JPH10149347A - Resource allocation method and device therefor - Google Patents

Abstract

PROBLEM TO BE SOLVED: To make a resource allocation plan at a high speed by preparing a step where the identifier of each node is changed and given between the nodes to which the same resources cannot be allocated among those nodes corresponding to each job to have different identifiers among the nodes. SOLUTION: A graph conversion means 11 produces the graph data 3 showing the relation between jobs to which the same resources cannot be allocated based on the job data 1 and the resource data 2. A searching range change means 12 registers or changes a part where the color sorting i.e., the allocation of resources is performed as a searching range in the data 3. A graph condensing means 13 condenses the redundant part of a graph according to the characteristic of a problem. An initial color sorting means 14 sets a color at each node of the graph as an initial state. A graph color sorting means 15 changes the color of each node so as to avoid the adjacent nodes of the same color. A plan conversion means 16 converts the color sorting result of the data 3 into a resource allocation plan. A plan change means 17 manually changes a plan according to the characteristic of a problem.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a resource allocation for creating or modifying a resource allocation plan for a given work group, such as equipment, personnel, and materials, required to execute each work. The present invention relates to a method and a resource allocation device.

[0002]

2. Description of the Related Art A resource allocating apparatus is a facility and a facility necessary for executing each work for a given work group.
This is a computer system that creates or modifies the allocation plan of resources such as personnel and materials. The resource allocation plan is required to satisfy various constraints such as the number of resources and compatibility with work, and to be the best combination for maximizing or minimizing a certain objective function.

As an example of using the resource allocation planning problem,
There are problems such as crew operation plans and vehicle operation plans in railway, bus, and transportation companies, and timetable creation problems.
For example, in a vehicle operation problem in a railway company, work is operation of a train, and work data is given by a set of a departure station / start time, an end station / arrival time. The resources are vehicles owned by a railway company, and the number of vehicles is determined. The resource allocation plan determines which vehicle is in charge of each operation in order to perform the operation of each train using a limited number of vehicles. In other words, this corresponds to assigning a vehicle to be used in each operation to each operation.

[0004] Typical constraints when allocating resources (here vehicles) are:-It is not possible to allocate the same resource to tasks that have overlapping work times. This condition is not only a matter of vehicle operation planning,
In general, this is a typical constraint for resource allocation planning problems. Other constraints specific to the vehicle operation planning problem include the following:-The same resource (vehicle) is used for two operations where the previous operation (operation) ends and the next operation (operation) starts at different places. Cannot be assigned ・ Resources (vehicles) are inspected at regular intervals. During this time, the resources may not be allocated to the work that corresponds to the traffic.

It is known that a graph coloring technique can be used to create such a resource allocation plan. Graph coloring is a method of coloring each node so that adjacent nodes always have different colors on a graph expressed by nodes and lines connecting them.
In the field of operations research, there is a wide range of research examples on map color coding.

In order to solve the resource allocation planning problem by the method of graph coloring, first, a given work is regarded as a node of a graph. Then, a line is connected between nodes corresponding to tasks for which the same resource cannot be allocated. The graph created in this way is color-coded so that adjacent nodes do not have the same color.

[0007] As a method of graph coloring, both an algorithm for coloring a graph to the minimum number of colors and an algorithm for coloring a graph to a specified number of colors are known.
The number of colors corresponds to the number of resources. When the number of resources is determined in advance, color coding is performed based on the number. When it is desired to reduce the number of resources as much as possible, color coding is performed so that the number of colors is reduced as much as possible. This choice depends on the nature of the intended field of application. In a general graph coloring algorithm, first, as a first step, a color is appropriately assigned to each node as an initial solution. Next, in step 2, the node in which the constraint violation has occurred, that is, the node having many branches where the adjacent node has the same color is changed to a color as different as possible from the adjacent node. By repeating the operation of step 2, a search is made for color classification in which adjacent nodes are all different colors.

[0008] The color-coded result of the graph obtained in this way is
By reading in the form of allocating the same resource to nodes of the same color, a resource allocation plan result can be obtained.

As an existing example of a resource allocation plan using graph coloring, there is an example dealing with a test timetable creation problem (A. Salazar, RVOakford: A Graph Formulation).
of aShool Scheduling Algorithm, Communications of t
he ACM, vol17, pp696-698 (1974)). In these examples,
The subject of the examination is regarded as work, and the time limit for conducting the examination is regarded as resources. Then, a graph is created in which the constraint condition that the same subject cannot be taken in the same time period is expressed by the edge between the nodes. If this graph is color-coded by the smallest number of colors n, it can be understood that all tests can be performed in n time periods by performing the test subjects assigned the same color in the same time period.

[0010]

However, the conventional resource allocation device using graph coloring has the following problems.

First, when the structure of the graph becomes dense,
That is, the number of input tasks (number of nodes in the graph)
When the constraint conditions (the number of sides connecting the nodes) increase, the existing graph coloring method makes it very difficult to classify the colors.

In general, a graph created from work as input data to a resource allocation device and constraints on resource allocation includes a complete subgraph (a structure in which nodes of the same number as the number of colors have all connections to each other). ) Is often included. If the coloring of the complete subgraph is performed by the existing coloring algorithm, a situation occurs in which all other colors are assigned to adjacent nodes when coloring has progressed to some extent.
At this time, the procedure of selecting an appropriate color in step 2 becomes almost impossible, and eventually step 2 is repeated indefinitely, resulting in a state where coloring is not completed.

Second, the resource allocation plan cannot be optimized. In other words, the method of graph coloring is
The graph can be painted with the least number of colors or with the specified number of colors, but only one result can be obtained. Normally, there should be multiple combinations of coloring results, and the objective function desired by the user should be maximized or minimized, such as leveling the work assigned to each resource and allocating as much of the task that each resource is good at. It is necessary to search for a combination of different coloring.

For example, taking a crew operation plan as an example, an operation plan that makes the total amount of work performed by each crew as uniform as possible is required. Such optimization is a content that cannot be expressed in a graph as a constraint such as the above-mentioned “the same resource cannot be allocated to work where the work time overlaps”, and resources that directly use graph coloring This cannot be done with the assignment device.

Third, it cannot be used for "resource reassignment planning" which is a situation-responsive scheduling. A resource reallocation plan is a plan that, when a resource allocation plan created in the past is partially changed at the stage of implementation, to satisfy the given constraints and with as few changes as possible. The other part is to modify. As described above, it is not possible to cope with the contingency by modifying the plan as small as possible with the conventional resource allocation apparatus using the graph coloring.

An example of using the resource reallocation plan is "modification of vehicle operation plan". This is a problem in that, when a planned vehicle operation plan is changed due to an accident or driving trouble, the operation plan after the date of the change is reviewed. In general, instead of completely recreating the plan after the trouble occurrence date, the plan is revised so as to return to the original operation plan as soon as possible. This is mainly because of the convenience of vehicle inspection, because an inspection plan according to the number of mileage is scheduled for each vehicle in advance for a long period of time, and it is difficult to change this.

An object of the present invention is to provide a resource allocation method and a resource which can reduce the load required for graph coloring when the number of input operations and constraints increase, and can speed up the creation of a resource allocation plan. To provide an assignment device.

Still another object of the present invention is to provide a resource allocating method and a resource allocating apparatus capable of efficiently searching for a plan having a good evaluation value and realizing optimization of the resource allocating plan.

Still another object of the present invention is to provide a resource allocating method and a resource allocating apparatus capable of performing a resource reallocation plan which is a situation-responsive scheduling.

[0020]

In order to achieve the above object, according to the first and seventh aspects of the present invention, when allocating resources to be used repeatedly for a plurality of tasks, each of the tasks corresponds to each task. A first relationship is set between nodes that cannot use the same resource among the nodes, and a second relationship is set such that nodes clearly assigned to the same resource are regarded as one node. Each node after having the relationship is given a number of types of identifiers according to the number of resources, and the identifier of each node is changed so that the nodes having the first relationship have different identifiers. Has changed.

That is, according to the first and seventh aspects of the present invention,
For example, by performing graph reduction by grouping nodes corresponding to tasks for which it is known that the same resource is allocated, when the graph structure becomes dense, that is, the number of input tasks (the number of tasks in the graph) When the number of nodes and the number of constraints (the number of sides connecting nodes) increase, the load required for graph coloring can be reduced, and the creation of a resource allocation plan can be accelerated.

According to the present invention as set forth in claims 2 and 8, the identifiers of the nodes are changed so that the nodes having the first relationship have different identifiers. Then, the resource allocation plan is evaluated, a plurality of evaluations of the resource allocation plan are obtained, and a resource allocation plan having a good evaluation value is searched.

As a result, since a plurality of evaluation values can be obtained, a plan having a good evaluation value can be searched, and optimization of a resource allocation plan is realized.

According to the third and ninth aspects of the present invention, when assigning a resource to be used repeatedly to a plurality of tasks, the same resource among nodes which cannot use the same resource among the nodes corresponding to the tasks is assigned. First to have a first relationship
A second step (means) of appropriately assigning a number of types of identifiers according to the number of resources to each of the nodes after the first relationship is established; A third step (means) of changing the identifier of each node so that the nodes having the relationship of 1 have different identifiers;
A fourth step (means) for evaluating the resource allocation plan; and returning to the second step, for each node having the first relationship, the number of types corresponding to the number of resources is determined. A fifth step (means) of appropriately assigning an identifier again and re-executing the third and fourth steps.

According to the third and ninth aspects of the present invention, since a plurality of evaluation values can be obtained and the steps can be calculated efficiently, a plan having a good evaluation value is efficiently searched for, Optimization of the allocation plan can be realized.

According to a fourth aspect of the present invention, when assigning a resource to be used repeatedly to a plurality of tasks, the nodes among the nodes corresponding to the tasks cannot use the same resource. A first step (means) for providing a first relationship, and a second step of providing each node having the first relationship with a number of types of identifiers corresponding to the number of resources. (Means), a third step (means) of changing the identifier of each node so that the nodes having the first relationship have different identifiers, and a fourth step of evaluating the resource allocation plan. Step (means)
And a fifth step (means) of changing the type of the identifier for an arbitrary node among the nodes to which the identifier has been assigned in the second step, and re-executing the third and fourth steps.

According to the present invention, a plurality of evaluation values can be obtained and the calculation can be efficiently performed in the same manner as described above. The search can be performed well and the resource allocation plan can be optimized.

According to the fifth and eleventh aspects of the present invention, when assigning a resource that is repeatedly used to a plurality of tasks, the same resource cannot be used among nodes corresponding to the tasks. A node having a first relationship, a node in a range in which the plan is to be modified among the nodes having the first relationship is identified, and the number of resources is determined for each node in the range in which the plan is to be modified. , And the identifiers of the respective nodes are changed so that the nodes in the range in which the plan is to be modified have different identifiers.

According to the fifth and eleventh aspects of the present invention, for example, the above-described processing is repeatedly executed while gradually changing the range to be corrected, thereby performing a resource reallocation plan which is a situation-responsive scheduling. Can be.

According to the present invention described in claims 6 and 12, the identifiers of the respective nodes are changed so that the nodes in the range in which the plan is to be modified have different identifiers. Then, the resource allocation plan is evaluated, a plurality of evaluations of the resource allocation plan are obtained, and a resource allocation plan having a good evaluation value is searched.

As a result, a plurality of evaluation values can be obtained, so that a plan with a good evaluation value can be searched, and optimization of a resource allocation plan is realized.

It should be noted that the system has a plan display function for presenting the resource allocation plan created by the method of the present invention to the user and a plan changing means having an editing function for the user to change the resource plan allocation plan. You may. As a result, the user can freely edit the created plan, and the operation can be performed according to the convenience of the user.

Further, when the user edits the plan, the presence or absence of a condition violation is determined on the graph connecting the nodes corresponding to the tasks for which the same resources cannot be allocated by lines, and You may make it notify.

Further, the system is provided with means for automatically performing a plan correction using a procedural correction procedure or a rewrite rule with respect to a result of resource allocation using the graph coloring technique. Is also good. Thereby, the quality of the generated plan can be improved.

The present invention can also be applied to a recording medium.

That is, when assigning a resource that is repeatedly used to each of a plurality of tasks, a first relationship is established between nodes corresponding to the respective tasks and to which the same resource cannot be assigned. Providing a second relationship that regards nodes to which the same resource can be allocated as one node; and providing each node having the second relationship with a number corresponding to the number of resources. Providing a recording medium, comprising: a step of assigning an identifier of a type; and a step of changing an identifier of each node so that nodes having the first relationship have different identifiers. It is. In this case, after changing the identifier of each node so that the nodes having the first relationship have different identifiers, evaluating the resource allocation plan; And searching for a resource allocation plan with a good evaluation value.

When allocating resources to be used repeatedly to a plurality of tasks, a first step of providing a first relationship between nodes that cannot be assigned the same resource among nodes corresponding to the tasks. And the first
A second step of appropriately assigning the number of types of identifiers corresponding to the number of resources to each of the nodes having the above relationship, and different identifiers between the nodes having the first relationship. As described above, the third step of changing the identifier of each node, the fourth step of evaluating the resource allocation plan, and returning to the second step, each of the nodes after having the first relationship. And a fifth step of appropriately re-assigning a number of types of identifiers corresponding to the number of resources to re-execute the third and fourth steps. It is possible.

When allocating resources to be used repeatedly to a plurality of tasks, a first step of providing a first relationship between nodes to which the same resource cannot be assigned among nodes corresponding to the tasks. And the first
A second step of assigning the number of types of identifiers corresponding to the number of resources to each of the nodes having the relationship of the above, so that the nodes having the first relationship have different identifiers. A third step of changing the identifier of each node, a fourth step of evaluating the resource allocation plan, and changing the type of the identifier for an arbitrary node among the nodes to which the identifier has been assigned in the second step. Change the third
And a fifth step of re-executing the fourth step.

A step of assigning a first relationship between nodes which cannot be assigned the same resource among nodes corresponding to each of the tasks when assigning resources to be used repeatedly for a plurality of tasks; Identifying the nodes in the range in which the plan is to be modified among the nodes after the first relation is established; and providing, for each node in the range in which the plan is to be modified, a number of types corresponding to the number of resources It is possible to provide a recording medium characterized by comprising a step of assigning an identifier and a step of changing the identifier of each node so that the nodes in the range in which the plan is to be modified are different identifiers. . In this case, after changing the identifier of each node so that the nodes in the range in which the plan is to be corrected have different identifiers, evaluating the resource allocation plan; and determining a plurality of evaluations of the resource allocation plan. And a step of searching for a resource allocation plan having a good evaluation value.

[0040]

Embodiments of the present invention will be described below in detail with reference to the drawings.

FIG. 1 is a block diagram showing a configuration of a resource allocation planning device according to one embodiment. The device according to the present embodiment is realized, for example, on a computer, and each function of the device is realized by operating the computer according to a predetermined procedure stored as a computer program. However, it is also possible to use a dedicated device. The resource allocation planning device shown in FIG. 1 includes a graph conversion unit 11 that creates, from work data 1 and resource data 2, graph data 3 that expresses the relationship of work in which the same resource cannot be allocated, A search range changing unit 12 for registering or changing a portion where color classification, that is, resource allocation is performed as a search range, and a graph reducing unit 1 for reducing a redundant portion of a graph according to a characteristic of a problem.
3, an initial color classifying unit 14 for setting a color as an initial state for each node of the graph, a graph color classifying unit 15 for changing the color of each node so that adjacent nodes do not become the same color, and a color classification of the graph data 3. Plan conversion means 16 for converting the result into a resource allocation plan, plan change means 17 for changing the plan heuristically or manually according to the problem characteristics, and plan evaluation means 18 for evaluating the created resource allocation plan. Have.

Here, a case is described in which the present apparatus assigns a train to a train. However, the present apparatus is of course applicable to other systems.

In the present device, the work is the operation of the train,
Work data is given by a set of departure station / start time and end station / arrival time. The resources are the rolling stocks owned by the railway company, and the number of rolling stocks is determined.

The purpose of the device is to determine which trains are responsible for which operations, ie for each operation,
The assignment of vehicles to be used in the operation. As a condition for optimizing the resource allocation, the work time allocated to each vehicle, that is, the running time, is to be leveled as much as possible.

The creation of the resource allocation plan is performed according to the procedure shown in FIG.

FIG. 3 shows work data in this example. FIG. 3 shows that there are a total of 16 operations in four days.
For example, on the 21st, 21-1, 21-2, 21-3, 21-
There are four tasks: In FIG. 3, one line indicates one operation, and the drawing position of the line indicates a time period during which the operation should be performed. In addition, symbols at both ends of the line indicate a place where work is started and a place where work is finished. That is, the work 21-1 is a work that leaves A in the morning of the 21st and returns to A around noon. Work 2
1-2 is a task to depart from A after noon on the 21st and end at B at night. Operation 21-3 is an operation to depart from B on the morning of the 21st and end at A at night. The work 21-4 is a work to leave A just after noon on the 21st and return to A at night.

FIG. 4 shows resource data in this example. FIG. 4 shows that there are three rolling-stock sets # 1, # 2, and # 3.

Hereinafter, using the flow chart of FIG. 2 and the data of FIG. 3 and FIG.
The specific procedure for creating a resource allocation plan up to 4 days is shown.

First, the work data of FIG. 3 is converted into graph data (step 21). For this, first, each work in the work data of FIG. 3 is represented by a vertex. This result is shown in FIG.
Shown in That is, each piece of work data represented by a line is represented by a node (vertex). Next, two vertices (operations) that cannot use the same resource (car formation) are connected by a line. For example, in the vehicle operation plan, the "combination running at the same time" and the "combination in which the same resource cannot be used due to the departure point and the end point" cannot use the same resource. Therefore, these vertices are connected by a line. As a result, the graph of FIG. 6 is obtained. The condition of the combination of the work in which the assignment of the same resource becomes impossible as described above can be freely configured according to the characteristics of the problem using the present apparatus.

Here, a portion surrounded by a dotted line in FIG. 6 will be described. The same resources (vehicles) cannot be allocated to 21-2 and 22-1 because they are adjacent operations across days and the end location B of 21-2 and the start location Α of 22-1 are different. Since 21-2 and 21-3 operate in the same time zone, the same resource (vehicle) cannot be allocated.

Next, this graph is contracted (step 2).
2). That is, operations that are clearly assigned the same resource (vehicle) are collected into one node. For example, in a vehicle operation plan, operations that cannot be assigned to other formations correspond to this because of the relationship between the start location and the end location. For example, in operation 21-2, the end location is B, and there is only operation 22-3 that departs from B on the 22nd of the next day. Therefore, it is necessary that 21-2 and 22-3 always assign the same vehicle formation. Therefore, these two nodes are combined into one node. Thus, 21-
2 and 22-3, 22-2 and 23-3, 23-2 and 24-
FIG. 7 shows the result of summarizing No. 3. The conditions for grouping the nodes can be freely configured depending on the characteristics of the problem using the present apparatus.

Next, initial color classification is performed on each vertex of the graph data by the number of colors of the resources (vehicles) (step 24).
This method may be performed using, for example, random numbers. FIG. 8 shows the result of appropriately coloring each node in FIG. 7 with three colors of black, gray, and white.
It is.

Next, the colors of the vertices are changed so that adjacent vertices of the graph data do not have the same color (step 25). There are various known methods,
The device can be used freely without departing from the purpose of the device. For example, "An Application of the Life Span Meth
Life Spa described in od to the Graph Coloring Problem, Statistical Mathematics Institute joint report 53, “Optimization: Modeling and Algorithm 4”, 105-137, (1994).
In the n Method, V: = {V ₁ ,..., V _n } = {the entire node of the graph} color (i): = color assigned to node Vi C (i, m): = node adjacent to node Vi And the color m
The number of nodes to which is assigned. bad-degree (i): = C (i, color (i)) = node V
Number of nodes of the same color adjacent to i. LS (i): = life span of node Vi (integer variable using i as an index) LTΜ (i, m): = number of times the color of node Vi has been changed to m. When the parameter min-tabu-length: = positive integer constant max-tabu-length: = positive integer constant larger than min-tabu-length α: = positive constant, the color of k color An example of a procedure for color-coding (integers 1,..., K) is as shown in FIG.

FIG. 10 shows the result of performing the color classification according to the procedure of FIG.

Next, the colors are associated with the resources (vehicle formation) based on the color classification results in FIG. 10, and resource allocation plan (vehicle operation plan) data is created (step 25). FIG.
FIG. 1 shows a resource allocation plan (vehicle operation plan) obtained from the results of FIG.

Next, the resource allocation plan created by the above procedure is evaluated (step 26). In this example,
Since the purpose is to equalize the work time allocated to each vehicle, the variance value of the total time of the work allocated to each of the vehicle formations # 1, # 2, and # 3 from the 21st to the 24th is the evaluation value for the plan. Is calculated as In this example, a plan with a smaller evaluation value can be regarded as a more preferable resource allocation plan (vehicle operation plan). The method of calculating the evaluation value can be freely configured depending on the characteristics of the problem using the present apparatus.

In step 26, when step 23 to step 25 are repeatedly executed, the resource allocation plan having the best evaluation value is held and compared with the evaluation value of the plan created this time. If you want to keep, exchange plans.

Next, it is determined whether the creation of the resource allocation plan is completed or continued (step 27). The end is determined by repeating the procedure from step 23 to step 26 a specified number of times, when the evaluation value falls below (or above) the desired value, or when the evaluation value is improved by repetition. Judgment is made based on conditions such as when it is no longer possible. This end determination method can be freely configured depending on the characteristics of the problem using the present apparatus.

If it is determined in step 27 that the creation of the resource allocation plan is to be continued, the process returns to step 23. In step 23, the colors of the nodes in the graph data are once erased, and the initial colors are classified by random numbers. In step 24, the graph is color-coded based on the new initial color-coding results.
In general, there are a plurality of types of color coding of a graph in which all adjacent nodes are different colors, so that the step 24 produces a different result depending on the result of the initial color coding.

When it is determined in step 27 that the process has been completed, the resource allocation plan held in step 26 becomes
This is the output result of this device.

Next, another embodiment of the present invention will be described.

In this embodiment, the resource allocation plan is modified. Modification of a resource allocation plan is to modify other parts of the resource allocation plan so as to satisfy given constraints and with as few changes as possible when the plan is partially changed. is there.

In the following, correction of the vehicle assignment of a train will be described as an example. In this example, similarly to the first example, the work is the operation of the train, and the work data is given by a set of the departure station / start time and the end station / arrival time. The resources are the rolling stocks owned by the railway company, and the number of rolling stocks is determined.

The purpose of this device is to review the operation plan after the date of the change when the planned vehicle operation plan is changed due to an accident or driving trouble. The condition of the resource reallocation optimization is to modify the plan so that the operation plan returns to the original operation plan as soon as possible. The creation of the resource allocation plan is performed according to the procedure shown in FIG.

Here, as input data, it is assumed that the same data in FIGS. 3 and 4 as the first example are given as work data and resource data, respectively. FIG. 13 shows an original vehicle operation plan before correction.

In the operation plan shown in FIG. 13, it is assumed that a vehicle replacement or the like at a station on the way occurs due to an accident or the like on the 21st and a change in vehicle operation occurs. As a result, 2
Assume that at the end of the day, the arrangements are arranged differently from the schedule. This change is shown in the 21st column of FIG.

The operation result on the 21st in FIG.
Problems will occur in the operation plan after the day. For example, the formation # 1 is berthed at A on the night of the 21st, but the schedule on the 22nd is operation 22-3, and this operation is an operation starting from B. Therefore, there is a defect in this part. In order to solve such a problem, it is necessary to correct the operation plan after the 22nd.

Hereinafter, a specific procedure for correcting the vehicle operation plan after 22 days by the present apparatus will be described.

First, the work data from the 21st to the 24th in FIG. 13 is converted into graph data (step 31).
This procedure is the same as step 21 in FIG. FIG. 5 shows the result.

Next, a search range affected by the plan correction is set (step 32). This search range always includes the initial event, that is, the work in which the vehicle operation (resource allocation) is changed (the operation on the 21st in the embodiment). The search range is initially set to be narrow. Then, in the following series of steps 33 to 37, it is checked whether the convergence of the plan correction is possible only within this search range. If he fails to do so, he gradually expands his search range until he succeeds.

Here, the day on which the operation change occurred, that is, the 21st day and the 22nd day of the next day are set as the first search range, and in a series of steps from step 33 to step 37, the plan correction is performed only within this search range. It is checked whether convergence is possible, and if it fails, the range is extended one day at a time. This search range setting method can be freely configured depending on the characteristics of the problem using the present apparatus.

In the following description, a search in which the target date is set to 24 days, that is, a search as to whether or not the plan can be corrected so that the operation on 24 days is as scheduled will be considered.

Next, for operations that are clearly assigned the same resource (vehicle), the nodes are put together (step 33). This is similar to step 22 in FIG. Next, the work data of FIG. 3 is converted into graph data (step 33). This is similar to step 21 in FIG. Thereby, the graph data of FIG. 6 is obtained. Further, the nodes to which the same vehicle organization is assigned on the day when the operation change occurs and the target day are combined into one. For example, FIG.
In 4, the # 2 organization is used for the operation 21-4 on the 21st, and the # 2 organization is assigned to the operation 24-3 on the 24th in the plan. Thus, “21-3, 24-1, 24-2”
"21-4, 24-3""21-1, 21-2, 24-"
4 "can be collected into one node.
The result is shown in FIG.

Next, initial color classification is performed on each vertex of the graph data by the number of colors of the resources (vehicles) (step 35).
This method uses the original plan, FIG. By doing so, it can be expected that the operation of resolving the constraint violation will be performed only on the portion related to the place where the change was made to the original plan, and the plan modification can be minimized.

FIG. 16 shows the result of assigning colors to the graph data of FIG. 15 based on the plan of FIG.

Next, the colors of the vertices are changed so that the adjacent vertices of the graph data do not have the same color (step 36). This method is the same as step 25 in FIG. FIG. 17 shows the result of performing color coding according to the procedure of FIG.
It is.

Based on the color classification results shown in FIG. 17, colors are associated with resources (vehicle formation), and resource allocation correction plan (vehicle operation correction plan) data is created (step 36). FIG. 18 shows a resource allocation correction plan (vehicle operation correction plan) obtained from the results of FIG. In FIG. 18, the shaded portion is the portion where the original plan is changed.

Next, it is determined whether to end or continue the creation of the resource allocation correction plan (step 37). The method of determining the end is determined by whether or not the color coding is successful in step 36. If the color coding fails in step 36, the process returns to step 32, where the search range is expanded and then the graph is colored. If the color coding is successful in step 36, the resource allocation modification plan obtained in step 36 is
This is the output result of this device.

The present invention is not limited to the above-described embodiment, but can be implemented in various modifications without departing from the gist of the present invention. Examples are also included.

For example, the resource allocation plan and the resource allocation correction plan to which the present invention is applied are not limited to the train operation plan of the train, but are not limited to the crew operation plan and the vehicle operation plan in the railway, bus, and transportation companies. Alternatively, it can be applied to a computer system that creates or corrects a resource allocation plan such as equipment, personnel, and materials necessary to execute each work for a given work group, such as a timetable creation problem. is there.

For example, in the initial color classification means,
In the embodiment, initial color coding is performed using random numbers,
In the second embodiment, the original plan before correction is used for the initial color coding. However, for example, a genetic algorithm described in “Genetic algorithm: edited by Japan Fuzzy Society, Asakura Shoten” can be used. In this method, a list of nodes included in the graph data is represented as a character string individual, and the character string represents an initial color classification for each node. In the initial color classification means using a genetic algorithm, a certain number of populations are held inside the initial color classification means, and reproduction is performed so that individuals with high evaluation values calculated by the plan evaluation means can survive with high probability, and Regeneration of the population is repeated by the operation of crossover and mutation.

For example, in the initial color classification means, for a task in which a user has a request for a resource to be allocated in advance, a color is applied to a state in which the resource is allocated to a corresponding node, and the graph color classification means is used. Let's not change the color of these nodes. As a result, the user can make a part of the resource allocation plan and make the rest of the plan with the present apparatus.

Further, for example, the graph color classification means may select a color to be exchanged with a random number in step 94 in the color classification procedure shown in FIG. In this case, since a different color classification result is obtained depending on the value of the random number, the same color classification result is not always obtained from the same initial state. Therefore, another resource allocation plan can be obtained by performing step 24 after step 27 of the procedure of FIG. In FIG. 12, step 35 can be executed after step 37 to obtain another resource allocation correction plan.

Further, after the graph data is changed to the resource allocation plan or the resource allocation correction plan by the plan conversion means, the plan changing means and the procedure for changing the plan of the portion where it is apparently necessary to perform additional correction are described. May be added. This means presents a resource allocation plan to the user, and adds a display device and a plan changing device that allow the user to make manual corrections freely. Alternatively, the plan is modified using a plan modification rule set by the user in advance.

The above-mentioned plan changing means may be placed before the graph changing means changes the graph data to the resource allocation plan. In this case, for example, when the plan changing unit is provided with a display device and a plan changing device that can present the resource allocation plan to the user and allow the user to make manual corrections, the manual correction performed by the user is not Added a function to check whether adjacent nodes on the graph data have been changed to the same color on the graph data to see if they violate the conditions of resource allocation, and issue a warning to the user when necessary. can do.

[0086]

As described above, according to the present invention, when the number of input tasks (the number of nodes of the graph) and the constraint conditions (the number of sides connecting the nodes) increase, the graph is classified into different colors. The required burden can be reduced, and the creation of the resource allocation plan can be accelerated (claims 1 and 7).

Further, a plan having a good evaluation value can be searched, and optimization of a resource allocation plan is realized.
8, 12).

Furthermore, it is possible to efficiently search for a plan having a good evaluation value and to optimize a resource allocation plan.

Further, it is possible to perform a resource reallocation plan which is a situation-responsive scheduling (claims 5 and 11).

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a configuration of a resource allocation planning device according to an embodiment of the present invention.

FIG. 2 is a flowchart showing the operation of the resource allocation planning device shown in FIG.

FIG. 3 is a diagram illustrating an example of work data according to an embodiment.

FIG. 4 is a diagram illustrating an example of resource data according to an embodiment.

FIG. 5 is a diagram showing a graph according to one embodiment.

FIG. 6 is a diagram showing a graph according to one embodiment.

FIG. 7 is a diagram showing a graph according to one embodiment.

FIG. 8 is a diagram showing a graph according to one embodiment.

FIG. 9 is a flowchart illustrating an example of a procedure for performing color classification according to an embodiment.

FIG. 10 is a diagram showing a graph according to one embodiment.

FIG. 11 is a diagram showing an example of a resource allocation plan (vehicle operation plan) according to one embodiment.

FIG. 12 is a flowchart illustrating an operation of a resource allocation planning device according to another embodiment.

FIG. 13 is a diagram illustrating an example of an operation plan according to another embodiment.

FIG. 14 is a diagram illustrating an example of an operation plan according to another embodiment.

FIG. 15 is a diagram showing a graph according to another embodiment.

FIG. 16 is a diagram showing a graph according to another embodiment.

FIG. 17 is a diagram showing a graph according to another embodiment.

FIG. 18 is a diagram illustrating an example of a resource allocation plan (vehicle operation plan) according to another embodiment.

[Explanation of symbols]

 REFERENCE SIGNS LIST 1 work data 2 resource data 3 graph data 11 graph conversion means 12 search range change means 13 graph reduction means 14 initial color classification means 15 graph color classification means 16 plan conversion means 17 plan change means 18 plan evaluation means

 ──────────────────────────────────────────────────の Continuing on the front page (72) Inventor Haruhiko Toyama 70 Yanagicho, Saiwai-ku, Kawasaki-shi, Kanagawa Prefecture

Claims (12)

[Claims] 1. A method for allocating resources that are repeatedly used for a plurality of tasks, wherein a first relationship is provided between nodes corresponding to the tasks, to which nodes the same resource cannot be assigned. And a step of providing a second relationship that regards nodes to which the same resource can be allocated as one node. The method according to the number of resources for each node after the second relationship is provided. A resource allocation method, comprising: assigning a number of types of identifiers; and changing an identifier of each node so that nodes having the first relationship have different identifiers. A step of evaluating the resource allocation plan after changing the identifier of each node so that the nodes having the first relationship have different identifiers; 2. The resource allocation method according to claim 1, further comprising a step of searching for a plurality of resource allocation plans having a good evaluation value. 3. A method for allocating resources that are repeatedly used for a plurality of tasks, wherein a node having the same resource among nodes corresponding to the tasks has a first relationship. A first step, a second step of appropriately assigning a number of types of identifiers according to the number of resources to each of the nodes after the first relation is provided, A third step of changing the identifier of each node so that different nodes have different identifiers; a fourth step of evaluating the resource allocation plan; and returning to the second step, A fifth step of appropriately re-assigning a number of types of identifiers corresponding to the number of resources to each of the nodes having the relationship, and causing the third and fourth steps to be re-executed. Resources characterized by Ri applied method. 4. A method of allocating resources that are repeatedly used for a plurality of tasks, wherein a first relationship is provided between nodes corresponding to the tasks and to which the same resource cannot be assigned. A first step, a second step of giving each node after the first relation the number of types of identifiers corresponding to the number of resources, and the first relation. A third step of changing the identifier of each node so that the nodes have different identifiers; a fourth step of evaluating the resource allocation plan; and a step of evaluating the node to which the identifier is assigned in the second step. And a fifth step of changing the type of the identifier for an arbitrary node and re-executing the third and fourth steps. 5. A method for allocating resources that are repeatedly used to a plurality of tasks, wherein a first relationship is provided between nodes corresponding to the tasks, to which nodes the same resource cannot be assigned. A step of specifying a node in a range in which the plan is to be corrected among the nodes after the first relation is provided; and for each node in the range in which the plan is to be corrected, A resource allocation method comprising: assigning a number of types of identifiers; and changing an identifier of each node so that nodes in a range in which the plan is to be modified have different identifiers. 6. A step of evaluating the resource allocation plan after changing the identifier of each node so that the nodes in the range in which the plan is to be modified have different identifiers. 6. The resource allocation method according to claim 5, further comprising the step of searching for a plurality of resource allocation plans having good evaluation values. 7. An apparatus for allocating resources to be used repeatedly for a plurality of tasks, wherein a node having the same resource among nodes corresponding to the tasks has a first relationship. Means, means for giving a second relationship between nodes to which the same resource can be assigned as one node, and each node having the second relationship has a number corresponding to the number of resources. A resource allocation apparatus comprising: means for assigning a number of types of identifiers; and means for changing the identifier of each node so that the nodes having the first relationship have different identifiers. 8. A means for evaluating the resource allocation plan after changing the identifier of each node so that the nodes having the first relationship have different identifiers, and evaluating the resource allocation plan. 8. The resource allocation apparatus according to claim 7, further comprising: means for searching for a plurality of resource allocation plans having good evaluation values. 9. An apparatus for allocating resources that are repeatedly used for a plurality of tasks, wherein a node having the same resource among nodes corresponding to the tasks has a first relationship. A first means, a second means for appropriately assigning a number of types of identifiers according to the number of resources to each of the nodes having the first relation, and the first relation Third means for changing the identifier of each node so that different nodes have different identifiers; fourth means for evaluating the resource allocation plan; and returning to the second means, the first means To each node after having the relationship, a number of types of identifiers corresponding to the number of resources are appropriately re-assigned, and the fifth and the fourth means are re-executed.
And a resource allocation device. 10. An apparatus for allocating resources that are repeatedly used for a plurality of tasks, wherein a node having the same resource among nodes corresponding to the tasks has a first relationship. A first means, a second means for assigning a number of types of identifiers according to the number of resources to each node after the first relation is provided, and the first relation is provided. A third means for changing the identifier of each node so that the nodes have different identifiers; a fourth means for evaluating the resource allocation plan; and a node for which the identifier is assigned in the second means. And a fifth means for changing the type of the identifier for an arbitrary node and re-executing the third and fourth means. 11. An apparatus for allocating resources to be repeatedly used for a plurality of tasks, wherein a node having the same resource among nodes corresponding to the tasks has a first relationship. Means for specifying a node in a range in which the plan is to be corrected among the nodes having the first relationship, and for each node in the range in which the plan is to be corrected, according to the number of resources. A resource allocation apparatus, comprising: means for assigning a number of types of identifiers; and means for changing the identifier of each node so that the nodes in the range in which the plan is to be modified are different identifiers. 12. A means for evaluating the resource allocation plan after changing the identifier of each node so that the nodes in the range in which the plan is to be modified have different identifiers. 12. The resource allocation apparatus according to claim 11, further comprising: means for searching for a plurality of resource allocation plans having a good evaluation value.