Allocation plan creation system, method, and program

Abstract

To improve validity of an allocation plan.SOLUTION: An allocation plan creation system 100 for creating an allocation plan in which candidate allocation targets are allocated to an allocation destination includes a transition likelihood calculation unit 114 and an allocation plan creation unit 115. The transition likelihood calculation unit 114 extracts transition of a plurality of positions for each personnel as transition context on the basis of a plan history in which a position for each personnel in the past is recorded to calculate likelihood from the extracted transition context. The allocation plan creation unit 115 specifies candidate personnel satisfying constraint conditions of attribute of personnel per position, calculates a degree of matching between the attribute of the identified candidate personnel and the position, calculates a degree of conformance of the candidate personnel relative to the position on the basis of the calculated degree of matching and likelihood to create an allocation plan in which the candidate personnel is allocated to the position on the basis of the calculated degree of conformance.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a system, method and program for creating an allocation plan. [Background technology]

[0002] There are many cases where optimal allocation planning is required for combinations of human resources, tasks, resources, or positions. Examples include the allocation of human resources to personnel posts or positions in personnel transfers within an organization, and the allocation of crew members to tasks such as aircraft flight schedules. When formulating such allocation plans, mathematical programming is often used to evaluate the compatibility between the human resource requirements expected for the task and the attributes of each human resource. Task requirements include the skills necessary to perform the task and the number of years of work experience.

[0003] By formulating the skills required to perform these tasks and the years of work experience as constraints, and then quantitatively evaluating the degree to which each person's attributes can meet the requirements of the task, as a match between each person's attributes and the constraints, personnel can be assigned to tasks. Quantitative evaluation makes it possible to identify personnel who are the best match for a given task.

[0004] However, if the task requirements are not sufficiently defined by the constraint formula, the creator of the allocation plan must manually select the final candidate talent in order of the highest match. Therefore, if the number of target talents or tasks is large, the selection decision becomes a burden for the creator of the allocation plan.

[0005] Therefore, Patent Document 1 discloses a GUI function that supports the creator of the allocation plan in selecting candidate personnel. The burden on the creator of the allocation plan is caused by the insufficient extraction of personnel requirements defined as constraint condition expressions, as described above.

[0006] In response to this, Patent Document 2 also discloses a technology that extracts the tacit knowledge of the creator of an unformulated allocation plan from past planning history and expresses the extracted tacit knowledge as a constraint condition formula. [Prior art documents] [Patent documents]

[0007] [Patent Document 1] Patent No. 6729922 [Patent Document 2] International Publication No. 2018 / 220885 Summary of the Invention [Problem to be solved by the invention]

[0008] Patent Document 1 discloses a technology that supports the creation of an allocation plan that combines human resources belonging to an organization with personnel posts or positions within the organization. The technology in Patent Document 1 supports the selection of candidate employees by creating an allocation plan that uses constraints to assign multiple candidate employees to each post.

[0009] In personnel transfers within an organization, placement is sometimes determined not only based on the suitability of the personnel for the post at a given time, but also on a career path that allows the personnel to hold more than two posts, either continuously or intermittently, from the perspective of mid- to long-term human resource development within the organization. Such placement decisions have traditionally been made using the tacit knowledge and past experience of personnel in the human resources department, and therefore cannot be made simply based on the suitability of the personnel for the post.

[0010] In Patent Document 1, the selection of candidate employees is supported by outputting combinations of selectable positions and candidate employees. However, since candidate employees are not extracted based on tacit knowledge that cannot be expressed in constraints, including the above-mentioned career paths, manual selection is required from multiple candidate employees.

[0011] Patent Document 2 discloses a technology that handles tacit knowledge using mathematical programming. The technology in Patent Document 2 extracts a planning pattern for a production sequence, which is a transition of tasks in a product manufacturing process that has been previously planned, for a product manufacturing plan. This planning pattern is a pattern of a back-and-forth production sequence. In the back-and-forth relationship, it is not possible to consider paths that include two or more sequences. Therefore, it is not possible to reflect tacit knowledge, such as the above-mentioned career path that involves holding more than two positions, in the planning.

[0012] The present invention has been made in view of the above-mentioned problems, and an object of the present invention is to provide a technique for improving the validity of an allocation plan. [Means for solving the problem]

[0013] In order to solve the above-mentioned problems, the present invention provides an allocation plan creation system that creates an allocation plan in which candidate allocation targets are assigned to allocation destinations, and includes: a transition likelihood calculation unit that extracts transitions of multiple allocation targets for each allocation target as transition contexts based on a plan history that records the allocation destinations for each past allocation target, and calculates likelihoods from the extracted transition contexts; and an allocation plan creation unit that identifies candidate allocation targets that satisfy constraints on attributes of the allocation targets for each allocation destination, calculates a degree of match between the attributes of the identified candidate allocation targets and the allocation destinations, calculates a suitability of the candidate allocation targets for the allocation destinations based on the calculated degree of match and the likelihood, and creates an allocation plan in which the candidate allocation targets are assigned to the allocation destinations based on the calculated degree of suitability. [Effects of the Invention]

[0014] According to the present invention, the validity of the allocation plan can be improved. [Brief explanation of the drawings]

[0015] [Figure 1] 1 is a schematic diagram of an allocation plan creation system according to a first embodiment. [Figure 2] 3 is a schematic diagram of a transition likelihood calculation unit and an allocation plan creation unit according to the first embodiment. [Figure 3] FIG. 3 is a diagram showing an example of the data structure of a plan history DB according to the first embodiment. [Figure 4] FIG. 3 is a diagram showing an example of the data structure of an allocation plan execution parameter DB according to the first embodiment. [Figure 5] FIG. 2 is a diagram showing an example of the data structure of a human resource attribute DB according to the first embodiment. [Figure 6] FIG. 2 is a diagram illustrating an example of a data structure of a position requirement DB according to the first embodiment. [Figure 7] FIG. 4 is a diagram illustrating an example of a data structure of a transition pattern DB according to the first embodiment. [Figure 8] FIG. 4 is a diagram illustrating an example of a data structure of a transition context DB according to the first embodiment. [Figure 9] FIG. 3 is a diagram showing an example of the data structure of candidate position information according to the first embodiment. [Figure 10] FIG. 4 is a diagram illustrating an example of a data structure of attribute match degree information according to the first embodiment. [Figure 11] FIG. 4 is a diagram illustrating an example of a data structure of maximum likelihood context information according to the first embodiment. [Figure 12] 1 is a flowchart showing the processing of the allocation plan creation system according to the first embodiment. [Figure 13] 1 is a flowchart showing the processing of the allocation plan creation system according to the first embodiment. [Figure 14] 10 is a flowchart showing an allocation plan creation process according to the first embodiment. [Figure 15] FIG. 10 is a diagram showing a mover determination screen according to the first embodiment. [Figure 16] FIG. 2 is a diagram illustrating an example of the hardware configuration of an information processing apparatus according to the first embodiment. [Figure 17] FIG. 10 is a schematic diagram of an allocation plan creation system according to a second embodiment. [Figure 18] FIG. 10 is a diagram showing an example of a data structure of a plan history according to the second embodiment. [Figure 19] FIG. 10 is a diagram showing an example of the data structure of a human resource attribute DB according to the second embodiment. [Figure 20] FIG. 10 is a diagram illustrating an example of a data structure of a task requirement DB according to the second embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0016] Hereinafter, specific examples of an allocation plan creation system according to an embodiment of the present invention will be described with reference to the drawings. Note that the present invention is not limited to the examples, but is defined by the claims. [Example]

[0017] FIG. 1 is a schematic diagram of an allocation plan creation system according to a first embodiment.

[0018] In this embodiment, an allocation plan creation system 100 is illustrated that creates an allocation plan (also called a personnel transfer plan) in which human resources, as an example of an "allocation target" in relation to a personnel transfer in an organization, are assigned to a personnel post or a personnel position, as an example of an "allocation destination." This embodiment relates to the creation of an allocation plan for tasks, positions, or resources that can be assigned to human resources, and personnel transfer is one application example of the invention.

[0019] The allocation plan creation system 100 is operated by a user 101. The allocation plan creation system 100 includes an operation terminal 111, a plan creation execution management unit 112, a table data operation terminal 113, a transition likelihood calculation unit 114, and an allocation plan creation unit 115. Furthermore, the allocation plan creation system 100 includes a plan history database (hereinafter, referred to as DB) 121, an allocation plan execution parameter DB 122, a human resource attribute DB 123, and a position requirement DB 124.

[0020] A user 101 is a user who instructs the assignment plan creation system 100 to create a personnel transfer plan and uses the created personnel transfer plan. The user 101 inputs information into a personnel attribute DB 123 that manages personnel attribute information necessary for creating a personnel transfer plan using mathematical programming, and a position requirement DB 124 that manages the requirements for personnel to be assigned to a position.

[0021] The operation terminal 111 draws the information input by the user 101 and the created personnel transfer plan. The operation terminal 111 accepts input of information necessary for creating a personnel transfer plan from the user 102. The operation terminal 111 outputs the accepted data to the plan creation execution management unit 112.

[0022] The plan creation and execution management unit 112 stores the parameters of the information input from the operation terminal 111 in the allocation plan execution parameter DB 122 , and then outputs the parameters to the transition likelihood calculation unit 114 and the allocation plan creation unit 115 .

[0023] The table data operation terminal 113 accepts input and modification of information stored in a personnel attribute DB 123 and a position requirement DB 124, which are necessary for executing a mathematical program, from the user 101. The personnel attribute DB 123 and the position requirement DB 124 may be created via the table data operation terminal 113, or may receive automatic input from a personnel management system provided by the organization.

[0024] The transition likelihood calculation unit 114 receives input of parameters from the plan creation execution management unit 112, data in the plan history DB 121, and data on candidate talent as an example of a "candidate allocation target" for a position created by the allocation plan creation unit 115. The transition likelihood calculation unit 114 creates a transition context based on these input parameters and data, and calculates the likelihood of position allocation taking into account the transition context for each candidate talent assigned to a position.

[0025] The allocation plan creation unit 115 receives parameter input from the plan creation execution management unit 112, data in the talent attribute DB 123, and data in the position requirement DB 124, and creates an allocation plan for positions for talent using mathematical programming. Furthermore, the allocation plan creation unit 115 receives transition likelihood as an example of "likelihood" for each combination of talent and position that takes into account the transition context, which is calculated by outputting data related to the created allocation plan to the transition likelihood calculation unit 114, and creates a final allocation plan.

[0026] The plan history DB 121 is a table that holds information on allocation plans created by the allocation plan creation unit 115 and personnel transfers that have been manually created in the past. An example of the table structure of the plan history DB 121 will be described with reference to FIG.

[0027] The allocation plan execution parameter DB122 is a table that manages parameters related to the creation of an allocation plan. The parameters related to the creation of an allocation plan include a list of personnel to be transferred and a list of vacant positions. Various setting values ​​and constraints used in mathematical programming may be input into the allocation plan execution parameter DB122. An example of the table structure of the allocation plan execution parameter DB122 is explained in FIG. 4.

[0028] The personnel attribute DB 123 is a table for managing the attribute information of each personnel belonging to an organization. An example of the table structure of the allocation plan execution parameter DB 123 will be described with reference to FIG.

[0029] The position requirement DB 124 is a table that manages the human resource requirements for each position that makes up an organization. The information managed in the table is used as constraints when creating an allocation plan using mathematical programming. An example of the table structure of the position requirement DB 124 is explained in FIG. 6.

[0030] FIG. 2 is a schematic diagram of a transition likelihood calculation unit and an allocation plan creation unit according to the first embodiment.

[0031] The transition likelihood calculation unit 114 includes a transition pattern extraction unit 201 , a transition information search unit 202 , a transition context extraction unit 203 , a maximum likelihood context calculation unit 204 , a transition pattern DB 221 , and a transition context DB 222 .

[0032] The transition pattern extraction unit 201 receives as input the plan history managed in the plan history DB 121, extracts transition patterns, and stores the extracted transition patterns in the transition pattern DB 221. The transition pattern extraction process by the transition pattern extraction unit 201 may be executed each time new data is registered in the plan history DB 121. A transition pattern is the continuous position changes of a given employee over a certain period of time. For example, if employee A transitions through the positions of general employee in the human resources department, assistant manager, department manager, and division manager over several years, the above-mentioned position transitions are the transition pattern of employee A.

[0033] The transition information search unit 202 receives input of information related to an allocation plan for candidates for any position, which is created using mathematical programming from the candidate position information 233. Based on the input information related to the allocation plan, the transition information search unit 202 searches the transition pattern DB 221 for a transition pattern required to create a transition context, and outputs the search result to the transition context extraction unit 203.

[0034] The transition context extraction unit 203 receives input of the list information of transition patterns obtained as a result of the search by the transition information search unit 202 and information related to the position and personnel allocation plan. The transition context extraction unit 203 extracts transition contexts using the transition patterns, stores the extracted transition contexts in the transition context DB 222, and also outputs them to the maximum likelihood context calculation unit 204. A transition context is a weighted directed graph made up of multiple transition patterns. An example of a transition context is described with reference to FIG. 8.

[0035] The maximum likelihood context calculation unit 204 identifies the longest path consisting of N-gram transitions from the transition context extracted by the transition context extraction unit 203 and calculates it as the maximum likelihood context. N-gram is a method of regarding changes between N connected positions as one context. The maximum likelihood context is the largest value among the total weight values ​​between positions included in individual N-gram paths. The maximum likelihood context can also be considered as the longest path when weighting is considered as the distance between nodes.

[0036] The transition pattern DB 221 is a table that manages the transition patterns extracted by the transition pattern extraction unit 201. An example of the table structure of the transition pattern DB 221 will be described with reference to FIG.

[0037] The transition context DB 222 is a table that manages the transition context extracted by the transition context extraction unit 203. An example of the table structure of the transition context DB will be described with reference to FIG.

[0038] The allocation plan creation unit 115 has a mathematical programming loop processing management unit 211, a position-talent attribute match degree calculation unit 212, a position-talent overall compatibility degree calculation unit 213, and a calculation result primary management unit 214. Furthermore, the allocation plan creation unit 115 has a target position primary DB 231, a target talent primary DB 232, candidate talent position information 233, attribute match degree information 234, and maximum likelihood context information 235.

[0039] The mathematical programming loop processing management unit 211 manages the loop execution of the process of allocating candidate talent to each vacant position in order for the list of vacant positions managed by the target position primary DB 231. The mathematical programming loop processing management unit 211 outputs information about the vacant position to be processed to the position-talent attribute match degree calculation unit 212.

[0040] The position-talent attribute match degree calculation unit 212 calculates attribute match degree information 234 as an example of the "match degree" of each candidate for a vacant position using mathematical programming, and ranks the candidates according to the magnitude of the attribute match degree information 234. The position-talent attribute match degree calculation unit 212 receives information on one vacant position to be assigned from the mathematical programming loop processing management unit 211 and receives a list of candidates from the target talent primary DB 232. Furthermore, the position-talent attribute match degree calculation unit 212 receives information on the requirements required for the talent to be assigned to the vacant position from the position requirement DB 124 and information on the attributes of each candidate from the talent attribute DB 123. Based on this received information, the position-talent attribute match degree calculation unit 212 applies mathematical programming to calculate a ranked list of candidates for the target vacant position. The candidate list may be narrowed down to the top few candidates by applying a threshold to the attribute match degree calculated for each candidate.

[0041] The position-talent overall compatibility calculation unit 213 uses the attribute match degree information 234 calculated by the position-talent attribute match degree calculation unit 212 and the most likely context information 235 to calculate the ranking of candidate talent for vacant positions while taking into account the most likely context, thereby calculating the overall compatibility as an example of "compatibility."

[0042] The calculation result primary management unit 214 manages information including the most likely context information for each candidate talent for each vacant position calculated by the position-talent overall compatibility calculation unit 213. As soon as the loop processing by the mathematical programming loop processing management unit 211 is completed, the calculation result primary management unit 214 combines the managed data into one plan and registers it in the plan history DB 121.

[0043] The target position primary DB 231 is a table having the same structure as part of the table data managed by the allocation plan execution parameter DB 122. Details of the target position primary DB 231 are shown in Figure 4 as target position information 401. The table structure of the target position information 401 will be explained with reference to Figure 4.

[0044] The target talent primary DB 231 is a table having the same structure as part of the table data managed by the allocation plan execution parameter DB 122. More details of the target talent primary DB 231 are shown in target talent information 431 in Fig. 4. The table structure of the target position information 431 will be explained in Fig. 4.

[0045] The candidate talent position information 233 is data indicating a candidate talent for any vacant position calculated by the position-talent attribute match degree calculation unit 212. An example of the data structure of the candidate talent position information 233 is shown in FIG.

[0046] The attribute match degree information 234 is data indicating, for a given position, candidates and the attribute match degree for each candidate calculated by the position-talent attribute match degree calculation unit 212. An example of the data structure of the attribute match degree information 234 is shown in FIG.

[0047] The most likely context information 235 is data including the likelihood of a transition path expected when a transition to a position to which a transfer is expected is realized based on the transition context for each candidate included in the candidate position information 233. An example of the data structure of the most likely context information 235 is shown in FIG.

[0048] FIG. 3 is a diagram illustrating an example of a data structure of the plan history DB according to the first embodiment.

[0049] The plan history DB121 is a table that stores, as item values ​​(column values), a plan ID 301, a position ID 302, a candidate ID 303, a suitability 304, a transition likelihood 305, an overall suitability 306, a final candidate 307, and a draft branch number 308. Each row 311 in the plan history DB121 indicates information about each candidate. A group 312 of rows 311 in the plan history DB121 indicates information generated in one plan creation process.

[0050] The plan ID 301 is an identifier assigned to an allocation plan created in a single plan creation process performed by the user 101. The position ID 302 is an identifier that identifies the position for which a candidate talent is indicated in the allocation plan. The candidate talent ID 303 is an identifier that identifies the candidate talent who is a candidate for allocation to that position. The compatibility 304 is the degree of compatibility between the position and talent attribute calculated by the position-talent attribute match degree calculation unit 212.

[0051] The transition likelihood 305 is a value calculated by the maximum context calculation unit 204. The overall compatibility 306 is a value obtained by multiplying the compatibility 304 and the transition likelihood 305. The final candidate 307 is flag information for identifying the candidate 303 with the highest overall compatibility 306 among the candidates 303 of the position ID 302. The draft branch number 308 is an identifier for identifying the local solution calculation process executed within the position-talent attribute match calculation unit 212. Note that the calculation method for the overall compatibility 306 is just an example, and the calculation may be performed using a process other than multiplication.

[0052] FIG. 4 is a diagram illustrating an example of a data structure of the allocation plan execution parameter DB according to the first embodiment.

[0053] The allocation plan execution parameter DB122 is a table that stores the target personnel and positions in one plan creation process. The allocation plan execution parameter DB122 is a table that stores, as item values ​​(column values), target position information 401 and target personnel information 431. The target position information 401 is list information of vacant positions for which candidate personnel need to be identified, and the target personnel information 431 is list information of personnel who may be assigned to the positions.

[0054] The target position information 401 is a table that stores, as item values ​​(column values), a plan ID 301 and a vacant position 411 that indicates a position to which a candidate talent should be assigned. Each row 421 of the target position information 401 indicates an individual position.

[0055] The target human resource information 431 is a table that stores, as item values ​​(column values), a plan ID 301 and a target human resource 441 that identifies human resources who may be transferred. Each row 451 of the target human resource information 431 indicates an individual target human resource.

[0056] FIG. 5 is a diagram illustrating an example of a table structure of the human resource attribute DB according to the first embodiment.

[0057] The human resource attribute DB 123 may refer to information external to the allocation plan creation system 100, such as general personnel information or career information of individual employees. The human resource attribute DB 123 is a table that stores, as item values ​​(column values), a human resource ID 501 that identifies a human resource and other attribute information related to the human resource. Here, an example of an attribute is given.

[0058] The human resource attribute DB 123 is a table that stores, as item values ​​(column values), the year 502 of the human resource, a current position ID 503 indicating the identifier of the current position, a grade 504 indicating the job title, and skills 505, 506 that contain information about the various skills required of the human resource. The skill information may have a true / false value or a numerical value indicating whether or not the skill is possessed. Each row 511 of the human resource attribute DB 123 indicates individual human resource information.

[0059] The talent attribute DB 123 is mainly used as a reference for the attribute information of each target talent when determining the degree of match between the talent attributes and the position requirements by mathematical programming.

[0060] FIG. 6 is a diagram illustrating an example of a data structure of the position requirement DB according to the first embodiment.

[0061] The position requirement DB 124 may refer to information managed as general personnel information outside the allocation plan creation system 100. The position requirement DB 124 is a table that stores, as item values ​​(column values), a position ID 601 that is an identifier for each position, and explanatory information and requirements for the position.

[0062] The position requirement DB 124 includes, as explanatory information for a position, a department 602 and a position 603. The explanatory information for a position in the position requirement DB 124 is an example, and other information such as job rank or grade may also be managed.

[0063] The position requirement DB 124 includes, as examples of position requirements, a minimum age 604 for assuming a position, a minimum grade 605, and required skills 606 and 607. The skill information may include a true / false value or a numerical value indicating whether or not the skill is possessed. Each row 611 of the position requirement DB 124 indicates information about each post.

[0064] FIG. 7 is a diagram illustrating an example of a data structure of the transition pattern DB according to the first embodiment.

[0065] The transition pattern DB 221 is a table that stores a pattern ID 701, a human resource ID 702, a previous position ID 703, a next position ID 704, and a plan ID 301 as item values ​​(column values).

[0066] The pattern ID 701 is the identifier of the pattern. The talent ID 702 is the identifier of the talent. The original position ID 703 is the source of each talent's past transfer. The destination position ID 704 is the destination of the transfer. The plan ID 301 is the identifier of the plan that includes the transfer plan. Each row 711 of the transition pattern DB 221 indicates one transfer (transition) of a talent.

[0067] The transition pattern 712 in the transition pattern DB 221 indicates any personnel whose transfer history is included in the plan history 121 and is grouped by personnel ID. The lower part of Fig. 7 is a graphical representation of the transition (movement) pattern of any personnel included in one pattern ID 701. The transition pattern may be created or updated every time the plan history DB 121 is updated.

[0068] FIG. 8 is a diagram illustrating an example of a table structure of a transition context DB according to the first embodiment.

[0069] The transition context DB 222 may manage data in multiple tables. In the example of Fig. 8, the transition context DB 222 has a context metadata management table 801 that manages metadata of the transition context, and context information 821 that indicates the context.

[0070] The context metadata management table 801 is a table that stores, as item values ​​(column values), a context ID 811, a reference pattern ID list 812, a starting position ID 813, and a graph ID 814. Each row 815 of the transition context DB 222 represents one piece of metadata. When creating context information for each candidate calculated by mathematical programming, the starting position ID 813 is set to the same value as the current position ID 503 of the candidate.

[0071] The context ID 811 is an identifier for each context. The reference pattern ID 812 is a list of pattern IDs used to create the context. The starting position ID list 813 is an identifier for the position that is the starting point of the context. The graph ID 814 is an identifier for graph data that indicates context information.

[0072] The context information 821 is data representing a weighted directed graph, and includes a graph ID 814 that identifies the graph, a transition destination 831 of the nodes that make up the graph, and a transition source 832. A transition context is created for each candidate for a specific position calculated using mathematical programming. Each row of the context information 821 represents a context example 841.

[0073] Here, an example of a method for creating a transition context will be described. First, the candidate's current position ID 503 is used as a search query to obtain multiple transition patterns in which the previous position ID 703 is included in the transition pattern DB 221. By combining the transition patterns, the example context shown in the example context 841 can be created. The example context 841 shows a weighting value for the transition between each node (corresponding to a position). However, this weighting value is calculated by dividing the total number of transition patterns that start from the same position by the total number of patterns that transition to the destination position. The method for calculating the weighting value is just one example. For example, a transition that occurs closer to the present may be considered to be more important by assigning a larger weighting value to a transition that was planned earlier in the transition pattern.

[0074] Furthermore, by taking into consideration the relationship between the candidate talent and the talent (identified from talent ID 702) whose transition pattern is included in the search results, a high weighting value may be assigned to similar (closely related) transition patterns. This is one method for reflecting tacit knowledge, for example, when there is a supervisor-supervisee relationship between employees and similarities can be seen in promotions or position changes within the organization. This improves the accuracy of transition context extraction. Tacit knowledge is the transition context that accompanies the transition over time or location of a combination of talent and position with certain attributes included in past planning history. This context includes more than two transitions from one position to another.

[0075] FIG. 9 is a diagram illustrating an example of a data structure of candidate position information according to the first embodiment.

[0076] The candidate position information 233 includes a position ID 302, a candidate 303, and a current position 503. Each row 901 of the candidate position information 233 indicates information about each candidate.

[0077] Position ID 302 indicates the position for which the candidate has been calculated by mathematical programming. Candidate 303 indicates the candidate's identifier. Current position 503 indicates the current position of each candidate.

[0078] FIG. 10 is a diagram illustrating an example of a data structure of attribute match degree information according to the first embodiment.

[0079] The attribute matching degree information 234 includes a position ID 302, a candidate 303, a current position 503, a suitability 304, and a draft branch number 308. Each row 1001 of the attribute matching degree information 234 indicates information about each candidate.

[0080] The position ID 302 indicates the position for which the candidate personnel is calculated by mathematical programming. Candidate talent 303 indicates the candidate talent's identifier. Current position 503 indicates the current position of each candidate talent. Matching score 304 indicates the degree of match between the position requirements of the above-mentioned position calculated by mathematical programming and the candidate talent's attributes. Draft branch number 308 is a branch number that identifies the local solution calculation process executed within the position-talent attribute matching score calculation unit 212.

[0081] FIG. 11 is a diagram illustrating an example of a data structure of the maximum likelihood context information according to the first embodiment.

[0082] The maximum likelihood context information 235 is information indicating the maximum likelihood context information of each candidate 303 of the position ID 302 included in the candidate position information 233, and includes a transition likelihood 305 and a context ID 811. Each row 1101 of the maximum likelihood context information 235 indicates the maximum likelihood context information of each candidate. The maximum likelihood context is a connection of the maximum weighted values ​​of each transition in a specified N-gram path in a graph of the context ID 811 created for each candidate.

[0083] The bottom of Figure 11 shows an example of a maximum likelihood context. The transition likelihood 305 described above is the sum of the weighted values ​​of each transition included in the maximum likelihood context. The transition likelihood does not have to be a simple sum. For example, a constant may be applied each time the path moves away from the starting position to increase or decrease the importance of distant paths.

[0084] 12 and 13 are flowcharts illustrating the processing of the allocation plan creation system according to the first embodiment.

[0085] In the plan creation process, subflows A and B are executed as processes for acquiring and managing necessary data before main flow C, which processes allocation plan creation by the allocation plan creation system 100. Subflow A is a process for creating a transition pattern DB 221. Subflow B is a process for creating a human resource attribute DB 123 and a position requirement DB 124.

[0086] The transition pattern extraction unit 201 acquires data from the plan history DB 121 and stores the extracted transition pattern information in the transition pattern DB 221 (S1211).

[0087] Next, the user 101 inputs the personnel attribute DB 123 and the position requirement DB 124 as information required for creating an allocation plan via the table data operation terminal 113 (S1221).

[0088] Next, the table data operating terminal 113 stores the input information in the personnel attribute DB 123 and the position requirements 124 (S1222).

[0089] Next, the user 101 inputs information into the target position primary DB 231 and the target human resource primary DB 232 as data related to planning to the planning execution management unit 112 via the operation terminal 111 (S1311).

[0090] Next, the plan creation execution management unit 112 stores the input information in the target position primary DB 231 and the target personnel primary DB 232, respectively, and then inputs this information to the transition likelihood calculation unit 114 and the allocation plan creation unit 115 (S1312).

[0091] Next, the transition information search unit 202 of the transition likelihood calculation unit 114 extracts one row of the candidate talent 303 from the input candidate talent position information 233, and executes a loop of extracting a transition context and calculating a maximum likelihood context. From this step to S1317, the process is executed in a loop for each row of the candidate talent 303 (S1313).

[0092] Next, the transition information search unit 202 searches the transition pattern DB 221 for a transition pattern that includes the candidate's current position information 503, and inputs the search results to the transition context extraction unit 203 (S1314).

[0093] Next, the transition context extraction unit 203 creates a directed transition graph shown in the transition context example 841 from the multiple transition patterns found in S1314, stores it in the transition context DB 222, and then inputs it to the maximum likelihood context calculation unit 204 (S1315).

[0094] Next, the maximum likelihood context calculation unit 204 acquires, from among the N-gram routes including paths connecting the candidate's current position 503 and the position ID 302 for the input information, the one with the largest sum of weighting values ​​of the paths it passes through as the maximum likelihood context (S1316).

[0095] Next, the transition information search function 202 determines whether or not there is an unprocessed row in the candidate position information 233, and if there is not (S1317: NO), ends the loop processing (S1317).

[0096] Next, the maximum likelihood context calculation unit 204 combines the maximum likelihood contexts corresponding to all the candidates included in the candidate position information 233 and inputs the combined maximum likelihood context information 235 to the position-talent overall compatibility calculation unit 213 (S1318).

[0097] The allocation plan creation unit 115 creates data corresponding to a group 312 of rows 311 in the plan history 121 from the target position primary DB 231, the target human resource primary DB 232, the human resource attribute DB 123, the position requirement DB 124, which are input from the plan creation execution management unit 112, and the input from the maximum likelihood context calculation unit 204. The allocation plan creation unit 115 registers the created data in the plan history DB 121 (S1319).

[0098] Next, the user 101 refers to and uses the created allocation plan via the operation terminal 111 (S1320).

[0099] FIG. 14 is a flowchart illustrating the allocation plan creation process according to the first embodiment.

[0100] The allocation plan creation unit 115 receives the data table of the target position primary DB 231 (equivalent to the target position information 401) and the data table of the target human resources primary DB 232 (equivalent to the target human resources information 431) as parameters related to plan creation from the plan creation execution management unit 112, and stores them as temporary data (S1411).

[0101] Next, the mathematical programming loop processing management unit 211 acquires one row of vacant position information 421 to which the candidate talent is to be assigned from the target position primary DB 231, and executes a loop to calculate the candidate talent for the position 411 of the acquired row (S1412).

[0102] Next, the post-talent attribute match degree calculation unit 212 acquires target talent information from the target talent primary DB 232, target talent attribute information from the talent attribute DB 123, and target position requirement information from the position requirement DB 124 (S1413).

[0103] Next, the position-talent attribute match degree calculation unit 212 calculates the attribute match degree of the candidate talent for the target position 411 using each acquired information and mathematical programming, and inputs the candidate talent position information 233 to the transition information search unit 202 and the attribute match degree information 234 to the position-talent overall compatibility degree calculation unit 213 (S1414).

[0104] Next, the post-personnel overall suitability calculation unit 213 receives the attribute match information 234 and the most likely context information 235 calculated by the most likely context calculation unit 204 (S1415).

[0105] Next, the post-talent overall compatibility calculation unit 213 calculates an overall compatibility 306 for each candidate talent 1001 by multiplying the compatibility 304 of the attribute match information 234 by the transition likelihood 305 of the most likely context information 235, and selects the candidate talent with the highest overall compatibility 306 as the final candidate 307 (S1416).

[0106] Next, the mathematical programming loop process execution unit 211 determines whether there are any unprocessed rows in the target position primary DB 231, and if there are no unprocessed rows (S1417: NO), ends the processing loop execution (S1417).

[0107] Next, the calculation result primary management unit 214 holds the output of the position-personnel overall compatibility calculation unit 213, combines the held information, and inputs it as a final plan into the plan history DB 121 (S1418).

[0108] FIG. 15 is a diagram illustrating a mover determination screen according to the first embodiment.

[0109] The allocation plan created using the allocation plan creation system 100 is drawn on the operation terminal 111. The drawing GUI on the operation terminal 111 includes allocation information 1501 that displays allocation information for each position, and final context information by candidate 1502 that shows the most likely context information of each candidate for that position.

[0110] The allocation information 1501 includes a vacant post 1511 indicating the target position, a final candidate 1512 indicating a truth value of whether the candidate is the final candidate, a candidate 1513, an overall suitability 1514, and a calculation process 1515 for the overall suitability. The calculation process 1515 includes an attribute match degree 1516 and a likelihood of the most likely context 1517. Furthermore, the most likely context information for each candidate 1502 includes transition context information 1421 for each candidate.

[0111] FIG. 16 is a diagram illustrating an example of a hardware configuration of the information processing device according to the first embodiment.

[0112] The allocation plan creation system 100 includes information processing devices 1600. Each information processing device 1600 includes a calculation device 1601, a main memory 1602, an auxiliary storage device 1603, a communication interface 1604, a media interface 1605, and an input / output device 1606. The calculation device 1601 is a microprocessor such as a CPU (Central Processing Unit).

[0113] The main memory 1602 is a random access memory (RAM) or a read only memory (ROM) in which programs, data, etc. are loaded. The auxiliary storage device 1602 is a hard disk drive (HDD), a solid state drive (SSD), or a flash memory. The communication interface 1603 is a network interface card that communicates with other information processing devices. The media interface 1604 is connected to an external storage medium 1607. The input / output device 1605 includes an input device such as a keyboard or a touch panel, and an output device such as a monitor or a display.

[0114] The arithmetic unit 1601, main memory 1602, auxiliary storage device 1603, communication interface 1604, media interface 1605, and input / output device 1606 are connected to one another via a bus. The external storage medium 1607 is, for example, a communication medium (i.e., a wired, wireless, or optical network, or a carrier wave or digital signal propagating through the network), or an item that is detachable from the media interface 1605. The information processing device 1600 is connected to a management terminal (not shown) via the input / output device 1606.

[0115] Each function of each information processing device 1600 described above is realized by dedicated hardware, or by the arithmetic device 1601 reading and executing a program stored in the main memory 1602 or the auxiliary storage device 1603. Furthermore, each program may be pre-recorded in an external storage medium 1607, or may be installed when needed via a predetermined communication network.

[0116] According to this configuration, the allocation plan creation system 100, which creates an allocation plan that assigns candidate talent to positions, includes a transition likelihood calculation unit 114 and an allocation plan creation unit 115. The transition likelihood calculation unit 114 extracts transitions between multiple positions for each candidate talent as transition contexts based on a plan history that records each candidate talent's past positions, and calculates likelihoods from the extracted transition contexts. The allocation plan creation unit 115 identifies candidate talent that meets the constraints on the candidate talent's attributes for each position, calculates the match between the identified candidate talent's attributes and the position, calculates the suitability of the candidate talent for the position based on the calculated match and likelihood, and creates an allocation plan that assigns candidate talent to the position based on the calculated suitability.

[0117] As a result, by extracting transition contexts of more than two positions from past plan histories, it becomes possible to create plans that take the transition context into consideration during the process of creating an allocation plan using mathematical programming. This makes it possible to improve the validity of the allocation plan. Furthermore, the transition context is not only created based on plan histories created manually in the past, but also by adding and accumulating allocation plans created by the allocation plan creation system 100 as history, the information that forms the basis of the context is expanded, making it possible to continuously improve the validity of the context and the plans calculated using it. [Example]

[0118] A specific example of an allocation plan creation system according to an embodiment of the present invention will be described with reference to the drawings. Note that the present invention is not limited to the examples, but is defined by the claims.

[0119] In this embodiment, an example of applying the system to planning of a content different from that of the first embodiment is disclosed, and it is shown that the system is not limited to a specific industry or field.

[0120] FIG. 17 is a schematic diagram of an allocation plan creation system according to the second embodiment.

[0121] In this embodiment, an allocation plan proposal creation system 1700 is illustrated that creates an allocation plan in which crew members working for a route airline are assigned to crew tasks as an example of "assignees" to be assigned to individual flight schedules. This embodiment relates to the creation of an allocation plan for tasks, positions, or resources that can be assigned to crew members, and the allocation of crew members to tasks that make up a flight schedule is one application example of the invention.

[0122] Furthermore, this embodiment is an example of application to planning different from that of the first embodiment. However, the functions and processing flow constituting the allocation plan creation system are common. Application to a different planning requires changes to the data required for calculating the allocation plan and the tables that manage the data. Therefore, in this embodiment, an example of a management table different from that of the allocation plan creation system 100 of the first embodiment is illustrated.

[0123] 17 and the allocation plan creation system 100 in Fig. 1, the allocation plan creation system 1700 has a plan history DB 1711, a human resource attribute DB 1701, and a task requirement DB 1702. The plan history DB 1711, the human resource attribute DB 1701, and the task requirement DB 1702 are substituted for the human resource attribute DB 123 and the position requirement DB 124 of the allocation plan creation system 100, respectively.

[0124] FIG. 18 is a diagram illustrating a data structure of a plan history according to the second embodiment.

[0125] The plan history DB 1711 is a table that stores, as item values ​​(column values), a plan ID 301, a task ID 1701, a candidate talent ID 303, a compatibility 304, a transition likelihood 305, an overall compatibility 306, a final candidate 307, and a draft branch number 308. Each row 1811 of the plan history DB 1711 indicates information about each candidate talent. A group 312 of rows 311 of the plan history DB 1711 indicates information generated in one plan creation process. Note that the calculation method for the overall compatibility 306 is just an example, and it may be calculated using a process other than multiplication.

[0126] The plan ID 301 is an identifier assigned to a plan created in a single plan creation process performed by the user 101. The task ID 1701 is an identifier that identifies the task to be assigned to a candidate talent in the allocation plan. The candidate talent ID 303 is an identifier that identifies a candidate talent who is a candidate for assignment to a task. The compatibility 304 is the degree of compatibility between the task and talent attribute calculated by the task-talent attribute match degree calculation unit 212.

[0127] The transition likelihood 305 is a value calculated by the maximum context calculation function 204. The overall compatibility 306 is a value obtained by multiplying the compatibility 304 and the transition likelihood 305. The final candidate 307 is flag information for identifying the candidate 303 with the highest overall compatibility 306 among the candidates 303 of the task ID 1701. The draft branch number 308 is an identifier for identifying the local solution calculation process executed within the task-talent attribute match calculation function 212.

[0128] FIG. 19 is a diagram illustrating an example of a data structure of the human resource attribute DB according to the second embodiment.

[0129] The personnel attribute DB 1701 is a table that stores personnel ID 1901, current base 1902, language 1903, grade 1904, and route (1) flight count 1906 as item values ​​(column values).

[0130] Personnel ID 1901 is an identifier for the personnel engaged in flight operations. Current base 1902 indicates the airport where the personnel will be staying when the plan is formulated or when the plan is implemented. Language 1903 indicates the languages ​​that the personnel can use. Grade 1904 indicates the upper limit of flight operations that the personnel can perform. Route (1) flight count 1905 indicates the proficiency level of flight experience for each airline route. Each row of personnel attribute DB 1701 indicates individual personnel information 1911, which is an attribute for each personnel. Note that the item values ​​(table columns) of personnel attribute DB 1701 may have attributes other than the item values ​​described above.

[0131] FIG. 20 is a diagram illustrating an example of a data structure of the task requirement DB according to the second embodiment.

[0132] The task requirement DB 1702 is a table that stores, as item values ​​(column values), a task ID 2001, a flight ID 2002, a class 2003, a language 2004, a minimum interval 2005, a departure base 2006, and an arrival base 2007. Each row of the task requirement DB 1602 indicates a task-specific requirement 1911, which is a requirement for each task.

[0133] Task ID 2001 is an identifier for an individual task that makes up the flight schedule. Flight ID 2002 is an identifier for the flight schedule to which the task is linked. Grade 2003 is the grade of the human resource required to perform the task. Language 2004 is the language required of the person performing the task. Minimum interval 2005 is the minimum interval that the person performing the task must have between tasks that they have previously performed. Departure base 2006 is the departure base of the flight schedule to which the task is linked. Arrival base 2007 is the arrival base of the flight schedule to which the task is linked. The item values ​​of task requirement DB 1702 may have attributes other than the item values ​​described above.

[0134] With this configuration, by extracting transition contexts of more than two tasks from the past planning history, it becomes possible to create a plan that takes the transition context into consideration during the process of creating an allocation plan using mathematical programming, thereby improving the validity of the allocation plan.

[0135] The present invention is not limited to the above-described embodiment, but includes various modifications.

[0136] For example, the transition context extraction unit 203 may extract a transition pattern based on the most recently recorded plan history, thereby improving the accuracy of the plan history. (Appendix 1) An allocation plan creation system that creates an allocation plan in which candidate allocation targets are assigned to allocation destinations, a transition likelihood calculation unit that extracts transitions of the plurality of allocation destinations for each allocation target as transition contexts based on a planning history that records the allocation destinations for each past allocation target, and calculates likelihoods from the extracted transition contexts; an allocation plan creation unit that identifies candidate allocation targets that satisfy constraints on the attributes of the allocation targets for each allocation destination, calculates a degree of match between the attributes of the identified candidate allocation targets and the allocation destinations, calculates a suitability of the candidate allocation targets for the allocation destinations based on the calculated degree of match and the likelihood, and creates an allocation plan in which the candidate allocation targets are assigned to the allocation destinations based on the calculated suitability. (Appendix 2) The transition likelihood calculation unit a transition context extraction unit that extracts, based on the planning history, transitions of a plurality of consecutive assignment destinations for each assignment target as a transition pattern, and extracts the transition context from the extracted plurality of transition patterns; a maximum likelihood context calculation unit that calculates, as the likelihood, the length of a longest path included in the transition context extracted by the transition context extraction unit, (Appendix 1) An allocation planning system as described in the above. (Appendix 3) the allocation plan creation unit creates a primary allocation plan based on the degree of match, and corrects the created primary allocation plan based on the degree of match to create the allocation plan. (Appendix 1) An allocation planning system as described in the above. (Appendix 4) a display terminal that displays the allocation plan created by the allocation plan creation unit; (Appendix 1) An allocation planning system as described in the above. (Appendix 5) the transition context extraction unit extracts a transition pattern of another candidate assignment target having an attribute similar to that of the candidate assignment target; (Appendix 2) An allocation planning system as described in the above. (Appendix 6) the transition context extraction unit extracts the transition pattern based on the plan history whose record is most recent; (Appendix 2) An allocation planning system as described in the above. (Appendix 7) An allocation plan creation method for creating an allocation plan in which candidate allocation targets are assigned to allocation destinations, extracting transitions of the plurality of allocation destinations for each allocation target as transition contexts based on a planning history that records the allocation destinations for each past allocation target, and calculating likelihoods from the extracted transition contexts; An allocation plan creation method that identifies candidate allocation targets that satisfy constraints on the attributes of the allocation targets for each allocation destination, calculates the degree of match between the attributes of the identified candidate allocation targets and the allocation destinations, calculates the suitability of the candidate allocation targets for the allocation destinations based on the calculated degree of match and the likelihood, and creates an allocation plan that allocates the candidate allocation targets to the allocation destinations based on the calculated degree of suitability. (Appendix 8) The step of creating the allocation plan includes creating a primary allocation plan based on the degree of match, and correcting the created primary allocation plan based on the degree of match to create the allocation plan. The method for creating an allocation plan as described in (Appendix 1). (Appendix 9) An allocation plan creation method for creating an allocation plan in which candidate allocation targets are assigned to allocation destinations, extracting transitions of the plurality of allocation destinations for each allocation target as transition contexts based on a planning history that records the allocation destinations for each past allocation target, and calculating likelihoods from the extracted transition contexts; an allocation plan creation program having a step of identifying candidate allocation targets that satisfy constraint conditions of the attributes of the allocation targets for each allocation destination, calculating a degree of match between the attributes of the identified candidate allocation targets and the allocation destinations, calculating a suitability of the candidate allocation targets for the allocation destinations based on the calculated degree of match and the likelihood, and creating an allocation plan in which the candidate allocation targets are assigned to the allocation destinations based on the calculated suitability. [Explanation of symbols]

[0137] 100... allocation plan creation system, 114... transition likelihood calculation unit, 115... allocation plan proposal creation unit, 203... transition context extraction unit, 204... maximum likelihood context calculation unit, 234... attribute match degree information, 235... maximum likelihood context information, 1411... vacant post, 1412... final candidate, 1413... candidate talent, 1414... overall suitability, 1416... attribute match degree, 1417... likelihood, 1700... allocation plan creation system,

Claims

1. An allocation plan creation system that creates an allocation plan in which candidate allocation targets are assigned to allocation destinations, a transition likelihood calculation unit that extracts transitions of a plurality of consecutive assignment destinations for each assignment target as transition patterns based on a planning history that records the assignment destinations for each past assignment target, creates a transition context from the extracted plurality of transition patterns, calculates a transition pattern that maximizes the sum of weighted values ​​assigned to each transition between the plurality of assignment destinations included in the created transition context as a most likely context, and calculates the sum of weighted values ​​of each transition included in the calculated most likely context as a likelihood; an allocation plan creation unit that identifies candidate allocation targets that satisfy constraints on the attributes of the allocation targets for each allocation destination, calculates a degree of match between the attributes of the identified candidate allocation targets and the allocation destinations, calculates a suitability of the candidate allocation targets for the allocation destinations based on the calculated degree of match and the likelihood, and creates an allocation plan in which the candidate allocation targets are assigned to the allocation destinations based on the calculated suitability.

2. the allocation plan creation unit creates a primary allocation plan based on the degree of match, and corrects the created primary allocation plan based on the degree of match to create the allocation plan. The allocation planning system of claim 1 .

3. a display terminal that displays the allocation plan created by the allocation plan creation unit; The allocation planning system of claim 1 .

4. the transition likelihood calculation unit extracts a transition pattern of other candidate assignment targets having attributes similar to those of the candidate assignment target; The allocation planning system of claim 1 .

5. the transition likelihood calculation unit extracts the transition pattern based on the most recently recorded planning history. The allocation planning system of claim 1 .

6. An allocation plan creation method by an allocation plan creation system that creates an allocation plan in which candidate allocation targets are assigned to allocation destinations, extracting, as a transition pattern, transitions of a plurality of consecutive assignment destinations for each assignment target based on a planning history that records the assignment destination for each past assignment target, creating a transition context from the extracted plurality of transition patterns, calculating, as a most likely context, the transition pattern that maximizes the sum of weighted values ​​assigned to each transition between the plurality of assignment destinations included in the created transition context, and calculating, as a likelihood, the sum of weighted values ​​of each transition included in the calculated most likely context; An allocation plan creation method that identifies candidate allocation targets that satisfy constraints on the attributes of the allocation targets for each allocation destination, calculates the degree of match between the attributes of the identified candidate allocation targets and the allocation destinations, calculates the suitability of the candidate allocation targets for the allocation destinations based on the calculated degree of match and the likelihood, and creates an allocation plan that allocates the candidate allocation targets to the allocation destinations based on the calculated degree of suitability.

7. The step of creating the allocation plan includes creating a primary allocation plan based on the degree of match, and correcting the created primary allocation plan based on the degree of match to create the allocation plan. The allocation planning method according to claim 6 .

8. An allocation plan creation program to be executed by an allocation plan creation system that creates an allocation plan in which candidate allocation targets are assigned to allocation destinations, The allocation plan creation system, a step of extracting, as a transition pattern, transitions of a plurality of consecutive assignment destinations for each assignment target based on a planning history that records the assignment destinations for each past assignment target, creating a transition context from the extracted plurality of transition patterns, calculating, as a most likely context, the transition pattern in which the sum of weighted values ​​assigned to each transition between the plurality of assignment destinations included in the created transition context is the largest, and calculating, as a likelihood, the sum of weighted values ​​of each transition included in the calculated most likely context; an allocation plan creation program for executing the steps of identifying candidate allocation targets that satisfy constraint conditions for the attributes of the allocation targets for each allocation destination, calculating the degree of match between the attributes of the identified candidate allocation targets and the allocation destinations, calculating the suitability of the candidate allocation targets for the allocation destinations based on the calculated degree of match and the likelihood, and creating an allocation plan in which the candidate allocation targets are assigned to the allocation destinations based on the calculated degree of suitability.

Images