JP6251663B2 - Task allocation device, task allocation method, and task allocation program - Google Patents

Description

  The present invention relates to a task assignment processing technique for distributing a plurality of tasks to a plurality of people using a computer.

  Currently, when assigning multiple jobs (hereinafter referred to as tasks) to multiple people (hereinafter referred to as workers), restrictions such as the upper and lower limits of the number of tasks per person and the upper and lower limits of the number of workers per task Efforts are being made to achieve efficient allocation by performing optimization calculations that maximize the number of worker-to-task allocations while satisfying conditions.

In addition, when a worker “i” is responsible for a task “j” and a cost “c _ij ” is required, an effort is made to realize efficient allocation by performing optimization calculation that minimizes the total cost. Has been done. The problem of finding the optimal allocation method for these worker-to-tasks is called the “allocation problem”.

  In general, the worker-to-task “maximum allocation problem” involves assigning workers and tasks to two vertex sets in a bipartite graph (a graph in which the vertex set is divided into two and the edges are limited to connecting the subsets). Corresponding and connecting edges between assignable workers and task vertices can be replaced with the “maximum matching problem for bipartite graphs”. The “maximum matching problem of bipartite graph” is to select a pair that maximizes the number of assignments of the bipartite graph under given constraints.

  Further, as shown in FIG. 1, the “maximum matching problem of bipartite graph” adds a new vertex as a source (src), adds an edge toward a vertex set corresponding to the worker, and a sink (sink). By adding new vertices and adding edges from the vertex set to which the task corresponds, it becomes possible to convert to a “maximum flow problem” that maximizes the flow rate from the source to the sink. Note that the “maximum allocation minimum cost problem” for worker vs. task is the “minimum cost flow problem” that selects the edge that minimizes the cost when the graph is converted to maximize the flow from the source to the sink. Can be converted.

  Both the “maximum flow problem” and the “minimum cost flow problem” can be solved by linear programming. In particular, a plurality of efficient algorithms for solving the “maximum flow problem” have been proposed in Non-Patent Document 1.

  In the following, background techniques for solving the worker-to-task assignment problem in the crowdsourcing service will be described. Here, crowdsourcing is a framework in which tasks are entrusted to an unspecified number of workers. For example, there are some that have already been established as commercial services, such as Lancers in Non-Patent Document 2.

  In particular, there is a case where a task is given a spatio-temporal condition and the task cannot be completed unless the worker physically moves to a designated location. In this case, non-patent literature solved by converting the “maximum allocation problem” of worker-to-task into “maximum flow problem” by defining the maximum allowable number and maximum allowable radius that a worker can perform tasks. There are three studies.

  Since this technology creates a graph using information on workers and tasks at a single point in time and performs optimization calculations, it can simultaneously handle the case where the cost of completing a worker and the task completion by the worker changes with time. Instead, it uses a greedy method of assigning tasks one by one in order.

Dinitz, Y., Garg, N., & Goemans, M. X. (1999) .On the single-source unsplittable flow problem.Combinatorica, 19 (1), 17-41. “Crosssourcing is Japan's largest“ Lancers ”” [online], searched on October 21, 2014, Internet URL <http://www.lancers.jp/> Kazemi, L., & Shahabi, C. (2012). Geocrowd: enabling query answering with spatial crowdsourcing.In Proceedings of the 20th International Conference on Advances in Geographic Information Systems (pp. 189-198).

  However, when assigning tasks at multiple points in time using the technology of Non-Patent Document 3 so that the task completion time is the fastest, even if the “minimum cost flow problem” is solved using the task completion time as the cost, Only a solution can be obtained.

  Therefore, there may be a case where the worker set and the cost of task completion by the worker change depending on the time cannot be handled at the same time. In addition, in the greedy method in which optimization is performed in order for each time point, a temporal global optimal solution cannot always be obtained, and only a local optimal solution may be obtained.

  The present invention has been made to solve such a conventional problem, and an object of the present invention is to obtain a global optimum solution in the time direction by performing optimization at a plurality of points in time.

  The task assignment device of the present invention includes a first database in which a time ID, a worker ID, and a task ID are associated with each other, a second database in which position information where the task occurs is stored, and a worker ID and a task from the first database. Means for extracting a pair of IDs, time ID of the pair, means for searching task position information from the second database using the extracted task ID as a key, and worker using the extracted worker ID as a key Means for retrieving the worker's position information for each time from the third database storing the position information for each time, and calculates the time when the worker moves to the position where the task occurs for each time indicated by the time ID. And means for selecting and outputting the earliest time among the calculated times, the worker corresponding to the time, the task, and the time ID of the time.

  The task assignment method of the present invention includes a first step of extracting a worker ID / task ID pair and a time ID of the pair based on a first database in which a time ID, a worker ID, and a task ID are associated with each other; A second step of searching for task position information from a second database in which position information where the task occurs is stored using the extracted task ID as a key, and a position for each worker time using the extracted worker ID as a key A third step of retrieving worker position information for each time from a third database in which information is stored; and a fourth step of calculating a time when the worker moves to a position where a task occurs at each time indicated by the time ID. And a fifth step of selecting and outputting the earliest time among the calculated times, the worker corresponding to the time, the task, and the time ID of the time. That.

  In addition, this invention is good also as an aspect of the program which functions a computer as the said task allocation apparatus. This program can be provided through a network or a recording medium.

  According to the present invention, it is possible to obtain a global optimum solution in the time direction by performing optimization at a plurality of points in time.

Conversion example of bipartite graph. The block diagram of the task assignment apparatus which concerns on embodiment of this invention. The flowchart which shows the process of an allocation search function part. The flowchart which shows the process of a cost calculation function part. The flowchart which shows the process of a constraint equation calculation function part. The capacity | capacitance example of each edge | side in a bipartite graph. Explanatory drawing of "δ + ν, δ-ν". The flowchart which shows the process of an allocation optimization function part.

≪Example of device configuration≫
A configuration example of the task assignment device according to the embodiment of the present invention will be described based on FIG. This task assignment device 001 is intended to solve the worker-to-task assignment problem in crowdsourcing.

  Specifically, the task assignment device 001 is configured by a computer and includes hardware leases such as a CPU, a main storage device (RAM, ROM, etc.), an auxiliary storage device (hard disk drive device, solid state drive device, etc.), and the like.

  As a result of cooperation between the hardware resource and the software resource (OS, application, etc.), the task assignment device 001 has a time interval DB005, a task DB010, a worker DB020, an assignment candidate set search function unit 030, an assignment candidate set DB040, and a cost calculation. A function unit 050, a cost DB 060, a constraint formula calculation function unit 070, a constraint formula DB 080, an allocation optimization calculation function unit 090, and an allocation set DB 100 are mounted.

  Each DB 005, 010, 020, 040, 060, 080, 100 is assumed to be constructed in the storage device. Note that the task assignment device 001 may be configured as a single computer, or the units 010 to 100 may be distributed among a plurality of computers. Hereinafter, processing contents of the components 010 to 100 will be described.

  << Time interval DB005, Task DB010, Worker DB020 >>

  Table 1 shows an example of the data structure of the time interval DB 005. The time interval DB 005 stores information on worker and task allocation intervals in a time format of “YYY-MM-DD”. For example, in Table 1, time information on February 2, 2014 is stored every other hour. In the time interval DB 005, a time interval ID is assigned to each record in order to designate the record from the outside.

  Table 2 shows an example of the data structure of the task DB 010. This task DB 010 stores information on the latitude and longitude of the task. In the task DB 010, a task ID is assigned to each record in order to designate the record from the outside.

  Table 3 shows an example of the data structure of the worker DB 020. The worker DB 020 stores latitude / longitude information, maximum moving distance information, information on the maximum number of accepted tasks, and moving means information for each worker time.

  The maximum movement distance in Table 3 indicates the maximum distance that the worker can move to perform a task at that time. The maximum number of accepted tasks indicates the maximum number of tasks that can be performed by a worker, and the same value is assigned to one worker and stored without changing every time.

  Further, the moving means indicates moving means information such as walking, car or bus. Here, it is assumed that the worker's future location, maximum moving distance, and moving means have already been input by the worker and that information is known. Alternatively, it is assumed that the future schedule is estimated using the worker's past log. The worker DB 020 also has a worker ID assigned to each record in order to designate the record from the outside.

<< Allocation Candidate Group Search Function Unit 030, Allocation Candidate Group DB040 >>
The allocation candidate group search function unit 030 receives the storage data of the DB 005, 010, 020 and outputs the processing information to the allocation candidate group DB 040 for storage. Here, the allocation candidate group search function unit 030 searches for a group (pair) of “time, worker, task” that can be allocated based on the data stored in the DB 005, 010, 020, and uses the search result as the allocation candidate group. .

  The processing contents of the allocation candidate group search function unit 030 will be described below with reference to FIG. First, when the process is started, a set N is prepared to be an empty set (S1-1). Thereafter, the record set is searched from the time interval DB 005, the search result is set to “H” (S1-2), the unprocessed element is extracted from the set H, and “h” is set as the extracted element (S1-3). A record set is searched from the task DB 010, and the search result is set to a set “T” (S1-4).

  Next, a record set whose time coincides with the element “h” is searched from the worker DB 020, and the search result is set to “W” (S1-5). An unprocessed element is extracted from the set “W” and “w” is obtained. (S1-6).

  Subsequently, “T” having the latitude and longitude as coordinates at the inner side of the circle having the radius of the maximum movement distance with the latitude and longitude of the element “w” as the center is searched. Thereby, an allocation candidate set of “time, worker, task” is specified, and each ID (time interval ID, worker ID, task ID) of the specified allocation candidate set is added to the set “N” (S1-7). .

  This process may be a circle inside / outside determination or a rectangle inside / outside determination with the maximum moving distance as a diagonal line in order to confirm whether or not the task position corresponds to the distance condition designated by the worker.

  Then, it is confirmed whether or not the unprocessed element “w” exists in the set “W” (S1-8). If it exists, the process returns to S1-6, and if it does not exist, the process proceeds to S1-9. In S1-9, it is confirmed whether or not an unprocessed element “h” exists in the set “H”. If it exists, the process returns to S1-3, and if it does not exist, the process proceeds to S1-10. In S1-10, the information of the set N is output and stored in the allocation candidate set DB 040, and the process ends.

  Table 4 shows an example of the data structure of the allocation candidate set DB040. The allocation candidate DB 040 stores the time interval ID, worker ID, and task ID of the allocation candidate group.

<< Cost calculation function unit 050, Cost DB 060 >>
The cost calculation function unit 050 receives the storage data of the allocation candidate set DB 040 as input, and outputs the processing data to the cost DB 060 for storage. Specifically, the cost calculation function unit 050 selects from among multiple pairs the “worker, task” pair of allocation candidate sets that has the minimum task completion time, and outputs the selected cost information to the cost DB 060. Hereinafter, the processing content of the cost calculation function unit 050 will be described with reference to FIG.

  S2-1: First, a set C is prepared when processing is started. The set C at this stage is an empty set.

  S2-2: Next, the allocation candidate group DB 040 is read, and unprocessed “worker, task” pairs are extracted and set as a set “P” for a plurality of time points. That is, unprocessed “worker, task” pairs for a plurality of times are collected to form a set P.

  This process will be described based on the data example in Table 4. Here, the time is represented by “τ”, the worker is represented by “w”, the task is represented by “t”, the subscript is represented by each ID, and the assignment candidate DB 040 is processed in order from the top row.

Therefore, (w ₁ , t ₁ ) is first selected as an unprocessed “worker, task” pair, and the set “P” for a plurality of time points is “P = {(τ ₁ , w ₁ , t ₁ ), ( τ ₂ , w ₁ , t ₁ ), (τ ₃ , w ₁ , t ₁ )} ”.

  S2-3: The record that matches the task ID of the set P is searched from the task DB 010, the record that matches the worker ID of the set P is searched from the worker DB 020, and the fastest task completion time “min_time” is calculated.

  Here, it is assumed that the task can be completed in a short time such as taking a picture of a specified place. Therefore, arriving at the location specified for a task is synonymous with or close to task completion.

The calculation of the fastest task completion time “min_time” is calculated based on “P = {(τ ₁ , w ₁ , t ₁ ), (τ ₂ , w ₁ , t ₁ ), (τ ₃ , w ₁ , t ₁ )}. First, the distance “dist ₁ , dist ₂ , dist ₃ ” at each time “τ ₁ , τ ₂ , τ ₃ ” is calculated using the latitude and longitude information of the worker and the task.

Here, it is assumed that the worker moving means is already input to the task assignment device 001 by the worker and stored in the worker DB 020. Therefore, the required time “duration _k ” required for the worker to move to the task position by the moving means at a certain time “τ _k ” can be obtained by equation (1).

Also, the time “time _i ” when the worker has moved to the task position at a certain time “τ _k ” by the moving means may be obtained by adding the movement start time to the required time “duration _k ”. 2).

  As a result, the earliest time “min_time” at which the worker completes the task can be obtained by taking these minimum values as shown in Expression (3).

  S2-4: The time “min_time”, the worker ID, the task ID, and the time ID of the time “min_time” are added to the set C. When there are a plurality of time IDs having the time “min_time”, a time ID to be added to the set C can be uniquely determined by providing a rule such as selecting the earliest time.

  S2-5, S2-6: It is confirmed whether or not there is an unprocessed set P (S2-5). If it exists, the process returns to S2-2, but if it does not exist, the process proceeds to S2-6. In S2-6, the information of the set C is output and stored in the cost DB 060, and the process is terminated.

  At this time, the time “min_time” is stored in the cost DB 060 as cost information. Table 5 shows an example of the data structure of the cost DB 060 in which the time ID, worker ID, task ID, and cost information are stored.

<< Constraint Expression Calculation Function Unit 070, Constraint Expression DB080 >>
The constraint equation calculation function unit 070 receives the storage data of the allocation candidate DB 040 as input, and outputs the processing information to the constraint equation DB 080 for storage. Here, the worker-to-task maximum assignment problem is replaced with the aforementioned minimum cost flow problem.

  Specifically, the constraint equation calculation function unit 070 calculates and outputs a capacity constraint equation and a flow rate storage constraint equation in the graph shown in FIG. The processing contents of the constraint equation calculation function unit 070 will be described below with reference to FIG.

  S3-1: First, when processing is started, a set “R” is prepared. The set “R” at this stage is an empty set.

  S3-2: Next, the allocation candidate set DB 040 is read to extract an unprocessed “worker, task” pair, and the extracted pair is set to “P ′”. This process will be described based on the data example in Table 4.

Here, since processing is performed in order from the upper row of the allocation candidate group DB 040, “P ′ = {w ₁ , t ₁ }” is extracted as the pair “P” of “worker, task” to be processed first, and thereafter “P ′ = {w ₂ , t ₁ }”, “P ′ = {w ₂ , t ₂ }”, “P ′ = {w ₂ , t ₃ }” are taken in order.

  S3-3: For the edge connecting each worker from the source (src) of the bipartite graph as shown in FIG. 1, a capacity restriction constraint formula with the maximum number of tasks accepted by the worker as an upper limit is created. Add to R.

At this time, the flow rate of water flowing from “node m” to “node n” is set to “x _{m, n} ” as a decision variable for performing the optimization calculation. Here, the capacity constraint means a constraint that the flow rate of each side does not exceed the capacity.

In other words, for each worker from the source, to create a capacity restriction constraint equation with the maximum number of accepted tasks of the worker as an upper limit, if the maximum number of accepted tasks of “worker w _i ” is “maxT _i ”, Equation (4) Will be created.

  For example, in the bipartite graph of FIG. 6, formula (5) is created.

  This process is the same process as in Non-Patent Document 1.

S3-4: A capacity constraint expression having an upper limit of “1” is created for the side connecting each task and the sink and the side connecting each worker and each task, and is added to “R”. When the decision variable “x _{m, n} ” is used in the same manner as in S3-3, a capacity constraint expression with “1” as the upper limit is created for the side connecting each task and the sink and the side connecting each worker and each task. Then, formula (6) is created.

  For example, in the bipartite graph of FIG. 6, formula (7) is created.

  This process is the same process as in Non-Patent Document 3.

  S3-5: A constraint expression for the flow rate conservation rule is created for worker vertices and task vertices and added to “R”. The flow conservation law means a constraint that the flow rate flowing into the apex is equal to the flow rate flowing out.

  That is, to create a constraint equation for the flow rate conservation rule for worker vertices and task vertices, as shown in FIG. If it is set as the starting point set of (8), Formula (8) will be created.

  For example, if it is the bipartite graph of FIG. 6, Formula (9) will be created.

  This process is the same process as in Non-Patent Document 3.

  S3-6: It is confirmed whether or not an unprocessed pair “P ′” exists in the allocation candidate group DB040. As a result of the confirmation, if it exists, the process returns to S3-2, and if it does not exist, the process proceeds to S3-7.

  S3-7: The information of the set R is output and stored in the constraint expression DB 080 (S3-8), and the process is terminated. Table 6 shows an example of the data structure of the constraint expression DB 080, and information on each constraint expression used when performing optimization calculation is stored in a text format. Each constraint equation is assigned a constraint equation ID for record search.

<< Allocation optimization calculation function unit 090 >>
The allocation optimization calculation function unit 090 receives the storage data of the cost DB 060 and the constraint equation DB 080 and outputs the processing information to the allocation set DB 100. Specifically, the allocation optimization calculation function unit 090 solves the “maximum flow problem” in the constraint expression given from the constraint expression DB 080 and uses the result to solve the “minimum cost flow problem”. The processing contents of the allocation optimization calculation function unit 090 will be described with reference to FIG.

  S4-1: A record set is searched from the constraint expression DB 080, and the search result is set to a set “R ′”.

  S4-2: The maximum flow rate flowing out from the source under the condition of the constraint equation “R” is obtained and is set to “maxF”. This maximum flow rate “maxF” is obtained from the above decision variable as shown in equation (10).

  The optimization calculation for obtaining the maximum flow rate “maxF” can be solved by linear programming, and a plurality of efficient algorithms have been proposed, so these may be used (for example, Non-Patent Document 1) ).

  S4-3: Create a constraint equation with “maxF” as the upper limit of the flow rate flowing out from the source and the upper limit of the flow rate flowing into the sink, and add it to “R ′”. Here, the constraint expression of Expression (11) may be created.

S4-4: A record set is searched from the cost DB 060 to create a cost function. The cost function is a sum of costs between assignments (workers, tasks) that is an optimization objective function. That is, if the worker set and the task set stored in the cost DB 060 are “w” and “T”, respectively, and the cost of the worker “w _i ” and the task “t _j ” is “c _ij ”, the formula ( 12) may be calculated.

However, the capacity between the worker and the side between the tasks is set to “1” or less because of the capacity restriction of S3-4. In addition, the cost DB 060 does not necessarily calculate the cost for every worker / task pair, but it can be calculated by substituting a sufficiently large number on the assumption that the cost is “c _ij = ∞”. . In the case of a nonexistent pair, it can be calculated by skipping the calculation.
S4-5: The cost function is minimized under the condition of “R ′”. Here, the minimization of the cost function can be solved by linear programming. A pair set of “workers and tasks” selected at the time of minimization, that is, a pair set of “workers and tasks” whose decision variable “x _{m, n} ” between worker and task is 1 or more is defined as “opt_p”.

  S4-6: Search for “opt_p” from the cost DB 060, and output the information on the pair of “time, worker, task” including the time ID at that time (time ID, worker ID, task ID) to the allocation group DB100. Store, and finish the process.

  Table 7 shows an example of the data structure of the allocation set DB 100. Here, “time ID, worker ID, task ID” is stored for a pair of “time, worker, task” that minimizes the sum of time until task completion as a result of optimization calculation.

  According to such a task assignment device 001, among the “worker, task” pairs extracted from the assignment candidate set “time, worker, task” of the assignment candidate set DB 040 in S2-3 of the cost calculation function unit 050. Since the minimum time of task completion times at a plurality of time points is calculated, it is possible to calculate a cost function with the fastest task completion time as a cost.

  As a result, optimization at a plurality of points in time can be performed at a time, and a global optimum solution in the time direction can be obtained. Therefore, even if the task completion time varies depending on the worker set or the worker completion time depending on the time like the crowdsourcing service where the space-time condition is given to the task, the “worker, task” pair that minimizes the task completion time It can be selected, and the fastest task completion time can be realized.

  As a result, the satisfaction of the task requester who wants to finish the work as soon as possible can be improved. The task assignment device 001 is particularly useful in an environment where the worker's future schedule can be grasped and predicted.

≪Program≫
The present invention is not limited to the above-described embodiments, and can be applied and modified within the scope of the claims. For example, the present invention can be configured as a document search program that causes a computer to function as a part or all of the units 005 to 100 of the task assignment device 001.

  According to this program, it is possible to cause a computer to execute part or all of S1-1 to S1-10, S2 · 1 to S2-6, S3-1 to S3-7, and S4-1 to S4-6. It becomes.

  The program can be provided through a network such as a website or e-mail. The program is stored in a recording medium such as a CD-ROM, DVD-ROM, CD-R, CD-RW, DVD-R, DVD-RW, MO, HDD, BD-ROM, BD-R, or BD-RE. It is also possible to record, save and distribute. This recording medium is read using a recording medium driving device, and the program code itself realizes the processing of the above embodiment, so that the recording medium also constitutes the present invention.

001 ... Task allocation device 001
005 ... Time interval DB
010 ... Task DB (second database)
010 ... Worker DB (third database)
030 ... Allocation candidate group search function unit 040 ... Allocation candidate group DB040 (first database)
050 ... Cost calculation function unit 050
060 ... Cost DB (4th database)
070: Constraint equation calculation function unit 080: Constraint equation DB (fifth database)
090 ... Allocation optimization calculation function unit 100 ... Allocation group DB

Claims (7)

A first database in which a time ID, a worker ID, and a task ID are associated;
A second database storing location information where the task occurs;
Means for extracting a worker ID and task ID pair and a time ID of the pair from the first database;
Means for retrieving task position information from the second database using the extracted task ID as a key;
Means for retrieving worker position information for each time from a third database in which the position information for each worker time is stored using the extracted worker ID as a key;
Means for calculating a time when the worker moves to a position where the task occurs at each time indicated by the time ID;
Means for selecting and outputting the earliest time among the calculated times, the worker and task corresponding to the time, and the time ID of the time;
A task assignment device comprising: The third database further stores moving means for each worker ID,
The time when the worker moves to the position where the task occurs at the time indicated by the time ID using the moving means,
2. The task assignment apparatus according to claim 1, further comprising means for calculating according to equations (1) and (2).

duration _k : time required for the worker to move to the task position at a certain time τ _k ,
dist _k : distance to the position where the task occurs at each time,
v _k : moving speed

The third database stores the maximum movement distance for each worker ID,
Means for determining whether there is a task within the maximum movement distance of the worker based on the two databases and the third database, and storing the time ID, worker ID and task ID in association with each other in the first database; ,
A fourth database for storing the selected earliest time, worker, task, and time ID in association with each other;
Means for creating a constraint equation given when the worker-to-task maximum allocation problem is replaced with a bipartite graph minimum cost flow problem based on the fourth database, and storing the created constraint equation in the fifth database;
Means for calculating a time / worker / task pair that minimizes the time until task completion for each time ID based on the fourth database and the fifth database;
The task assignment apparatus according to claim 2, further comprising: A task assignment method for distributing a plurality of tasks to a plurality of workers using a computer,
A first step of extracting a worker ID / task ID pair and a time ID of the pair based on a first database in which a time ID, a worker ID, and a task ID are associated;
A second step of retrieving task position information from a second database in which position information where the task occurs is stored using the extracted task ID as a key;
A third step of retrieving worker position information for each time from a third database in which position information for each worker time is stored using the extracted worker ID as a key;
A fourth step of calculating a time when the worker moves to a position where the task occurs at each time indicated by the time ID;
A fifth step of selecting and outputting the earliest time among the calculated times, the worker and task corresponding to the time, and the time ID of the time;
A task assignment method characterized by comprising: In the case where the third database further stores moving means for each worker ID,
In the fourth step, the time when the worker moves to the position where the task occurs at the time indicated by the time ID using the moving means,
5. The task assignment method according to claim 4, further comprising a step of calculating according to equations (1) and (2).

duration _k : time required for the worker to move to the task position at a certain time τ _k ,
dist _k : distance to the position where the task occurs at each time,
v _k : moving speed

The third database stores the maximum movement distance for each worker ID,
In the case where the selected earliest time, worker, task, and time ID are associated and stored in the fourth database,
Determining whether a task exists within the maximum movement distance of the worker based on the second database and the third database, and storing the time ID, worker ID, and task ID in the first database in association with each other. When,
Creating a constraint equation given when replacing the worker-to-task maximum allocation problem with a bipartite graph minimum cost flow problem based on the fourth database, and storing the created constraint equation in the fifth database;
Calculating a time / worker / task pair that minimizes the time until task completion for each time ID based on the fourth database and the fifth database;
The task assignment method according to claim 5, further comprising:   A task assignment program for causing a computer to function as the task assignment device according to claim 1.

Images