JP5113023B2 - Work plan creation device - Google Patents

Description

  The present invention relates to a work plan creation device.

A system that automatically creates a work plan that satisfies various constraints required for each work content (such as the qualifications of a worker to comply with laws and regulations) has already been developed (for example, Patent Document 1).
JP 2002-149929 A

  However, the creation of a work plan is often used after being corrected based on original knowledge such as many years of experience of the creator. These findings can be regarded as restrictions not related to laws and regulations in the system, but there is a problem that it is difficult to incorporate such knowledge into the system.

  The present invention has been made in view of such problems, and reflects the knowledge of the person in charge based on the change history of the person in charge for the automatically generated work plan information and the work plan information created in the past. It is an object of the present invention to provide a work plan creation device that can automatically create work schedule information.

  In order to solve the above-described problem, the work plan creation device according to the present invention includes a work plan automatic creation unit that creates work plan information based on predetermined constraints, and the work plan information is changed by a planner's operation. Sometimes, the change history is output from the work plan manual correction unit that outputs the information of the change and the finalized work plan information and the change information output from the work plan manual correction unit. A change history extraction unit for extracting information, an assignment information extraction unit for extracting past assignment information from the finally confirmed work plan information output from the work plan manual correction unit, a change history information and a past The assignment trend matrix generator that generates an assignment trend matrix that shows the assignment trend by the planner from the assignment information of the planner, and the work schedule information that was created by the work plan automatic creation part based on the assignment trend matrix Ean has a automatic correction unit.

  In such a work plan creation apparatus according to the present invention, the work plan information includes at least two or more items, and the change history information and the past allocation information are any of the two or more items included in the work plan information. It is preferable that these two items be managed in a matrix having rows and columns. Then, the change history extraction unit adds 1 to the matrix component determined by the items corresponding to the rows and columns included in the change history, and the allocation information extraction unit adds the finally determined work plan information to the work plan information. Preferably, 1 is added to the components of the matrix determined by the items corresponding to the included rows and columns.

  At this time, it is preferable that the change history information and the past allocation information include two or more matrices in which at least one of the row and column items is different from the other.

  Further, at this time, the allocation trend matrix generation unit is configured to perform scaling for aligning the sizes of the rows and columns and scaling for aligning the sizes of the components when the items of the rows and columns of two or more matrices are different. It is preferable.

  Further, in such a work plan creation device according to the present invention, the work plan information includes at least work content, a worker, and the qualification that the worker has, and each of the change history information and past allocation information includes Qualification-work content table with qualification as row and work content as column, worker-work content table with worker as row and work content as column, and work attribute in which work content is classified as worker as row It is preferable that it is comprised from the worker-work attribute table which uses as a column.

によりパラメータをα、β、γ、δ、ε、ξを算出し、さらに、直前に作成された勤務計画情報から資格−作業内容テーブル、作業員−作業内容テーブル、及び、作業員−作業属性テーブルを生成して、当該テーブルの各々に対する最適化問題の最適解の平均値δ′、ε′、ξ′を算出し、割当傾向行列Ｄを次式

但し、Ｄ_ijは割当傾向行列の成分
により算出するように構成されることが好ましい。 And this time, allocation tends matrix generation unit, qualified in a past assignment information - each component of the work table a _ij, dispersed with V _a, the worker - the component b _ij of work tables, the variance V _b , each component of the worker-work attribute table is c _ij , the variance is V _c, and each component of the qualification-work content table in the change history information is d _ij , and the variance is V _d , When each component of the work content table is e _ij , the variance is V _e , each component of the worker-work attribute table is f _ij , and the variance is V _f ,

[Alpha], [beta], [gamma], [delta], [epsilon], [xi] are calculated from the above, and the qualification-work content table, worker-work content table, and worker-work attribute table are calculated from the work plan information created immediately before. Are calculated, and average values δ ′, ε ′, ξ ′ of the optimal solution of the optimization problem for each of the tables are calculated.

However, D _ij is preferably configured to be calculated from the components of the allocation tendency matrix.

  If the work plan creation device according to the present invention is configured as described above, the history of corrections made by the planner using the work plan manual correction unit for the work plan information created by the work plan automatic creation unit And since the assignment tendency of this planner can be extracted as an assignment tendency matrix by the assignment tendency matrix generation unit from the work plan information finally determined, the knowledge of this person in charge can be accumulated as data, In addition, the work plan information corrected by the work plan draft auto-correction unit reflects the knowledge of the person in charge that has been accumulated. Can be planned.

  Hereinafter, preferred embodiments of the present invention will be described with reference to the drawings. First, the configuration of the work plan creation apparatus according to the present invention will be described with reference to FIG. The work plan creation device 1 includes a work plan automatic creation unit 11 that creates work plan information 30 based on input information, and an assignment trend matrix storage unit 25 that includes the work plan information 30 created by the work plan automatic creation unit 11. The work plan automatic correction unit 12 that corrects the corrected work plan information in the corrected work plan plan storage unit 21 and corrects the work plan information to the corrected work plan information. For example, the planner reads out from the storage unit 21 and displays it on a display unit such as a display, and the planner provides a function of correcting using the input unit such as a mouse or a keyboard. Change history information is extracted from the work plan manual correction unit 13 stored in the final work plan storage unit 22 as work plan information and the correction information corrected by the work plan manual correction unit 13. Then, the final work plan information corrected by the change history extraction unit 14 and the work plan manual correction unit 13 stored in the assignment change history information storage unit 23 is read from the final work plan plan storage unit 22 and past allocations are made. Information is extracted and stored in the past allocation information storage unit 24, and the change history information stored in the allocation change history information storage unit 23 and the past history information stored in the past allocation information storage unit 24 are stored. The allocation trend matrix generation unit 16 calculates the planner's allocation trend from the allocation information as an allocation trend matrix and stores it in the allocation trend matrix storage unit 25. Note that these functions can be implemented in a computer having a hard disk or the like as a storage unit.

  Then, the process performed with this work plan preparation apparatus 1 is demonstrated. First, the work plan automatic creation unit 11 fulfills various restrictions based on input information such as workers, work contents, number of persons and qualifications required for work, and vacation requests, which are input from the input means by the planner. Information 30 is created. Further, the work plan information 30 is corrected by the work plan automatic correction unit 12 and stored in the corrected work plan plan storage unit 21. Here, as shown in FIG. 2, the work plan information 30 has at least records including information on the work content 30a, the work date and time 30b, the worker 30c, and the qualification 30d possessed by the worker as many as the number of work contents. Configured. The processing of the work plan automatic correction unit 12 will be described later.

  Then, as described above, the planner corrects the work plan by the work plan manual correction unit 13 using the work plan information corrected in this way. At this time, the corrected work plan information is transferred to the change history extraction unit 14 and the assignment change history information is extracted. For example, as shown in FIG. 3A, in the work plan, the worker of the work plan information 31 (“Koda” in this example) and the worker of the work plan information 32 (“Otohara” in this example) And are replaced. As shown in FIG. 3B, the allocation change history information storage unit 23 is configured to store the change history information as a table for managing workers as rows and managing work contents as columns ( This table is referred to as “worker-work content table 23b” in the following description). Therefore, when the change history extraction unit 14 receives the information on the work plan table corrected from the work plan manual correction unit 13, the change history extraction unit 14 sets 1 to the component specified by the worker and the work content in the worker-work content table 23b. Perform the addition process. When the correction of FIG. 3A is performed, 1 is added to the component “1” specified by the worker “Koda” and the work content “periodic inspection” to become “2”, 1 is added to the component “3” specified by the worker “Otohara” and the work content “civil engineering” to become “4”. In FIG. 3A, “◯” indicates a value of a component not related to the above description, and an arbitrary value is set (the same applies to the following description).

  In this allocation change history storage unit 23, as shown in FIG. 4, in addition to the worker-work content table 23b described above, the change history information is managed as qualifications as rows and qualifications for managing work content as columns. -A work content table 23a and a worker-work attribute table 23c that manages workers as rows and manages work attributes as columns are managed. Therefore, every time the correction history extraction unit 14 receives the correction information of the above work plan information, in addition to the worker-work content change history, the qualification-work content change history and the worker-work attribute change history. Are extracted and stored in the corresponding tables 23a to 23c. Here, the work attribute is information for grouping work contents, and is a superordinate concept of the work contents. For example, the work content “electric work” is classified into a work attribute “work”, and the work content “periodic inspection” is classified into a work attribute “inspection”. The change history extraction unit 14 can specify the work attribute from the work content by storing the correspondence between the work content and the work attribute in a storage unit or the like in advance. Further, the qualification can be specified by information included in the work plan information 30. When the change history extraction unit 14 extracts the qualification-work content change history from the work correction information, if the change history extraction unit 14 always corresponds to the qualification necessary for the work content, the change history extraction unit 14 described above. It is configured not to perform addition processing. This is for excluding necessary qualification information in order to extract the relationship between the qualification and the work content.

  As described above, the qualification-work content table 23a, the worker-work content table 23b, and the worker-work attribute table 23c can be expressed by a matrix composed of rows and columns. d), also referred to as matrix (e) and matrix (f).

  On the other hand, when the work plan information finally determined is stored in the final work plan plan storage unit 22 by the work plan manual correction unit 13, the allocation information extraction unit 15 uses the finally determined work plan information from the work plan information. Similar to the above-described change history extraction unit 14, all work plan information included in the finally determined work plan information is extracted, and qualification-work content, worker-work content for each work plan information. And the number of worker-work attributes is counted and added to the table of the past allocation information storage unit 24. As shown in FIG. 5, the past assignment information storage unit 24 stores a qualification-work content table 24a, a worker-work content table 24b, and a worker-work attribute table 24c. The extraction method and the addition method are the same as those of the change history extraction unit 14 described above. These tables are called a matrix (a), a matrix (b), and a matrix (c), respectively.

  After the finally determined work plan information is extracted by the allocation information extraction unit 15 and stored in the past allocation information storage unit 24, the allocation trend matrix generation unit 16 generates an allocation trend matrix. Is called. The process of this allocation tendency matrix generation part 16 is demonstrated using FIG. The allocation trend matrix generation unit 16 first reads the above-described matrices (a) to (f) from the allocation change history information storage unit 23 and the past allocation information storage unit 24 (step S100).

  By the way, these matrices (a) to (f) have different sizes (rows × columns) because the numbers of items (qualifications, work contents, workers) constituting the rows and columns are different. For this reason, a size scaling process for aligning the size of the matrix is performed (step S110). Specifically, the qualification-work content tables 23a and 24a and the worker-work attribute tables 23c and 24c are converted into the form of the worker-work content tables 23b and 24b, respectively, and temporarily stored in the storage unit or the like. To do. That is, in the case of the qualification-work content tables 23a and 24a, the “qualification” that is the item in the row is replaced with “worker”, and the component set in the qualification is changed to the work having the qualification for each work content. For each worker, a component determined by the worker and the work content is set. For example, as shown in FIG. 7A, “Koda” and “Yodogawa” have the qualification of “electric engineer”, and “5” is set for the component determined by the electrician and the periodic inspection. If so, “5” is set for each of the components determined by Koda and Yodogawa in the converted table and periodic inspection. Similarly, in the case of the worker-work attribute tables 23c and 24c, “work attribute” which is an item in the column is replaced with “work content”, and the component set in the work attribute is changed for each worker. For each work content classified as a work attribute, a component determined by the work content and the worker is set. For example, as shown in FIG. 7B, “Civil work” and “Electrical work” are classified into the work attribute “Construction”, and “4” is set as the component determined by Yodogawa and construction. If it is, “4” is set for each of the components determined by the Yodo River, civil engineering work, and electrical work in the converted table. By performing the above processing, all of the matrices (a) to (f) are stored in the form of a “worker × work content” matrix.

In addition, when the above matrix is added, if the value of one matrix component is large and the value of the other matrix component is small, adding these matrices causes the matrix with a small component value to affect the whole. Will be lower. Therefore, component scaling processing is performed on the matrix whose size has been scaled in step S110 (step S120). Specifically, the sum of all the components included in each of the matrices (a) to (f) is A to F, and the largest value among the sums A to F obtained in this way is M. (M = max {A, B, C, D, E, F}). Then, using these values, the matrices (a) to (f) are subjected to scaling processing according to the following equation (1) and temporarily stored in a storage unit or the like. Here, a _{ij to} f _ij represent respective components of the matrices (a) to (f).

When the size and component scaling processing is completed as described above, parameters α to ξ corresponding to the scaled matrices (a) to (f) are determined. Here, the parameter α corresponds to the matrix (a), the parameter β corresponds to the matrix (b), the parameter γ corresponds to the matrix (c), the parameter δ corresponds to the matrix (d), and the matrix (e ) Corresponds to parameter ε, and parameter ξ corresponds to matrix (f). Specifically, first, the respective variances V _{a to} V _f in the respective matrices (a) to (f) are obtained, and the reciprocal number V of these is obtained by the following equation (2).

  The parameters α to ξ are determined by the following equation (3) based on the reciprocal V and the variances Va to Vf of the matrices (a) to (f).

  The method of determining the parameters α to ξ is based on the idea that the influence on the assignment of the person in charge is strong in the matrix that the variance regarding the components in the matrix is large.

  Next, with respect to the matrices (d) to (f) stored in the allocation change history information storage unit 23 extracted as the work history change history, the optimum solution when the optimization problem is applied to each of these matrices is determined. Parameters δ ′, ε ′, and ξ ′ that are average values are calculated (step S140). Here, how to determine the parameter δ ′ will be described. First, the number of appearances of the qualification-work content is extracted from the work plan information generated immediately before, similarly to the allocation information extraction unit 15 described above, and a qualification-work content matrix (X) is generated. The matrix (X) is subjected to processing similar to that shown in step S110 to perform size scaling processing to form a worker × work content matrix. Then, an optimization problem expressed by the following equation (4) is considered for the matrix X thus scaled.

（４）

(4)

The optimal solution δ _ij ″ of the problem shown in this equation (4) is expressed by the following equation (5).

Then, the average value of the optimum solution δ _ij ″ is calculated as δ ′. Note that the parameter ε ′ also generates a worker-work content matrix (Y) from the work schedule information generated immediately before. The optimal solution ε _ij ″ is calculated from the matrix (Y) by the above-described process, and is obtained from this average value. The parameter ξ ′ is also obtained from the work plan information generated immediately before the worker-work attribute matrix (Z). Then, the optimal solution ξ _ij ″ is calculated from the matrix (Z) by the above-described process and obtained from this average value. Before solving the optimization problem shown in the above equation (4), the matrix (X ), (Y), (Z), it is preferable to perform the component scaling processing shown in step S120.

Finally, the allocation tendency matrix (D) is determined using the matrices (a) to (f) and the parameters α to ξ and δ ′ to ξ ′ determined in steps S130 and S140 described above (step S150). It stores in the allocation tendency matrix storage unit 25 (step S160). The assignment tendency matrix (D) is a matrix of worker × work contents, and D _ij indicates each component of the assignment tendency matrix (D). Here, since the matrixes (d) to (f) are related to information in which the creator who creates the work plan has changed the assignment, the matrices (d) to (f) are adjusted using special parameters δ ′ to ξ ′.

  The above-mentioned work plan automatic correction unit 12 corrects the work plan information 30 created by the work plan automatic creation unit 11 using the assignment tendency matrix (D) determined in this way. Specifically, the total value of the components is calculated for each column of the assignment trend matrix (D), that is, for each work content, and the assignment is corrected in the order of the work content having the largest total value. This large sum means that the work is more likely to be assigned.

  For example, when the values shown in FIG. 8A are set in each component of the assignment tendency matrix (D) in which five work contents are assigned to four workers, the components for each work contents Therefore, the allocation work is performed in the order of work 2, work 1, work 4, work 3, and work 5. In addition, in the assignment of workers to each work content, correction is performed so as to assign to a worker having a large numerical value (component) of the worker for the work content. For example, as shown in FIG. 8B, the number of persons required for the work 2 is 2, and the value of each worker component for the work 2 in the assignment tendency matrix (D) is (6, 2, 3, 7 ), The ideal assignment is preferably assigned to the worker 4 and the worker 1 in descending order of the component values (FIG. 8B shows a case where work is assigned to the worker. “1” is set, and “0” is set when no assignment is made). If this ideal assignment is different from the work plan information 30 created by the work plan automatic creation unit 11 (in the case of FIG. 8B, worker 1 and worker 2 The work plan automatic correction unit 12 performs assignment change processing according to the contents of the assignment tendency matrix (D). As an allocation correction method, for example, as shown in FIG. 8C, in the allocation tendency matrix (D), allocation is performed in descending order of worker component values in the corresponding work contents, and various constraint conditions (laws) This is done by allocating the required number of people until it meets the restrictions for complying with the rules).

  If the work plan creation device 1 is configured as described above, the history of the revision made by the planner using the work plan manual manual modification unit 13 with respect to the work plan information created by the work plan automatic creation unit 11 and the final The assignment tendency of the planner can be extracted as an assignment tendency matrix by the assignment tendency matrix generation unit 16 from the work plan information that has been determined. Therefore, the knowledge of the person in charge can be accumulated as data, and the work plan information corrected by the work plan automatic correction unit 12 can be corrected based on the allocation trend matrix based on this knowledge. The knowledge is reflected, and the work plan can be made efficiently with less manual correction work by the planner.

It is a block diagram which shows the structure of the work plan preparation apparatus which concerns on this invention. It is a data structure figure which shows the data structure of work schedule information. It is explanatory drawing for demonstrating the extraction method of the change history when work plan information is corrected, Comprising: (a) shows the content of the changed work plan information, (b) is management of the extracted information The method is shown. It is explanatory drawing which shows the data structure of the table managed by the allocation change log | history information storage part. It is explanatory drawing which shows the data structure of the table managed by the past allocation information storage part. It is a flowchart which shows the flow of a process in an allocation tendency matrix production | generation part. It is explanatory drawing which shows the scaling method of the size of a matrix, Comprising: (a) shows the example of scaling of a qualification-work content table, (b) shows the example of scaling of a worker-work attribute table. It is explanatory drawing which shows the correction method of the work plan information in a work plan automatic correction part, Comprising: (a) shows the method of determining the priority of the work content to correct from an allocation tendency matrix, (b) The method for assigning workers is shown, and (c) shows the flow of judgment when assigning workers.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Work plan creation apparatus 11 Work plan automatic creation part 12 Work plan automatic correction part 13 Work plan draft manual correction part 14 Change history extraction part 15 Allocation information extraction part 16 Allocation trend matrix generation part

Claims (6)

A work plan automatic creation unit that creates work plan information based on predetermined constraints,
When the work plan information is changed by the planner's operation, the work plan manual correction unit that outputs the information of the change and outputs the finally confirmed work plan information;
A change history extraction unit that extracts history information of the change from the change information output from the work plan manual correction unit;
An allocation information extraction unit that extracts past allocation information from the finally determined work plan information output from the work plan draft manual correction unit;
An allocation trend matrix generating unit that generates an allocation trend matrix indicating an allocation trend by the planner from the history information of the change and the past allocation information;
A work plan creation apparatus comprising: a work plan plan automatic correction unit that corrects the work plan information created by the work plan automatic creation unit based on the allocation trend matrix. The work plan information includes at least two or more items,
The change history information and the past allocation information are managed in a matrix having rows and columns of any two items of the two or more items included in the work plan information,
The change history extraction unit adds 1 to a component of the matrix determined by an item corresponding to the row and column included in the change history,
The allocation information extraction unit is configured to add 1 to a component of the matrix determined by an item corresponding to the row and column included in the finally determined work plan information. The work plan creation device described.   The work history creation apparatus according to claim 2, wherein the change history information and the past allocation information include two or more matrices in which at least one of the row and column items configured is different from the other.   The allocation tendency matrix generation unit is configured to perform scaling for aligning the sizes of the rows and columns and scaling for aligning the sizes of components when the items of the rows and columns of the two or more matrices are different. The work plan creation device according to claim 3. The work plan information includes at least work content, a worker, and qualifications the worker has,
Each of the change history information and the past allocation information includes a qualification-work content table having the qualification as a row and the work content as a column, and a worker-work having the worker as a row and the work content as a column. The work plan creation device according to claim 3 or 4, comprising a content table and a worker-work attribute table having a column of work attributes in which the worker is classified as a row and the work content is classified. The allocation trend matrix generation unit includes:
In the past allocation information, each component of the qualification-work content table is a _ij , the variance is V _a , each component of the worker-work content table is b _ij , and the variance is V _b, and the worker-work Each component of the attribute table is c _ij and the variance is V _c .
In the change history information, each component of the qualification-work content table is d _ij , the variance is V _d , each component of the worker-work content table is e _ij , and the variance is V _e, and the worker-work When each component of the attribute table is f _ij and the variance is V _f ,

To calculate the parameters α, β, γ, δ, ε, ξ,
Further, the qualification-work content table, the worker-work content table, and the worker-work attribute table are generated from the work plan information created immediately before the optimization problem for each of the tables. The average values δ ′, ε ′, and ξ ′ of the optimal solution are calculated, and the allocation tendency matrix D is

However, the work plan creation device according to claim 5, wherein D _ij is configured to be calculated from a component of an allocation tendency matrix.

Images