responsibility section) of product management responsibility indicated by THEN is RU1. Serial # is recorded from # 1.

  Further, the example of the IF to THEN rule shown in FIG. 12 shows the rule for rule # 2 in FIG. In FIG. 12, as the condition indicated by the IF, the value of the record of the request source is between zn and zn + 1, the value of the record of the record source is between dn and dn + 1, and the item code When the value of the record is between cn and cn + 1 and the value of the record of the cost element is between dn and dn + 1, the value of the record of the management source code (responsibility section) is RN0 as a result indicated by THEN Means that

Next, processing of the rule verification device 10 according to the second embodiment will be described.
However, the difference between the processing performed by the rule verification device 10 according to the second embodiment and the processing performed by the rule verification device 10 according to the first embodiment is the calculation performed by the rule verification device 10 according to the difference in the number of conditions indicated by IF. And the display unit 110 displays differently depending on the order of the calculation.
Accordingly, only the calculation performed by the duplication determination unit 102 of the rule verification device 10 and the display displayed by the display unit 110, which are the main differences, will be described here.

  The calculation performed by the inclusive convex hull calculation unit 106 of the rule verification device 10 is a four-dimensional calculation of a request source, a recording source, an item code, and a cost element, which are the four conditions indicated by the IF. However, the inclusive convex hull calculating unit 106 according to the second embodiment may use, for example, a GJK (Gilbert-Johnson-Keerthy) algorithm, similarly to the inclusive convex hull calculating unit 106 according to the first embodiment. In the case of n dimensions, a two-dimensional plane algorithm may be calculated n (n−1) ÷ 2 times. Therefore, in the case of four dimensions, the two-dimensional algorithm is calculated six times.

Next, display in which the display unit 110 displays four conditions will be described.
The inconsistent relationship is determined when the inclusion convex hull calculating unit 106 determines that all the ranges indicated by the four conditions are overlapped. However, the display unit 110 can only perform three-dimensional display. Therefore, the display unit 110 displays, for example, all combinations that select three of the four ranges. Also, priorities are set in advance in the conditions, and the display unit 110 displays the three ranges in three dimensions based on the priorities. Further, the user may arbitrarily select it.

The rule verification device 10 according to the first embodiment of the present invention has been described above.
The rule verification device 10 according to the embodiment of the present invention includes a convex hull generation unit 101, an overlap determination unit 102, and an overlap rule specification unit 103. The convex hull generating unit 101 generates a convex hull based on each rule of a plurality of rules in which combinations of record value ranges are associated with attributes of the record. The overlap determination unit 102 determines whether there is an overlap between the plurality of convex hulls generated by the convex hull generation unit 101. The duplication rule specifying unit 103 specifies a combination of rules indicated by the convex hulls that the duplication determination unit 102 has determined to have duplication.
In this way, when IF-THEN format rules are added sequentially, there is a contradiction or redundancy relationship between the added IF-THEN format rules and the previously added IF-THEN format rules. It is possible to identify rules that may have occurred.
Further, the rule verification device 10 further includes a relationship determination unit 104, a rule division unit 105, an inclusive convex hull calculation unit 106, a rule presentation unit 107, and a display unit 120, so that the contradiction relationship specified by the duplication rule specification unit 103 It is possible to visually notify a user of a rule that may have a redundant relationship and correct it to a rule having no contradiction or redundant relationship.

Note that the processing of the rule verification device 10 according to the embodiment first obtains a convex hull for each rule, obtains an overlapping area, and then identifies a rule that may have a contradiction or redundancy relationship. explained. However, the processing of the rule verification device 10 according to the embodiment is not limited to this.
For example, the processing of the rule verification device 10 may be optimized by obtaining a contradiction or redundancy relationship for each condition, and obtain a convex hull for the optimized combination of conditions.
In the processing flow according to the embodiment of the present invention, the order of processing may be changed within a range in which appropriate processing is performed.
When there is no contradiction between IF to THEN rules, a convex hull generating unit 101 that generates a convex hull based on each rule of a plurality of rules that associates a combination of record value ranges and attributes of the record; An inclusion convex hull calculation unit 106 that calculates an inclusion convex hull that includes a convex hull generated based on a rule related to the same attribute of the record, and a rule that displays the convex hull or the rule indicated by the inclusion convex hull on the display unit 120 The redundancy relationship may be specified by the rule verification device 10 including the presentation unit 107. Moreover, the rule verification apparatus 10 may correct the specified redundant relationship.

  In addition, although embodiment of this invention was described, the above-mentioned rule verification apparatus 10 has a computer system inside. The process described above is stored in a computer-readable recording medium in the form of a program, and the above process is performed by the computer reading and executing this program. Here, the computer-readable recording medium means a magnetic disk, a magneto-optical disk, a CD-ROM, a DVD-ROM, a semiconductor memory, or the like. Alternatively, the computer program may be distributed to the computer via a communication line, and the computer that has received the distribution may execute the program.

  The program may be for realizing a part of the functions described above. Furthermore, what can implement | achieve the function mentioned above in combination with the program already recorded on the computer system, and what is called a difference file (difference program) may be sufficient.

  Although several embodiments of the present invention have been described, these embodiments are presented by way of example and are not intended to limit the scope of the invention. Various omissions, replacements, and changes can be made without departing from the scope of the invention.

DESCRIPTION OF SYMBOLS 10 ... Rule verification apparatus 101 ... Convex hull production | generation part 102 ... Duplication determination part 103 ... Duplication rule specific | specification part 104 ... Relation determination part 105 ... Rule division part 106 ... Inclusion convex Package calculation unit 107 ... rule presentation unit 110 ... storage unit 120 ... display unit

Claims (10)

A convex hull generating unit that generates a convex hull based on each rule of a plurality of rules associating a combination of record value ranges and attributes of the record;
An overlap determination unit that determines whether there is an overlap between the plurality of convex hulls generated by the convex hull generation unit;
A duplication rule identifying unit that identifies a combination of rules indicated by the convex hull that is determined to be duplicated by the duplication judgment unit;
By excluding any one of the ranges overlapping with other rules from the combination of the value ranges of the records constituting one rule of the combination of rules specified by the duplicate rule specifying unit, A rule verification device comprising: a rule division unit that generates a new rule for division to be resolved. A relationship determination unit for determining whether the two rules are inconsistent or redundant based on attributes indicated by two rules of the rule combination indicated by the convex hull specified by the duplication rule specifying unit; The rule verification device according to claim 1 provided. The relationship determination unit determines that the two rules are in a contradiction when two of the rules indicated by the convex hull specified by the duplication rule specification unit have different attributes. The rule verification device described. When the relation determination unit determines that two rules out of a combination of rules indicated by the convex hull specified by the duplication rule specifying unit are in a contradictory relationship, the rule dividing unit uses the two rules as a user instruction. The rule verification device according to claim 2, wherein the rule verification device is divided based on the rule. The relationship determination unit determines that the two rules are in a redundant relationship when the attributes indicated by two rules of the combination of rules indicated by the convex hull specified by the duplication rule specification unit are the same. The rule verification device according to any one of claims 2 to 4. When the relationship determination unit determines that two rules of the rule combination indicated by the convex hull specified by the duplication rule specifying unit are in a redundant relationship, the two convex hulls corresponding to the two rules are included. The rule verification device according to any one of claims 2 to 5, further comprising an inclusion convex hull calculating unit that calculates an inclusion convex hull. A rule presenting unit for displaying a rule indicated by the convex hull or the containing convex hull on a display unit;
The rule presenting unit determines whether or not there is an overlap between the inclusion convex hull calculated by the inclusion convex hull calculation unit and another convex hull, and the duplication determination unit determines whether the overlapping determination hull is the inclusion convex hull. The rule which the said inclusion convex hull when it determines with there being no duplication with other convex hulls, and the rule which the said convex hull which the said rule division | segmentation part divided | segmented are displayed on a display part. Verification device. An inclusive convex hull calculating unit for calculating an inclusive convex hull including a convex hull generated based on a rule relating to the same attribute of the record;
The rule verification device according to claim 1, further comprising: a rule presenting unit that displays the convex hull or the rule indicated by the containing convex hull on a display unit. A convex hull is generated based on each rule of a plurality of rules in which a combination of record value ranges and attributes of the record are associated,
Determine if there is overlap between multiple convex hulls,
Identify the combination of rules indicated by the convex hull that is determined to be duplicated,
By excluding any one of the ranges overlapping the other rules from the combination of the value ranges of the records constituting one rule of the identified rule combinations , a new division to eliminate the duplication Generate rules Rule validation method. The computer of the rule verification device
A convex hull generating means for generating a convex hull based on each rule of a plurality of rules associating a combination of record value ranges and attributes of the record;
A duplication determination unit that determines whether there is an overlap between the plurality of convex hulls generated by the convex hull generation unit;
A duplication rule specifying means for specifying a combination of rules indicated by the convex hull determined by the duplication judgment means to be duplicated;
By excluding any one of the ranges overlapping with other rules from the combination of the value ranges of the records constituting one rule of the combination of rules specified by the duplicate rule specifying means, A program that functions as a rule division unit that generates new rules for division to be resolved.

Images