JP2746351B2 - Job / resource allocation optimization device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an apparatus for optimizing the assignment between a plurality of jobs and a plurality of resources by a computer , for example, to efficiently create a schedule of shift work for humans. The present invention relates to an effective job resource allocation optimizing device .

[0002]

2. Description of the Related Art Assignment between a plurality of jobs and a plurality of resources under certain constraints includes, for example, assignment of shift work, creation of a machine maintenance plan, creation of a transport dispatch plan, and production plan. This is a routine work in a wide variety of fields, such as creation and aircraft layout.

Conventionally, such assignment creation has been performed exclusively by hand using expert know-how. However, in recent years, automation has been advanced with the development of computer technology. As a method of performing the assignment by the computer, there are an approach based on a mathematical programming such as a linear programming and a knowledge engineering approach. The former is effective only when the format of the constraint condition is established, such as when the constraint condition can be described by an equation or an inequality. However, in an actual problem, the constraints are often described in a more complicated format. Therefore, the heuristic method based on the latter approach has a wider range of application.

FIG . 4 shows a conventional assignment processing procedure based on a knowledge engineering approach. (1) A plurality of jobs are sorted and stored according to the priority of each job. (Step 101) (2) Check if there is a job for which resource allocation has not been performed, and if there is, check out one of the jobs. (Step 1
02, 103) (3) If there is no unassigned job, the process is terminated. (4) Check if there is a resource that can be allocated to the job extracted in (2), and if so, select one of the resources. (Steps 104 and 105) (5) If there is no resource that can be allocated in (4), the resource allocation to the job is determined to have failed, and the process returns to (2) to select the next unallocated job. (6) It is checked whether the resource extracted in step 105 satisfies the constraint condition of the job selected in step 103. (Step 106) (7) If the resource does not satisfy the constraint condition of the job in the check of Step 106, another resource is taken out and the constraint condition check of Step 106 is performed again. (8) If the extracted resource satisfies the constraint,
Record the job and resource assignments. (Step 107) After (9) and (8), return to step 102 and repeat the process until there are no unassigned jobs.

[0005] In such an assignment, the order in which jobs are assigned has a great effect on the efficiency of the assignment. If this order is inappropriate, the backtracking process in which the once determined schedule is broken down and the allocation is performed again frequently occurs. Conventionally, such a problem has been mainly solved by the following method. (1) The difficulty of assignment of each job is determined in advance, and assignment is performed in order from the job with the greatest difficulty. (2) Recalculate the difficulty of assigning each job each time the assignment of one job is determined.

[0006] As a known example related to these, there is, for example, Japanese Patent Application Laid-Open No. 63-163669 "assignment determination method".

[0007]

In the preparation of various schedules, various restrictions on resource allocation are often imposed between jobs that are temporally adjacent to each other. For example, in the case of staffing of shift work, there is often a constraint that night shift cannot be assigned to the same person continuously. Further, even in the production process schedule creation, there may be a restriction on the types of jobs that can be continuously executed by the machine as a resource.

In the above-mentioned conventional method, the assignment to each job is made in order to give priority to the job which is most difficult to assign.
In some cases, assignment is not performed after a temporally continuous job, and it is not possible to realize an assignment that sufficiently satisfies the constraints imposed on temporally continuous jobs. For example, when the conventional method is applied to the example of FIG. 5, the resource allocation order for each job is as follows: job 2 → job 12 → job 3 → job 11 → job 4 →, and good schedule allocation cannot be realized. That is, when the resources are allocated to the jobs in such a complete intermittent order, the schedule creation result has high discontinuity and is difficult to use.

To avoid this drawback, when several consecutive jobs are grouped and assigned, the constraints imposed on temporally successive jobs are achieved,
Since the difficult part of the assignment is postponed, assignment failures and retries are likely to occur.

SUMMARY OF THE INVENTION An object of the present invention is to automatically assign a plurality of jobs and a plurality of resources by using a computer , and to impose on a job which is temporally continuous without delaying a portion difficult to assign. It is an object of the present invention to provide a job / resource allocation optimizing apparatus which improves allocation accuracy and reduces allocation failures by reflecting a highly localized restriction condition such as a restriction condition to be applied to the determination of a job selection order. .

[0011]

In order to achieve the above object, the present invention provides a method for determining the assignment between a plurality of jobs and a plurality of resources according to a constraint condition. In some cases, priority is given to the difficulty in allocating resources to each job and the allocation of jobs that are temporally adjacent to each other.

[0012]

According to the present invention, the assignment of a plurality of jobs and a plurality of resources is determined by the following procedure. (1) For each job, find a numerical value indicating the difficulty of assignment. (2) Select the job with the highest numerical value indicating difficulty. (3) Select a peripheral job temporally adjacent to the job having the highest difficulty. (4) Among the selected peripheral jobs, the most difficult
Resource allocation to the job in the order close to the job with the highest time . (5) Repeat the process from (2) for the remaining jobs.

By using the present invention, when allocating a plurality of jobs to a plurality of resources, the job which is most difficult to be assigned is prioritized, and at the same time, jobs which are temporally adjacent to each other are sequentially assigned. The schedule result can satisfy various local constraints.

[0014]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS One embodiment of the present invention will be described below with reference to the drawings.

FIG. 1 is a system configuration diagram of an embodiment of the present invention. The system includes a job storage unit 1, a resource storage unit 2, a constraint information storage unit 3, a job selection device 4, a resource selection device 5, an assignment determination device 6, and an assignment status storage device 7. It further comprises an assignment difficulty calculation section 41, an assignment order determination section 42, and an assignment difficulty recording section 43.

As shown in FIG. 2, the job storage unit 1
Information on each job arranged in the order of job execution time is stored. For example, the job number 1 stores information such as the work C in the place B, which has the job name A and requires four qualified B personnel in the morning of the first day.
The resource storage unit 2 stores information on each resource. For example, as shown in FIG. 3, information on resources stores that resource A has qualifications B and C, and that there are no restrictions on available locations A, E, F and available time as restrictions on available places. Further, the constraint information storage unit 3 stores various constraint information (restriction conditions other than those shown in FIG. 3) when allocating jobs and resources as necessary.

The operation outline of FIG. 1 is as follows. First, in the job selection device 4, the assignment difficulty calculation unit 4
1 calculates the resource allocation difficulty for each job from the information of each job in the job storage unit 1. There are various calculation methods for the assignment difficulty. For example, (a) a job with a smaller number of resources that can be assigned to the job has a higher difficulty level, and (b) a job with a greater number of constraints imposed on the job. (C) Priorities are determined in advance, and the higher the priority of the job, the higher the difficulty. The assignment difficulty recording unit 43 records the assignment difficulty of each job calculated by the assignment difficulty calculation unit 41 in the order of job execution time. FIG. 5 shows a specific example of the correspondence between jobs and assignment difficulty. On the basis of the assignment difficulty of each job recorded in the assignment difficulty recording unit 43, the assignment order determination unit 42 determines the job selection order and selects the jobs one by one. The processing in the allocation order determining unit 42 relates to the present invention, and will be described later. On the other hand, the resource selection device 5 reads the resources from the resource storage unit 2 and selects them one by one in order. The assignment determining device 6 checks whether the job and the resource selected by the assignment order determining unit 42 and the resource selecting device 5 of the job selecting device 4 do not violate the constraints stored in the constraint information storage unit 3, and determines If all match, allocation is performed, and the allocation result is stored in the allocation status storage device 7. If they do not match, the allocation determination device 6 checks the constraint again using another resource selected by the resource selection device 5.

Next, the processing in the assignment order determining unit 42 according to the present invention will be described in detail. Assignment order determination unit 42
Is based on the assignment difficulty of each job recorded in the assignment difficulty recording unit 43, and sequentially selects the job having the highest difficulty and the peripheral jobs temporally adjacent thereto. Various methods can be considered as such a job selection method. Here, the following two methods will be described as examples. In the first method, a job is assigned a certain distance (a certain range) from the job with the highest difficulty in order.
The second method is a method of selecting each job including the mountain in a group of mountains having the job as a peak in order from the job having the highest difficulty. .

FIG. 6 is a diagram for explaining the first method. An example of a selection order in this case is as follows: job 5 → job 4 → job 6 → job 3 → job 7 → job 2 → job 8 → job 18 → job 19 → job 17 → job 20 → job 16 → job 21 → Job 15 → ... This is because, in the vicinity of the first peak job 5, the jobs that are temporally adjacent to the first job 5 and alternately from the earlier job are alternately selected, and the next second peak job is selected. In the vicinity of 18, the job 18 is similarly selected alternately from the temporally adjacent jobs centered on the job 18 and from the temporally later job.

FIG. 7 is a diagram for explaining the second method.
An example of the selection order in this case is as follows: job 5 → job 4 → job 3 → job 2 → job 1 → job 6 → job 7 → job 8 → job 9 → job 10 → job 18 → job 17 → job 16 job 15 → Job 14 → Job 19 → Job 20 → Job 21 → Job 22 → Job 2
3 → ... This means that the job centered on the first peak job 5 and the next second peak job 18 is first selected from the job on the left side of the mountain in the order of closest to the peak job, and then the job on the right side is selected. Are similarly selected in the order from the closest to the peak job.

FIG. 8 is a flowchart showing the operation of the assignment order determining unit 42 in the case of the first method.

First, the job chain of the selection result (for this, a register or the like in the assignment order determining unit 42 is used)
Is emptied for initialization (step 201). next,
It is checked whether there is any unselected job in the assignment difficulty recording unit 43 (step 202). In the initial state, all jobs are in an unselected state. Therefore, in this state, one job having the highest difficulty level is selected and added to the end of the selected job chain (step 203). here,
The selected job is assumed to be A. In the first selection, this job A is at the head of the chain (in the example of FIG. 6, job 5 corresponds to this). Next, the assignment difficulty recording unit 4
The job A in No. 3 is changed from an unselected state to a selected state (step 204). Thereafter, with respect to the jobs in the assignment difficulty recording unit 43, unselected jobs within a certain range from the job A are selected in the order of time closest to the job A and added to the end of the chain. The process of changing the selected job from the unselected state to the selected state is repeatedly performed (loop of steps 205 to 208).

In this way, when there is no unselected job within a certain range from job A, the process returns to step 202, and if there is any unselected job remaining in the assignment difficulty recording unit 43, To select the job that is the most difficult to assign (the job with the greatest difficulty) and repeat the same processing.

FIG. 9 is a flowchart showing the operation of the assignment order determining unit 42 in the case of the second method.

First, the selection result job chain in the assignment order determination unit 42 is emptied for initialization (step 3).
01). Next, it is checked whether there are any unselected jobs in the assignment difficulty recording unit 43 (step 302). In the initial state, all jobs are in an unselected state. One job having the greatest difficulty is selected from the unselected jobs (step 303). Here, this selected job is assumed to be A. This job A is added to the end of the selection result job chain (step 304), and the job A in the assignment difficulty recording unit 43 is changed from an unselected state to a selected state (step 305). In the example of FIG.
Corresponds to job A.

Next, the selected job A having the highest priority
Is I and the previous job is J (step 30).
6), determine if J should be selected next. That is, if the job J is unselected and (I difficulty)-(J difficulty) is positive, the job J is selected and added to the end of the chain, and the assignment difficulty recording unit 43 Is changed to a selected state (steps 307 to 309).
Next, I = I-1, J = J-1 (step 31)
0), move the target job forward by one, and
The process is repeated from 07. Thus, step 30
By repeating the processing of steps 307 to 310 while the condition of No. 7 is satisfied, the jobs on the left side of the mountain having the peak of job A are selected in order from the closest to job A.

If the condition of step 307 is no longer satisfied, I = A, J = A + 1 (step 311),
Steps 307 to 307 for the job following job A
A process corresponding to 310 is performed. That is, the job J immediately after the job A is in the unselected state, and (the degree of difficulty of I) −
If the (J difficulty level) is positive, the job J is selected and added to the end of the chain, and the job J in the assignment difficulty level recording unit 43 is changed to the selected state (steps 312 to 312).
14). Next, assuming that I = J + 1 and J = J + 1 (step 315), the target job is moved backward by one, and the processing is repeated from step 312. By repeating the processing of steps 312 to 314 while the condition of step 312 is satisfied in this manner, the jobs on the right side of the mountain having the peak of job A are selected in order from the closest to job A.

If the condition of step 312 is no longer satisfied, the flow returns to step 302, and the assignment difficulty recording section 43
If there are any unselected jobs remaining in the list, the job that is most difficult to be assigned among the unselected jobs is selected, and the same processing is repeated.

The embodiment of the present invention has been described above. The point is that the order of allocating resources to jobs is determined by the priority of the difficulty of allocating resources to each job and the allocation of temporally adjacent jobs. It is sufficient that the method can be determined, and the method of realizing the method is not limited to the method described in the embodiment.

As described above, according to the present invention, when a plurality of jobs and a plurality of resources are allocated, a job having a high degree of difficulty and a job temporally adjacent thereto are preferentially allocated. As an effect of job priority with high priority, failure of allocation can be avoided, resource allocation to jobs in a discrete order can be avoided, and a schedule with high continuity can be created.

[Brief description of the drawings]

FIG. 1 is a system configuration diagram of an embodiment of the present invention.

FIG. 2 is a diagram illustrating a specific example of a job.

FIG. 3 is a diagram illustrating a specific example of a resource.

FIG. 4 is a processing flowchart of a conventional job / resource allocation.

FIG. 5 is a diagram illustrating an example of correspondence between a job and difficulty in resource allocation.

FIG. 6 is a diagram illustrating a first method of job selection according to the present invention.

FIG. 7 is a diagram illustrating a second method of job selection according to the present invention.

FIG. 8 is a processing flowchart corresponding to the first method of FIG. 6;

FIG. 9 is a processing flowchart corresponding to the second method in FIG. 7;

[Explanation of symbols]

 REFERENCE SIGNS LIST 1 job storage unit 2 resource storage unit 3 contract information storage unit 4 job selection device 41 assignment difficulty calculation unit 42 assignment order determination unit 43 assignment difficulty recording unit 5 resource selection device 6 assignment determination device 7 assignment status storage device

Claims (2)

(57) [Claims] 1. A job allocation optimizing apparatus for allocating a plurality of jobs and a plurality of resources according to a constraint condition using an electronic computer. A job storage unit for storing; a resource storage unit for storing information of resources to be allocated to the jobs; an assignment difficulty calculation unit for calculating a resource assignment difficulty for each job from the information in the job storage unit; the assignment difficulty calculating unit, and the assigned difficulty recording unit that records the job assignment difficulty calculated execution time order of the job, the assignment difficulty of each job recorded in the allocation difficulty recording unit analyzes, first , Most assigned from unselected jobs
Select the job with the highest difficulty, and then select the most
For jobs with high assignment difficulty, in order of time
The process of selecting a predetermined range of jobs is
Repeat until there are no more jobs
A job order allocation optimizing device, comprising: an allocation order determining unit for determining an order. 2. The job / resource allocation optimizing device according to claim 1, wherein the allocation order determining unit selects from the job having the highest assignment difficulty selected to the job successively having the lowest assignment difficulty. Determine the range of
A job / resource allocation optimizing apparatus, wherein a job having the highest degree of difficulty in allocation within the determined range is selected in the order of temporal proximity.