Placement optimization device, placement optimization method, and placement optimization program

Abstract

A placement optimization device according to one embodiment of the present invention comprises: an acquisition unit that acquires spot information including a network topology indicating node spots and depot spots having recovery resources standing by for when there is a failure at a node, a network parameter, and the latitude and longitude of the node spots and the depot spots; a movement time matrix calculation unit that calculates, on the basis of the spot information, a movement time matrix indicating the movement time between the node spots and the depot spots; a formulation unit that formulates a mixed integer programming problem discretely expressing the placement of the recovery resources, on the basis of the network topology, the network parameter, and the movement time matrix; an optimization unit that optimizes the mixed integer programming problem by means of a heuristic method using the augmented Lagrangian function method, and determines optimal placement information indicating the optimal placement of the recovery resources; and an output control unit that outputs the optimal placement information.

Description

[Technical Field]

[0001] This invention relates to a layout optimization device, a layout optimization method, and a layout optimization program. [Background technology]

[0002] Traditionally, network administrators dealt with network failures by determining the recovery order of the affected areas using servers and other devices, and then dispatching recovery personnel to the affected areas according to that recovery order.

[0003] For example, Patent Document 1 discloses a technique for creating an optimal deployment plan for recovery personnel, taking into account the travel time required for recovery personnel to reach the fault location and the communication speed in the network. For example, Non-Patent Document 1 discloses a technique for determining the optimal placement of SDN controllers using gradient descent and genetic algorithms. Furthermore, for example, Non-Patent Document 2 discloses a technique for determining the optimal recovery sequence in the event of a large-scale failure by sequential linear relaxation. [Prior art documents] [Patent Documents]

[0004] [Patent Document 1] International Publication No. 2022 / 168328 [Non-patent literature]

[0005] [Non-Patent Document 1] V. Huang, G. Chen, Q. Fu and E. Wen, "Optimizing Controller Placement for Software-Defined Networks," 2019 IFIP / IEEE Symposium on Integrated Network and Service Management (IM), 2019, pp. 224-232. [Non-Patent Document 2] D. Zad Tootaghaj, N. Bartolini, H. Khamfroush and T. La Porta, "On Progressive Network Recovery From Massive Failures Under Uncertainty," in IEEE Transactions on Network and Service Management, vol. 16, no. 1, pp. 113-126, March 2019, doi: 10.1109 / TNSM.2018.2880155. [Overview of the Initiative] [Problems that the invention aims to solve]

[0006] However, Patent Document 1 is formulated as a nonlinear program, which has the problem of rounding errors in the placement variables. Furthermore, Patent Document 1 calculates the impact using virtual traffic volume on the network and does not consider probabilistic evaluation. Non-Patent Document 1 uses a penalty function method with fixed coefficients, which has the problem of slow convergence and unstable solutions. Also, Non-Patent Document 2 calculates the impact using virtual traffic volume on the network and does not consider probabilistic evaluation. Although it defines the optimal placement problem for SDN controllers, it has the problem of not being applicable to failure occurrence and the behavior of recovery personnel.

[0007] This invention was made in view of the above circumstances, and its purpose is to provide a technology that can formulate fault response more precisely by discretely representing the allocation of recovery resources and formulating it into a mixed integer program. [Means for solving the problem]

[0008] To solve the above problems, an arrangement optimization device according to one aspect of the present invention comprises: a network topology indicating the locations of nodes and depot locations where recovery resources wait in the event of a failure at a node; network parameters; an acquisition unit that acquires location information including the latitude and longitude of the node locations and the depot locations; a travel time matrix calculation unit that calculates a travel time matrix indicating the travel time between the node locations and the depot locations based on the location information; a formulation unit that formulates a mixed-integer programming problem that discretely represents the arrangement of the recovery resources based on the network topology, network parameters and the travel time matrix; an optimization unit that optimizes the mixed-integer programming problem using a heuristic method with an extended Lagrangian function method to determine optimal arrangement information indicating the optimal arrangement of the recovery resources; and an output control unit that outputs the optimal arrangement information. [Effects of the Invention]

[0009] According to one aspect of this invention, it becomes possible to create an optimal deployment plan for recovery personnel, taking into account the travel time required for recovery personnel to reach the affected area and the communication speed in the network. [Brief explanation of the drawing]

[0010] [Figure 1] Figure 1 shows an example of a network topology. [Figure 2] Figure 2 shows an example of the behavior of recovery resources from the occurrence of a failure to the recovery of the failure. [Figure 3] Figure 3 shows an example of the behavior of recovery resources when allocation probability is introduced. [Figure 4] Figure 4 shows an example of an overview for calculating the expected value of the impact. [Figure 5] Figure 5 shows the binary representation of the xvk configuration and an example of a specific configuration corresponding to that binary representation. [Figure 6] Figure 6 shows an example of travel time between the depot and the communications building. [Figure 7] Figure 7 shows an example of the failure rate and the response time from the occurrence of a failure until the resolution is completed. [Figure 8] Figure 8 shows an example of the recovery time and the conditions under which a failure occurs in a communications building. [Figure 9] Figure 9 shows an example of a method for calculating the expected value of the impact. [Figure 10] Figure 10 shows an example of an overview for optimizing a planning problem. [Figure 11] Figure 11 shows an example of optimizing placement using a genetic algorithm. [Figure 12] Figure 12 is a block diagram showing an example of the hardware configuration of a placement optimization device according to an embodiment. [Figure 13] Figure 13 is a block diagram showing the software configuration of the placement optimization device in the embodiment, in relation to the hardware configuration shown in Figure 12. [Figure 14] Figure 14 is a flowchart illustrating an example of the general operation of a placement optimization device to obtain the optimal placement. [Figure 15] Figure 15 shows an example of the network topology used for comparison. [Figure 16] Figure 16 shows the results of the simulation. [Modes for carrying out the invention]

[0011] Embodiments of this invention will be described below with reference to the drawings. [composition] Figure 1 shows an example of a network topology. In the network topology shown in Figure 1, communication buildings n, recovery resources v, and depots k are arranged. Here, n, v, and k are variables that are arbitrary integers, including 0. For example, if N communication buildings (nodes) are arranged on the network topology, then n can take any value from 0 to N-1. Similarly, if V recovery resources are arranged on the network topology, then v can take any value from 0 to V-1, and if K depots are arranged, then k can take any value from 0 to K-1. A communication building is a node in a network topology that provides communication services. Failures in the network topology occur at the communication building level; for example, if a failure occurs in communication building n, it will affect communications passing through communication building n.

[0012] Recovery resources are the resources actually used to restore the system, such as power generators and workers, while depot k is the location where these recovery resources are stationed. For example, if a failure occurs in communication building n, recovery resources v, which were stationed at depot k, are dispatched to restore the system. Once the failure is resolved, communication in communication building n is restored. Note that there may be cases where recovery resources v are not stationed at depot k.

[0013] (Modeling of disaster recovery) The following describes in detail the modeling process in which, when a failure occurs in communication building n, a recovery resource v waiting at depot k travels to the failed communication building n and restores it.

[0014] Figure 2 shows an example of the behavior of recovery resource v from the occurrence of a failure to its recovery. As shown in Figure 2, the recovery resource v simplifies its behavior as follows. As shown in Figure 2(1), the recovery resource v detects that a failure has occurred in communication building n. Here, it is assumed that the failure in communication building n is notified to the recovery resource v waiting at depot k without delay. As shown in Figure 2(2), the recovery resource v moves from depot k to communication building n. Details such as the travel time will be described later. As shown in Figure 2(3), the recovery resource v performs recovery work at the communication building n where the failure occurred. Details such as the time required for the recovery work will be described later. As shown in Figure 2(4), after the recovery work is completed, the recovery resource v returns to depot k.

[0015] In the example shown in Figure 2, the recovery resource v is assumed to travel back and forth between depot k and the communication building n where the failure occurred after the recovery work is completed. That is, the recovery resource v will not go from the communication building n where the failure occurred and the recovery work has been completed to another communication building n+1 where a different failure has occurred, nor will it return to another depot k+1, etc.

[0016] Figure 3 shows an example of the behavior of a recovery resource v with allocation probability introduced. As shown in Figure 3, in this embodiment, in order to make the problem easier to treat as an optimization problem, an allocation probability is introduced when the recovery resource v goes to a communication building. For example, in Figure 3, the allocation probability to the first communication building is shown as 0.8, and the allocation probability to the second communication building is shown as 0.2. Here, the allocation probability can be calculated using the same method as in Non-Patent Document 1, so a detailed explanation is omitted here.

[0017] Furthermore, the failure rate in communication building n is λ n For example, in Figure 3, the failure rate λ of communication building n is n =0.1 items / hour, and λ n = 0.03 cases / hour. In this embodiment, the failure rate λ in each communication building n in the network topology. n This document describes methods for minimizing (optimizing) the impact on communication services (data traffic) in a model that takes into account fault response.

[0018] (Formulation of failure response) Next, the formulation of failure response will be described in detail.

[0019] Figure 4 is a diagram showing an example of an outline for calculating the expected value of the impact. First, the state where the recovery resource v is arranged in the depot k is expressed as the arrangement x vk and the allocation probability of the recovery resource v to the communication building n is expressed as the probability p vn .

[0020] As shown in (1) of Figure 4, the operating rate π 0n of the communication building is calculated from the failure occurrence rate λ n of each communication building n, the arrangement x vk of the recovery resource v, and the allocation probability p vn . Here, the details of the calculation method of the operating rate π 0n will be described later.

[0021] As shown in (2) of Figure 4, the reduction amount of traffic volume (impact a i ) for each combination of failure occurrence locations is calculated. The upper row in (2) of Figure 4 represents the probability p i when a failure occurs in the communication building n and the communication building m is operating normally, the middle row represents the probability p i when the communication building n is operating normally and a failure occurs in the communication building m, and the lower row represents the probability p i when failures occur in both the communication building n and the communication building m. The detailed calculation method of each probability p i and the calculation method of the impact a i will be described later. The impact a i is calculated based on this probability p i . And the expected value E[a] of the impact is calculated based on the impact a i .

[0022] Next, the expression of the arrangement x vk will be described. In this embodiment, the arrangement x vk is represented by a binary variable x vkIt is represented as ∈{0,1}. Therefore, the location x of the recovery resource v. vk This can be expressed as follows:

[0023]

number

[0024] Figure 5 shows arrangement x vk This figure shows the binary representation and an example of a specific arrangement corresponding to that binary representation.

[0025] Figure 5 shows an example where three recovery resources (Recovery Resource 0, Recovery Resource 1, and Recovery Resource 2) are placed in four depots (Depot 0, Depot 1, Depot 2, and Depot 3). In the example in Figure 5, Recovery Resource 1 and Recovery Resource 2 are placed in Depot 0, and Recovery Resource 0 is placed in Depot 1. Recovery resources are not placed in Depots 1 and 2. When the above arrangement is represented using binary variables, it can be shown as the matrix in Figure 5.

[0026] Thus, in this embodiment, the placement x of the recovery resource v vk By discretizing it using a binary representation, it is formulated as a mixed-integer programming problem.

[0027] Next, we will describe the travel time of the recovery resource v located at depot k to the communication building n where the failure occurred.

[0028] Figure 6 shows the travel time d between depot k and communication building n. kn This is a diagram illustrating an example. As shown in Figure 6, the time required to travel from depot k to communication building n is the travel time d. kn This is expressed as follows: In this case, arrangement x vk The travel time d required for recovery resource v to move to communication building n. kn This can be expressed as follows:

[0029]

number

[0030] Next, the failure rate λ of all communication buildings that recovery resource v is responsible for v We will also explain the time it takes when a failure occurs.

[0031] Figure 7 shows the failure rate λ. v This diagram also shows an example of the response time from the occurrence of a problem until the resolution is completed. The failure rate λ is calculated as the failure rate for all communication buildings that recovery resource v is responsible for. v Expressed as, the failure rate λ v The assignment probability p vn Weighted failure rate λ n It is calculated from the failure rate λ. v This can be expressed as follows:

[0032]

number

[0033] As shown in Figure 7(b), the time required for the recovery resource v to address the failure is expressed as the work time r. In this case, the time required for one failure (failure to communication building n) is the travel time d for the recovery resource v located at depot k to move to communication building n. vn The work time r for recovery resource v and the travel time d for recovery resource v to move from communication building n to depot k are also included. vn The sum of (i.e., 2d vn It can be expressed as +r). And the allocation probability p to the recovery resource v. vn Considering this, the time required for recovery resource v to respond to the failure is response time μ. v -1 Expressed as such, the response time μ v -1 This can be expressed as follows:

[0034]

number

[0035] Also, the waiting time in the queue W v This can be represented by the M / M / 1 queue as follows:

[0036]

number

[0037] Next, the utilization rate π of the communications building n. 0n This section will explain the details. The recovery time S is the time it takes for communication building n to recover after a failure occurs. n This is how it is expressed.

[0038] Figure 8 shows the recovery time S. n This diagram shows an example of the conditions under which a failure occurs in communication building n. As shown in Figure 8(a), recovery time S n Considering the allocation probability of recovery resource v, the recovery time S n This can be expressed as follows:

[0039]

number

[0040] As shown in the above equation, recovery time S n Waiting time W v And travel time d vn This is expressed as the sum of the time and the work time r. Also, as shown in Figure 8(b), if a failure occurs in communication building n, communication building n will not relay communications while the failure is occurring.

[0041] Here, the utilization rate π 0n This is the probability that no failure has occurred in communication building n. Therefore, the availability rate π 0n The recovery time is S n and failure rate λ n It can be expressed as follows using .

[0042]

number

[0043] In other words, the utilization rate π 0n This is defined as the probability that there are no problems in the queue, and on average represents the proportion of communication buildings n that are operating normally.

[0044] Next, we will explain in detail how to evaluate the reliability of communication services. Figure 9 shows an example of a method for calculating the expected value of the impact, E[a]. As shown in Figure 9(a), if no failure occurs (availability π 0n Compared to (1) (left side of Figure 9(a)), the decrease in the total bandwidth o that each communication building n can communicate with at the POI (Point of Interface) is Δo, and the impact a i Let's assume that.

[0045] Impact level a i The (decrease amount Δo) can be calculated in advance based on the combination of fault locations, as shown in Figures 9(b) to (d). For example, as shown in Figure 9(b), the operational rate π of communication building n. 0n and the operating rate of communication building m π 0m Let's assume that this is the case.

[0046] In this case, as shown in the upper panels of Figures 9(b) and (c), the probability p is that a failure occurs in communication building m, while communication building m is operating normally. i is, p i =(1-π 0n )π 0m As shown in the middle sections of Figures 9(b) and (c), the probability p when communication building n is operating normally and communication building m is experiencing a failure is as follows. i is, p i =π 0n (1-π 0m Similarly, as shown in the lower panels of Figures 9(b) and (c), the probability p when there is a failure in communication building n and communication building m is as follows. i is, p i =(1-π0n )(1-π 0m ) means that the probability p i is, in terms of impact a i This represents the probability of that happening.

[0047] Pre-calculated impact a i and its probability p i Based on this, the expected value of the influence E[a] is calculated as follows.

[0048]

number

[0049] Furthermore, impact level a i The pattern in which it saturates is an influence of a i This eliminates the need to calculate the impact a. i Since it does not affect anything, the calculation can be omitted.

[0050] (Optimization of planning problems) Finally, the details of the optimization method for the planning problem will be explained. The planning problem in this embodiment is mixed-integer and nonlinear. Therefore, existing solvers cannot solve this problem. Thus, in this embodiment, a metaheuristic method is applied. For example, in this embodiment, similar to Non-Patent Document 1, optimization is performed by combining a genetic algorithm (GA) and a gradient method.

[0051] Figure 10 shows an example of an overview for optimizing a planning problem. As shown in Figure 10, first, the placement of various recovery resources v x vk Generate the following. Assume that depot k is given. Then, determine the location x of the recovery resource v. vkSearch with the genetic algorithm GA. Next, the allocation probability p of the restoration resource v to the communication building n is determined so as to minimize the expected value E[a] of the influence degree. vn and the flow f mn is optimized by the gradient method. The flow is a non-negative real value set on each edge of the topology of the communication building that virtually represents the communication volume between communication buildings, and is obtained by the gradient method described above. For example, the placement x of integer-valued restoration resources vk is fixed and optimized by the gradient method. The influence degree a of the optimized placement x of the restoration resources vk is calculated, and the placement x of the restoration resources of the next generation is generated based on the calculated influence degree a i The specific method for generating the placement x of the restoration resources of the next generation will be described later. i vk vk

[0052] Here, when optimizing the allocation probability p of the restoration resources vn and the flow f mn by the gradient method, the extended Lagrange method that can obtain a solution more stably and quickly without using the penalty method may be used.

[0053] FIG. 11 is a diagram showing an example of optimizing the placement by the genetic algorithm. First, the placement x of the restoration resources is represented by genes. The gene expression may be any expression as long as it is, for example, an array of the same size as the number of restoration resources and can hold the number of the depot to which it is to be placed. vk

[0054] Then, the placement of the restoration resources expressed by genes is optimized using the gradient method as described with reference to FIG. 10. Then, the influence degree a in the optimized placement x of the restoration resources vk is calculated. Then, genetic operations are performed to generate the placement x of the restoration resources of the next generation i Here, the genetic operations include mutation, uniform crossover, elitist selection, etc. vk [[ID=4!]]

[0055] As described above, the allocation probability p of the restoration resources​​​​​vn and flow f mn By optimizing this using the gradient method, we can determine the placement with the least impact.

[0056] (composition) Next, we will describe the hardware and software configurations of the placement optimization device for implementing the method described above. Figure 12 is a block diagram showing an example of the hardware configuration of the placement optimization device 1 according to this embodiment. The placement optimization device 1 is a computer that analyzes input data, generates output data, and outputs it. For example, the placement optimization device 1 is installed in any location set by the administrator who manages the placement optimization device 1.

[0057] As shown in Figure 12, the placement optimization device 1 comprises a control unit 10, a program storage unit 20, a data storage unit 30, a communication interface 40, and an input / output interface 50. The control unit 10, the program storage unit 20, the data storage unit 30, the communication interface 40, and the input / output interface 50 are connected to each other via a bus so as to be able to communicate with each other. Furthermore, the communication interface 40 may be connected to an external device so as to be able to communicate with it via a network. In addition, the input / output interface 50 is connected to an input device 2 and an output device 3 so as to be able to communicate with it.

[0058] The control unit 10 controls the placement optimization device 1. The control unit 10 includes a hardware processor such as a central processing unit (CPU). For example, the control unit 10 may be an integrated circuit capable of executing various programs.

[0059] The program storage unit 20 can use a combination of non-volatile memory that allows writing and reading at any time, such as EPROM (Erasable Programmable Read Only Memory), HDD (Hard Disk Drive), and SSD (Solid State Drive), as a storage medium, and non-volatile memory such as ROM (Read Only Memory). The program storage unit 20 stores programs necessary to execute various processes. In other words, the control unit 10 can realize various controls and operations by reading and executing programs stored in the program storage unit 20.

[0060] The data storage unit 30 is a storage device that uses a combination of non-volatile memory, such as an HDD or memory card, which allows for writing and reading at any time, and volatile memory, such as RAM (Random Access Memory), as storage media. The data storage unit 30 is used to store data acquired and generated during the process in which the control unit 10 executes a program and performs various processing.

[0061] The communication interface 40 includes one or more wired or wireless communication modules. For example, the communication interface 40 includes a communication module that connects to an external device via a network, either wired or wirelessly. The communication interface 40 may also include a wireless communication module that connects to an external device wirelessly, such as a Wi-Fi access point and a base station. Furthermore, the communication interface 40 may include a wireless communication module that connects to an external device wirelessly using short-range wireless technology. In other words, the communication interface 40 can be any general communication interface that can communicate with an external device and send and receive various types of information under the control of the control unit 10.

[0062] The input / output interface 50 is connected to the input device 2 and the output device 3, etc. The input / output interface 50 is an interface that enables the transmission and reception of information between the input device 2 and the output device 3. The input / output interface 50 may be integrated with the communication interface 40. For example, the placement optimization device 1 and at least one of the input device 2 and the output device 3 are wirelessly connected using short-range wireless technology, and information may be transmitted and received using said short-range wireless technology.

[0063] The input device 2 may include, for example, a keyboard or pointing device for the user to input various information to the placement optimization device 1. The input device 2 may also include a reader for reading data to be stored in the program storage unit 20 or data storage unit 30 from a memory medium such as a USB memory, or a disk device for reading such data from a disk medium.

[0064] Output device 3 includes a display, etc., for displaying the results calculated by the control unit 10. Output device 3 also includes a printer, etc., for printing the information displayed on the display.

[0065] Figure 13 is a block diagram showing the software configuration of the placement optimization device 1 in the embodiment, in relation to the hardware configuration shown in Figure 12. The control unit 10 comprises an acquisition unit 101, a movement time matrix calculation unit 102, a planning problem formulation unit 103, an optimization unit 104, and an output control unit 105.

[0066] The acquisition unit 101 includes a network topology acquisition unit 1011, a parameter acquisition unit 1012, and a location information acquisition unit 1013.

[0067] The network topology acquisition unit 1011 acquires the network topology. The network topology acquisition unit 1011 acquires the network topology either from input by the administrator or stored in the data storage unit 30. Here, the network topology includes information such as the place names where communication buildings are located, how the communication buildings are connected to each other, the place names where depots are located, and the number of recovery resources.

[0068] The parameter acquisition unit 1012 acquires network parameters. The parameter acquisition unit 1012 acquires network parameters either by input by the administrator or stored in the data storage unit 30. Here, the network parameters include the communication bandwidth required by each city, the upper limit of the amount of communication passing through the city, the failure rate that occurs in communication buildings, the allocation of recovery resources, the probability of allocating recovery time, etc.

[0069] The location information acquisition unit 1013 acquires location information. For example, the location information acquisition unit 1013 uses a map API or the like to obtain the latitude and longitude of communication buildings and depots from the received network topology. Then, it acquires information about the latitude and longitude of these communication buildings and depots as location information.

[0070] The travel time matrix calculation unit 102 calculates a travel time matrix. The travel time matrix calculation unit 102 calculates the travel time between the communication building and the depot in matrix form from the location information stored in the acquired data storage unit 301. For example, the travel time matrix calculation unit 102 calculates the travel distance between the communication building and the depot using a map API or the like, and calculates the travel time between the communication building and the depot using the calculated distance and a predetermined average speed.

[0071] The planning problem formulation unit 103 formulates the planning problem. The planning problem formulation unit 103 retrieves the network topology and network parameters stored in the acquired data storage unit 301. Based on the network topology, network parameters, and travel time matrix, the planning problem formulation unit 103 formulates a mixed integer programming problem, as described above, that is, a discrete representation of recovery resources.

[0072] The optimization unit 104 optimizes the planning problem. The optimization unit 104 optimizes the planning problem using the optimization method described above. Specifically, the optimization unit 104 optimizes using a heuristic method with an extended Lagrangian function method and determines optimal placement information that indicates the optimal placement of recovery resources.

[0073] The output control unit 105 outputs the optimized layout to the output device 3. The output control unit 105 controls the display of the output device 3 to show the optimized layout.

[0074] The acquired data storage unit 301 is used to store the data acquired by the acquisition unit 101 (network topology, network parameters, location information, etc.).

[0075] [Operation] Figure 14 is a flowchart showing an example of the general operation of the placement optimization device 1 to obtain the optimal placement. The operation of this flowchart is realized when the control unit 10 of the placement optimization device 1 reads and executes the program stored in the program storage unit 20.

[0076] This operation is initiated by a predetermined input from the administrator (user) of the placement optimization device 1, indicating that they want to know the optimized placement.

[0077] In step ST101, the network topology acquisition unit 1011 of the acquisition unit 101 acquires the network topology. The network topology acquisition unit 1011 acquires the network topology either from input by the administrator or stored in the data storage unit 30. The network topology acquisition unit 1011 outputs the acquired network topology to the location information acquisition unit 1013 and then stores it in the acquired data storage unit 301. Here, the network topology includes information such as the place names where communication buildings are located, how the communication buildings are connected to each other, the place names where depots are located, and the number of recovery resources.

[0078] In step ST102, the parameter acquisition unit 1012 acquires network parameters. The parameter acquisition unit 1012 acquires network parameters either entered by the administrator or stored in the data storage unit 30. The parameter acquisition unit 1012 stores the acquired network parameters in the acquired data storage unit 301. Here, the network parameters include the communication bandwidth required by each city, the upper limit of the amount of communication passing through the city, the failure rate that occurs in communication buildings, and the working time.

[0079] In step ST103, the location information acquisition unit 1013 acquires location information. For example, the location information acquisition unit 1013 uses a map API or the like to acquire the latitude and longitude of communication buildings and depots on the received network topology. Then, it acquires information about the latitude and longitude of these communication buildings and depots as location information. The location information acquisition unit 1013 stores the acquired location information in the acquired data storage unit 301.

[0080] In step ST104, the travel time matrix calculation unit 102 calculates a travel time matrix. The travel time matrix calculation unit 102 calculates the travel time between the communication building and the depot in matrix form from the location information stored in the acquired data storage unit 301. For example, the travel time matrix calculation unit 102 calculates the travel distance between the communication building and the depot using a map API or the like, and calculates the travel time between the communication building and the depot using the calculated distance and a predetermined average speed. Then, the travel time matrix calculation unit 102 outputs the travel time matrix to the planning problem formulation unit 103.

[0081] In step ST105, the planning problem formulation unit 103 formulates the planning problem. The planning problem formulation unit 103 acquires the network topology and network parameters stored in the acquired data storage unit 301. Based on the network topology, network parameters, and travel time matrix, the planning problem formulation unit 103 formulates the planning problem as described above. The planning problem formulation unit 103 randomly determines the initial values ​​of the recovery resource allocation and recovery time allocation probability when the calculation starts. Then, the planning problem formulation unit 103 outputs the planning problem to the optimization unit 104.

[0082] In step ST106, the optimization unit 104 optimizes the planning problem. The optimization unit 104 optimizes the planning problem using the optimization method described above. That is, the optimization unit 104 determines the optimized placement of recovery resources. The optimization unit 104 outputs information about the optimized placement to the output control unit 105.

[0083] As mentioned above, steps ST105 and ST106 may be repeated a predetermined number of times for the placement of genetically modified recovery resources from the optimized placement.

[0084] In step ST107, the output control unit 105 outputs the optimized layout to the output device 3. The output control unit 105 controls the display of the output device 3 to show the optimized layout.

[0085] As described above, the placement optimization device 1 can determine the optimal placement of recovery resources that have the least impact on a given network topology.

[0086] [Comparison of impact levels] Next, we will describe an example of comparing the influence of the optimized placement achieved by the placement optimization device 1 described above with other methods. In this embodiment, we compare this embodiment with a random method and the Facility Layout Problem (FLP).

[0087] The random method randomly allocates recovery resources. The condition is that the allocation probability is set according to the reciprocal of the distance from the recovery resource to each communication building.

[0088] In the Facility Placement Problem (FLP), placement is optimized using a mixed-integer linear programming approach. The condition is to minimize the sum of the distances to the recovery resources and the communication buildings. For example, this can be weighted by the communication bandwidth required and the allocation probability of the communication buildings. Furthermore, the number of communication buildings each recovery resource is responsible for is limited to equalize the utilization rate of the recovery resources. This condition can be expressed as follows:

[0089]

number

[0090] Next, I will explain the settings used for the comparison. Figure 15 shows an example of the network topology used for comparison. The network topology shown in Figure 15 is part of the AT&T network in the Internet Topology Zoo. This network topology is an IP / MPLS backbone connecting major cities in the United States.

[0091] Furthermore, the number of nodes (communication buildings) is assumed to be 10, and the number of recovery resources is assumed to be 5. Depots are located in each city, and the travel time between the same cities is assumed to be 0.

[0092] Furthermore, the communication bandwidth required by each city is set at 1 Gbps, and the upper limit for the amount of data traffic passing through a city is set at 3 Gbps.

[0093] As described above, the travel time matrix calculation unit 102 calculates the travel time of vehicles between cities using a map API.

[0094] The simulation assumes that 0.1 outages occur per hour across all communication buildings combined. The occurrence rate is determined by randomly assigning outages to each communication building and using 10 different seeds for the simulation.

[0095] Figure 16 shows the results of the simulation. As shown in Figure 16, the random method has an impact of 7.98 Gbps and a standard deviation of 2.8, the FLP method has an impact of 1.48 Gbps and a standard deviation of 0.348, and the method according to this embodiment has an impact of 1.11 Gbps and a standard deviation of 0.0833. In other words, the results in Figure 16 show that this embodiment has the smallest impact compared to the other two, and the small standard deviation also indicates that it consistently has the smallest impact.

[0096] [Effects and Effects] According to this embodiment, the placement optimization device 1 can reduce the impact on communication services by determining the optimal placement of recovery resources, taking into account the effects of probabilistically occurring network failures.

[0097] [Other embodiments] In the above embodiment, arrangement x vk An example of optimizing using the fixed gradient method has been explained, but optimization can be performed using any gradient method or other methods.

[0098] Furthermore, the method described in the above embodiment can be distributed by storing the program (software means) that can be executed by a computer in a storage medium such as a magnetic disk (floppy disk, hard disk, etc.), optical disk (CD-ROM, DVD, MO, etc.), or semiconductor memory (ROM, RAM, flash memory, etc.), and by transmitting it via a communication medium. The program stored on the medium also includes a configuration program that configures the software means (including not only the execution program but also tables and data structures) to be executed by the computer. The computer realizing this device reads the program stored in the storage medium and, if necessary, constructs the software means using the configuration program, and executes the above-mentioned processing by controlling the operation of this software means. The storage medium referred to in this specification is not limited to distribution mediums, but also includes storage mediums such as magnetic disks and semiconductor memory provided inside the computer or in devices connected via a network.

[0099] In short, this invention is not limited to the embodiments described above, and can be modified in various ways during implementation without departing from its essence. Furthermore, each embodiment may be combined as appropriately as possible, in which case the combined effects can be obtained. Moreover, the embodiments described above include inventions at various stages, and various inventions can be extracted by appropriate combinations of the multiple constituent elements disclosed. [Explanation of Symbols]

[0100] 1…Placement Optimization Device 2…Input device 3…Output device 10…Control Unit 101…Acquisition Department 1011...Network topology acquisition unit 1012...Parameter acquisition section 1013...Location Information Acquisition Unit 102...Movement Time Matrix Calculation Unit 103...Formalization of the Planning Problem 104...Optimization Department 105…Output control unit 20...Program memory unit 30...Data storage unit 301...Data Acquisition Storage Unit 40…Communication Interface 50… Input / Output Interface

Claims

1. A network topology indicating the location of a node and a depot location where recovery resources are on standby in the event of a node failure; network parameters; and an acquisition unit that acquires location information including the latitude and longitude of the node location and the depot location. A travel time matrix calculation unit calculates a travel time matrix showing the travel time between the node location and the depot location based on the location information, A formulation unit that formulates a mixed-integer programming problem that discretely represents the placement of the recovery resources based on the network topology, network parameters, and the travel time matrix, An optimization unit that optimizes the aforementioned mixed-integer programming problem using a heuristic method with an extended Lagrangian function method and determines optimal placement information that indicates the optimal placement of the recovery resources, An output control unit that outputs the aforementioned optimal placement information, A placement optimization device equipped with the following features.

2. The arrangement optimization device according to claim 1, wherein the discrete arrangement of the recovery resources is represented using binary variables.

3. The network topology indicates the nodes to which the recovery resources are responsible, and the network parameters include at least the failure rate of the nodes and the time spent working on the nodes. The configuration unit randomly determines the initial values ​​of the placement of the recovery resources and the assignment probability to the nodes that the recovery resources are responsible for, determines the time it takes for the nodes to recover from a failure based on the location information, the assignment probability, the failure rate, and the work time, and calculates the operational rate, which is the probability that no failure has occurred at the node, based on the assignment probability and the determined time, as described in claim 1.

4. The configuration unit calculates, based on the operating rate, the impact level, which is the amount of decrease in communication volume compared to when no failure occurs, for each combination of failure locations, and calculates the expected value of the calculated impact level, as described in claim 3.

5. The placement optimization device according to claim 1, wherein the optimization unit determines the optimal placement information by searching for the placement of the recovery resources using a genetic algorithm and optimizing the allocation probability of the recovery resources using a gradient method.

6. The arrangement optimization device according to claim 5, wherein the optimization unit represents the arrangement of the recovery resources as an array of the same size as the number of recovery resources and holds the number of the destination depot, and generates the next generation gene expression by performing genetic manipulation on the optimized value using the gradient method.

7. A placement optimization method performed by the processor of an optimization device, The process involves obtaining a network topology that shows the location of a node and a depot location where recovery resources are on standby in the event of a node failure, network parameters, and location information including the latitude and longitude of the node location and the depot location. Based on the aforementioned location information, calculate a travel time matrix showing the travel time between the node location and the depot location, To formulate a mixed-integer programming problem that discretely represents the placement of the recovery resources based on the network topology, network parameters, and the travel time matrix, The aforementioned mixed-integer programming problem is optimized using a heuristic method with the extended Lagrangian function method, Determining optimal placement information that indicates the optimal placement of the aforementioned recovery resources, Outputting the aforementioned optimal placement information, A method for optimizing placement, comprising the following features.

8. A placement optimization program comprising instructions to be executed by the processor of an optimization device, wherein the instructions are: The process involves obtaining a network topology that shows the location of a node and a depot location where recovery resources are on standby in the event of a node failure, network parameters, and location information including the latitude and longitude of the node location and the depot location. Based on the aforementioned location information, calculate a travel time matrix showing the travel time between the node location and the depot location, To formulate a mixed-integer programming problem that discretely represents the placement of the recovery resources based on the network topology, network parameters, and the travel time matrix, The aforementioned mixed-integer programming problem is optimized using a heuristic method with the extended Lagrangian function method, Determining optimal placement information that indicates the optimal placement of the aforementioned recovery resources, Outputting the aforementioned optimal placement information, A placement optimization program equipped with the following features.

Images