Optical path design device, optical path design method and program

Abstract

According to the present invention, a demand acquisition unit acquires a path demand which indicates a start point, an end point, and required communication capacity of communication. A graph generation unit generates an auxiliary graph in which a plurality of nodes, which constitute an optical network, are connected through existing edges that indicate a wavelength path to which a path demand can be allocated, in a wavelength path set in optical communication channels, and through new edges that indicate a candidate of a wavelength path that can be newly set. In the auxiliary graph, a transmission mode, a frequency of the wavelength path, and weights of the edges are set to each of the edges. A search unit searches the auxiliary graph for a route from the start point to the end point on the basis of the weights, the route being indicated by the path demand. A design unit sets a wavelength path in an optical communication channel in a transmission mode pertaining to the new edges included in the searched route, and accepts the path demand in the wavelength path at a frequency pertaining to each of the edges included in the route.

Description

[Technical Field]

[0001] The present invention relates to an optical path design device, an optical path design method, and a program. [Background technology]

[0002] A multi-layer network is known that consists of a wavelength division multiplexing (WDM) path in which one or more wavelengths are multiplexed and transmitted on an optical fiber, and one or more time division multiplexing (TDM) paths in which each wavelength multiplexing path is accommodated in a unit of fixed-length time slots.When a demand for multi-layer paths arises in such a multi-layer network, it is necessary to accommodate the demand for multi-layer paths by determining the route and frequency of the wavelength multiplexing path and the route of the time division multiplexing path.

[0003] Non-Patent Document 1 discloses a method for performing multi-layer path accommodation design using an auxiliary graph. Non-Patent Document 1 discloses that when a demand for a multi-layer path between two points occurs, an appropriate route and frequency can be selected by searching for the shortest route in an auxiliary graph that combines edges formed by existing wavelength multiplexing paths that can accommodate the demand and edges formed by new paths. [Prior art documents] [Non-patent literature]

[0004] [Non-Patent Document 1] Shuqiang Zhang, Charles Martel, Biswanath Mukherjee, “Dynamic traffic grooming in elastic optical networks,” IEEE Journal on selected areas in communications, vol. 31, no. 1. Jan. 2013. Summary of the Invention [Problem to be solved by the invention]

[0005] On the other hand, recent advances in digital coherent optical communication technology have made it possible to arbitrarily set modulation methods, symbol rates, etc. in a single device, which has made design more complex. Generally, in digital coherent optical transceivers, to simplify state management, modulation methods, symbol rates, etc. are selected from transmission modes, which are combinations of these. For example, if QPSK and 16QAM are available as modulation methods and there are two bit rates for each, there are four transmission modes. In reality, there may be even more combinations of transmission modes than this.

[0006] However, no method has been proposed for determining the transmission mode simultaneously with the route and frequency of a multi-layer path. If the transmission mode is designed separately from the route and frequency, the range of possible solutions may be narrowed, and the output solution may deviate from the optimal solution. Furthermore, Non-Patent Document 1 also does not disclose a method for selecting the best transmission mode from a large number of transmission modes.

[0007] An object of the present invention is to provide an optical path design device, an optical path design method, and a program that can determine a transmission mode simultaneously with a route in an optical network that includes wavelength multiplexed paths with multiple transmission modes. [Means for solving the problem]

[0008] One aspect of the present invention is an optical path design device that designs paths based on path demand in an optical network having an optical communication path on which one or more time-divided wavelength paths are set, the optical path design device comprising: a demand acquisition unit that acquires the path demand, which indicates the start point and end point of communication, and a required communication capacity; a graph generation unit that generates an auxiliary graph that connects multiple nodes constituting the optical network by existing edges, which are edges that indicate wavelength paths to which the path demand can be assigned among the wavelength paths already set on the optical communication path, and new edges, which are edges that indicate candidates for wavelength paths that can be newly set, and in which the transmission method and frequency of the wavelength path and the weight of the edge are set on each edge; a search unit that searches the auxiliary graph based on the weight for a route from the start point to the end point indicated by the path demand; and a design unit that sets a wavelength path on the optical communication path using the transmission method associated with the new edge included in the searched route, and generates path design information for accommodating the path demand on the wavelength path at the frequency associated with each edge included in the route.

[0009] One aspect of the present invention is an optical path design method for designing paths based on path demand in an optical network having an optical communication path in which one or more wavelength multiplexing paths are set, the optical path design method including: a demand acquisition step for acquiring the path demand, which indicates the start point and end point of communication, and a required communication capacity; an auxiliary graph generation step for generating an auxiliary graph connecting multiple nodes constituting the optical network using existing edges, which are edges indicating wavelength multiplexing paths to which the path demand can be assigned, among the wavelength multiplexing paths already set in the optical communication path, and new edges, which are edges indicating candidates for wavelength multiplexing paths that can be newly set; an edge setting step for setting, for each edge of the auxiliary graph, a transmission mode of the wavelength multiplexing path indicated by the edge, a frequency to which the path demand is to be assigned, and a weight; a search step for searching, based on the weight, for a route from the start point to the end point indicated by the path demand in the auxiliary graph; and a design step for generating path design information, which sets a wavelength multiplexing path in the optical communication path in the transmission mode associated with the new edge included in the searched route, and assigns the path demand to the wavelength multiplexing path at a frequency associated with each edge included in the route.

[0010] One aspect of the present invention is a program that causes a computer to function as the optical path design device according to the above aspect. [Effects of the Invention]

[0011] According to the above aspect, in an optical network including wavelength multiplexed paths with multiple transmission modes, the transmission mode can be determined simultaneously with the route. [Brief explanation of the drawings]

[0012] [Figure 1] 1 is a diagram illustrating an optical path design device and a multi-layer network NW according to an embodiment. [Figure 2] FIG. 10 is a diagram illustrating an example of a method for accommodating multi-layer path demand in an embodiment. [Figure 3] 1 is a schematic block diagram illustrating a configuration of an optical path design apparatus according to a first embodiment. [Figure 4] 3 is a flowchart showing a multi-layer path design method according to the first embodiment. [Figure 5] 4 is a flowchart showing a method for generating an existing graph according to the first embodiment. [Figure 6] 1 is a flowchart showing a method for generating a new graph according to the first embodiment. [Figure 7] FIG. 2 is a diagram illustrating an example of a configuration of a physical network according to the first embodiment. [Figure 8] FIG. 2 is a diagram showing an example of an existing graph according to the first embodiment. [Figure 9] FIG. 2 is a diagram showing an example of a newly created graph according to the first embodiment. [Figure 10] FIG. 2 is a diagram illustrating an example of an auxiliary graph according to the first embodiment. [Figure 11] FIG. 4 is a diagram showing a route search result according to the first embodiment. [Figure 12] FIG. 10 is a schematic block diagram illustrating a configuration of an optical path design apparatus according to a second embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0013] Hereinafter, an embodiment of an optical path design device, an optical path design method, and a program will be described. Fig. 1 is a diagram showing an optical path design device 10 and a multi-layer network NW according to an embodiment. When a new multi-layer path demand (connection request) occurs in the multi-layer network NW, the optical path design device 10 according to the embodiment designs a multi-layer path that accommodates the multi-layer path demand.

[0014] A multi-layer network NW is a network in which multiple nodes N are connected via optical communication paths such as optical fibers. Nodes N may be digital coherent optical transceivers. Each node N establishes a wavelength multiplexing path between itself and the opposing node, which multiplexes optical signals of one or more wavelengths. A wavelength multiplexing path is a path that multiplexes one or more wavelength paths. A wavelength path is assigned to a frequency band consisting of a group of frequency slots among multiple frequency slots that are obtained by dividing the frequency band available in the optical communication path. The frequency bandwidth of a wavelength path is determined by the modulation method and bit rate. Combinations of modulation method and symbol rate of signals that can be set in a wavelength multiplexing path are prepared in advance as transmission modes. The modulation method, symbol rate, and transmission mode are examples of transmission methods.

[0015] In a multi-layer network NW, wavelength multiplexing paths are time-division multiplexed into multiple time slots. In other words, a multi-layer network NW is composed of optical communication paths with one or more time-division wavelength paths set. A combination of wavelength path and time slot is assigned to a multi-layer path demand.

[0016] Here, the accommodation of the multi-layer path demand will be explained. Fig. 2 is a diagram showing an example of a method for accommodating the multi-layer path demand in the embodiment. In the example shown in Figure 2, the multi-layer network NW has the following configuration. The multi-layer network NW includes nodes N1, N2, and N3. Nodes N1 and N2 are connected via an optical communication path, and nodes N2 and N3 are connected via an optical communication path. A wavelength path P1 using frequency slot F2 is set between nodes N2 and N3.

[0017] Here, when a demand arises for a multi-layer path that starts at node N1 and ends at node N3, the demand for the multi-layer path can be accommodated as follows. The first accommodation method is to set up a new wavelength path P2 with a hop count of 1 using frequency slot F1 between node N1 and node N2, and accommodate the multilayer path demand in any time slot of the new wavelength path P2 and in an unused time slot of the existing wavelength path P1. In this case, electrical relay processing is performed at node N2. The second accommodation method is to set up a new wavelength path P3 with a hop count of 2 using frequency slot F1 between node N1 and node N3, and accommodate the multi-layer path demand in any time slot of the new wavelength path P3. In this case, node N2 transfers the optical signal input from node N1 to node N3 without performing opto-electrical conversion.

[0018] When a demand for a multi-layer path occurs, the optical path design device 10 according to the embodiment designs a route that appropriately accommodates the demand for a multi-layer path.

[0019] First Embodiment FIG. 3 is a schematic block diagram showing the configuration of the optical path design apparatus 10 according to the first embodiment. The optical path design device 10 includes a demand acquisition unit 11, a graph generation unit 12, a search unit 13, a design unit 14, and a storage unit 15. The demand acquisition unit 11 acquires multi-layer path demands from devices connected to the multi-layer network NW. The multi-layer path demands are transferred by, for example, a node N connected to the device. The multi-layer path demands include a start node, an end node, and a required bit rate (required communication capacity).

[0020] The graph generation unit 12 generates an auxiliary graph consisting of nodes N that configure the multi-layer network NW and edges that indicate wavelength multiplexing paths that are constructed between the nodes N and that satisfy the required bit rate of the multi-layer path demand. The graph generation unit 12 includes an existing graph generation unit 121, a new graph generation unit 122, and a weight determination unit 123.

[0021] The existing graph generation unit 121 generates an existing graph having wavelength paths, among existing wavelength paths, to which a multi-layer path demand can be additionally allocated as edges. A wavelength path to which a multi-layer path demand can be additionally allocated is a wavelength path that has enough available time slots to satisfy the requested bit rate of the multi-layer path demand.

[0022] The new graph generation unit 122 generates a new graph having edges corresponding to candidates for newly set wavelength paths. The new graph generation unit 122 includes a transmission mode determination unit 1221 that determines parameters such as the transmission mode and frequency of the wavelength path candidates related to the edges of the new graph, and a frequency determination unit 1222 that assigns frequencies to the wavelength path candidates.

[0023] The weight determination unit 123 determines edge weights of the auxiliary graph.

[0024] The search unit 13 searches for a route connecting a start node and an end node of a multi-layer path demand.

[0025] Based on the searched route, the design unit 14 generates multi-layer path design information including instructions to set new wavelength paths and instructions to accommodate the multi-layer path demand by specifying frequencies and time slots.

[0026] The storage unit 15 stores information necessary for generating an auxiliary graph. The storage unit 15 has storage areas related to a transmission mode storage unit 151, a topology storage unit 152, and a path storage unit 153.

[0027] The transmission mode storage unit 151 stores, for each transmission mode, a modulation method, a symbol rate, a required number of frequency slots (bandwidth), and a maximum number of hops that can be transmitted in association with each other. The maximum number of hops is an example of the distance between nodes. In other embodiments, the transmission mode storage unit 151 may store other distances, such as the path length in the physical topology, instead of the maximum number of hops. Furthermore, the higher the frequency efficiency of a transmission mode, the smaller the maximum number of hops that can be transmitted. The transmission mode storage unit 151 may store available transmission modes for each node N.

[0028] The topology storage unit 152 stores information on the physical topology of the optical transmission paths and nodes N that constitute the multi-layer network NW. That is, the topology storage unit 152 stores information indicating the connection relationships between the nodes N.

[0029] The path storage unit 153 stores information about existing multi-layer paths. For example, for an existing wavelength path, the path storage unit 153 stores the node N constituting the wavelength path, the transmission mode, the frequency slot to be used, and the time slot in use.

[0030] FIG. 4 is a flowchart showing a multi-layer path design method according to the first embodiment. When a multi-layer path demand is input, the optical path design apparatus 10 attempts to accommodate the multi-layer path demand in accordance with the procedure shown in FIG. First, the existing graph generating unit 121 generates an existing graph indicating wavelength paths to which the requested bit rate of the multi-layer path demand can be allocated from among the wavelength paths that have already been allocated (step S1).

[0031] FIG. 5 is a flowchart showing a method for generating an existing graph according to the first embodiment. The existing graph generation unit 121 first generates an existing graph that includes only node N with no edges (step S11). Next, the existing graph generation unit 121 identifies all existing wavelength paths based on the information held by the path holding unit 153, selects the existing wavelength paths one by one, and executes the following processes from step S13 to step S16 (step S12).

[0032] The existing graph generation unit 121 identifies the number of time slots required to satisfy the requested bit rate based on the transmission mode of the selected wavelength path (step S13). The existing graph generation unit 121 determines whether the identified number of time slots are available in the selected wavelength path (step S14). If the identified number of time slots are available (step S14: YES), the existing graph generation unit 121 adds an edge connecting the start point and end point of the selected wavelength path to the existing graph (step S15). The weight determination unit 123 determines the weight of the added edge according to the following equation (1) (step S16).

[0033]

number

[0034] If the specified number of time slots are not available in the selected wavelength path (step S14: NO), the next wavelength path is selected without adding an edge related to the wavelength path.

[0035] In this way, the existing graph is created according to the procedure shown in Fig. 5. If there are multiple edges between the same node pair in the existing graph, the existing graph generation unit 121 may leave only the edge with the smallest weight and delete the other edges. Next, the new graph generation unit 122 generates a new graph indicating candidate wavelength paths to which the requested bit rate of the multi-layer path demand can be allocated (step S2).

[0036] FIG. 6 is a flowchart showing a method for generating a new graph according to the first embodiment. The new graph generation unit 122 first generates a new graph that has only nodes N with no edges (step S21). Next, the new graph generation unit 122 identifies all patterns of pairs of nodes N based on the information stored in the topology storage unit 152, selects pairs of nodes N one by one, and executes the following processes from step S23 to step S26 (step S22).

[0037] The new graph generation unit 122 searches for the shortest path on the physical network connecting the selected pair of nodes N based on the information stored in the topology storage unit 152 (step S23). The search for the shortest path may be performed using a search algorithm such as Dijkstra's algorithm or A* algorithm, with distance as the edge weight. The new graph generation unit 122 refers to the transmission mode storage unit 151 and determines whether there are one or more transmission modes that can be set for the wavelength path connecting the identified shortest path (step S24). In other words, the new graph generation unit 122 determines whether there is a transmission mode whose maximum hop count is equal to or greater than the hop count of the shortest path.

[0038] If there are one or more configurable transmission modes (step S24: YES), the transmission mode determination unit 1221 of the new graph generation unit 122 selects an appropriate transmission mode from the one or more configurable transmission modes based on a predetermined policy (step S25). For example, from the one or more configurable transmission modes, the transmission mode determination unit 1221 may select the transmission mode with the smallest difference between the number of hops of the path and the maximum number of hops, the transmission mode with the largest transmission capacity, or the transmission mode with the smallest required frequency band.

[0039] Based on the information stored in the path holding unit 153, the establishment graph generation unit 122 determines whether or not there is an available frequency slot that can secure the frequency bandwidth required to set the selected transmission mode in the optical communication path connecting the nodes on the shortest path (step S26). An available frequency slot is a frequency slot that is not assigned to another wavelength path. If there is an available frequency slot (step S26: YES), the frequency determination unit 1222 of the establishment graph generation unit 122 selects an appropriate frequency slot from the available frequency slots (step S27). For example, the frequency determination unit 1222 may select the frequency slot with the smallest frequency slot number from among the available frequency slots. Then, the establishment graph generation unit 122 adds an edge connecting the selected pair of nodes N to the establishment graph (step S28). The weight determination unit 123 determines the weight of the added edge according to the following equation (2) (step S29).

[0040]

number

[0041] If there is no transmission mode that can be set for the selected pair of nodes N (step S24: NO), or if there is no available frequency slot (step S26: NO), the next pair of nodes N is selected without adding an edge connecting the pair of nodes N.

[0042] When the new graph is created according to the procedure shown in FIG. 6, the graph generation unit 12 generates an auxiliary graph by combining the existing graph and the new graph (step S3). If there are multiple edges between the same node pair in the auxiliary graph, the graph generation unit 12 may leave only the edge with the smallest weight and delete the other edges. Next, the search unit 13 searches for a route on the auxiliary graph that connects the start node N and the end node N of the multi-layer path demand and minimizes the total weight (step S4). The route search may be performed using a search algorithm such as the Dijkstra algorithm or the A* algorithm, with distance as the edge weight. As a result of the search, the search unit 13 determines whether or not a route connecting the start node N and the end node N exists (step S5).

[0043] If a route exists (step S5: YES), the design unit 14 generates path design information for realizing the route (step S6). Specifically, for edges on the searched route that are related to wavelength path candidates, the design unit 14 generates a setting instruction for newly setting a wavelength path based on the path, frequency slot, and transmission mode set for the edge. The design unit 14 also generates an allocation instruction for a time slot that realizes the requested bit rate for the wavelength path related to each edge on the searched route. The design unit 14 transmits the setting instruction and allocation instruction to the corresponding node N.

[0044] If no route exists (step S5: NO), the design unit 14 rejects the multi-layer path demand because it cannot be set (step S7). At this time, the design unit 14 may notify the sender of the multi-layer path demand of the rejection.

[0045] Here, a multi-layer path design method will be explained using a specific example. In the example shown below, the following two transmission modes can be set as the transmission mode. Transmission mode M1 uses 16QAM (Quadrature Amplitude Modulation) as the modulation method, a bit rate of 200Gbps, and a maximum number of hops of 1. Transmission mode M2 ​​uses QPSK (Quadrature Phase Shift Keying) as the modulation method, a bit rate of 100Gbps, and a maximum number of hops of 2.

[0046] Fig. 7 is a diagram illustrating an example of the configuration of a physical network according to the first embodiment. As illustrated in Fig. 7, the multi-layer network NW includes six nodes N, namely, node N1, node N2, node N3, node N4, node N5, and node N6. Node N1 is physically connected to nodes N2 and N6 via optical transmission paths. Node N2 is physically connected to nodes N1, N3, and N6 via optical transmission paths. Node N3 is physically connected to nodes N2, N4, and N5 via optical transmission paths. Node N4 is physically connected to nodes N3 and N5 via optical transmission paths. Node N5 is physically connected to nodes N3, N4, and N6 via optical transmission paths. The node N6 is physically connected to the nodes N1, N2, and N5 via optical transmission paths. In such a multi-layer network NW, it is assumed that a demand for a multi-layer path between node N1 and node N4 occurs.

[0047] The existing graph generation unit 121 generates an existing graph. FIG. 8 is a diagram showing an example of an existing graph according to the first embodiment. In this example, it is assumed that there are existing wavelength paths with different timeslots between node N1 and node N3, between node N3 and node N5, and between node N4 and node N5. Note that the wavelength path between node N1 and node N3 is a two-hop path via node N2. In the example shown in FIG. 8, x app =0.5, y app = 0.5, which means that the weight of a 1-hop edge is 1.0 and the weight of a 2-hop edge is 1.5.

[0048] The new graph generation unit 122 generates a new graph. FIG. 9 is a diagram showing an example of a new graph according to the first embodiment. In this example, one-hop edges are provided between node N1 and node N2, between node N2 and node N3, between node N3 and node N4, between node N4 and node N5, between node N5 and node N6, between node N1 and node N6, between node N2 and node N6, and between node N3 and node N5, and two-hop edges are provided between node N1 and node N3, between node N2 and node N4, between node N1 and node N5, and between node N4 and node N6. The transmission mode determination unit 1221 of the new graph generation unit 122 determines the transmission mode with priority given to the bit rate. Therefore, 16QAM transmission mode 1 is associated with the one-hop edge, and QPSK transmission mode 2 is associated with the two-hop edge. In the example shown in FIG. 9, x est =1.0, y est = 0.5, which means that the weight of a 1-hop edge is 1.5 and the weight of a 2-hop edge is 2.0.

[0049] The graph generation unit 12 generates an auxiliary graph by combining the existing graph and the new graph. FIG. 10 is a diagram illustrating an example of an auxiliary graph according to the first embodiment. When there are multiple edges for the same node N pair, the edge with the smaller weight is selected. The search unit 13 calculates a route that accommodates the multi-layer path demand. FIG. 11 is a diagram illustrating a route search result according to the first embodiment. As shown in FIG. 11, the search unit 13 detects a route that passes through the edge between node N1 and node N3, which has the smallest total weight, and the edge between node N3 and node N4. As a result, the design unit 14 adds a time slot to the two-hop wavelength path between node N1 and node N3, and generates path design information for establishing a wavelength multiplexing path of transmission mode 2 between node N3 and node N4.

[0050] As described above, according to the first embodiment, the optical path design device 10 generates an auxiliary graph having existing edges indicating wavelength paths to which the path demand can be allocated among existing wavelength paths, and new edges indicating candidate wavelength paths that satisfy the requested bit rate related to the multi-layer path demand. The optical path design device 10 sets the transmission mode and frequency of the wavelength path and the weight of the edge to each edge of the auxiliary graph. The optical path design device 10 sets the wavelength path to the new edge included in the route searched by the auxiliary graph, and generates path design information for accommodating the multi-layer path demand to the wavelength path related to each edge. In other words, by setting a transmission mode and frequency to a newly established edge in an auxiliary graph, the optical path design device 10 can simultaneously determine the wavelength path to be newly established to accommodate the multi-layer path demand and its transmission mode.

[0051] In the first embodiment, an edge is added by selecting one appropriate transmission mode from the transmission modes that can be set in step S25, but this is not limited to this. For example, in other embodiments, an edge may be added for each transmission mode with a weight according to the transmission mode. In this case, the weight coefficient x est or y estmay be set to a value corresponding to the frequency bandwidth or the bit rate, or the edge weight may be calculated by adding a term corresponding to the frequency bandwidth or the bit rate to Equation (2).

[0052] In the first embodiment, an appropriate frequency slot is selected from multiple frequency slots in step S27 to add an edge, but this is not limited to this. For example, in other embodiments, weights according to the center frequencies of multiple frequency bands may be calculated, and the edge with the smallest weight may be added. In this case, the edge weight is calculated by adding a term for the weight according to the center frequency to equation (2). The weight according to the center frequency may be smaller as the frequency slot number becomes smaller, or may be smaller as the number of consecutive free frequency slots remaining after the addition of the wavelength path becomes larger.

[0053] In the first embodiment, the optical path design device 10 configures all nodes in the network as the nodes N for which the existing graph and the new graph are to be constructed, but this is not limited to this. For example, in another embodiment, the amount of calculation may be reduced by limiting the nodes N for which the graph is to be constructed to some nodes. For example, in another embodiment, the optical path design device 10 pre-calculates a K-Shortest Paths algorithm that calculates K shortest paths between the start point and end point of a multi-layer path, and only edges included in the nodes N that pass through this algorithm are configured as the existing graph and the new graph, thereby reducing the amount of calculation while leaving the nodes N necessary for calculating the optimal solution.

[0054] Second Embodiment FIG. 12 is a schematic block diagram showing the configuration of an optical path design apparatus 10 according to the second embodiment. The optical path design device 10 according to the second embodiment determines edge weights using a reinforcement learning model. That is, the optical path design device 10 according to the second embodiment differs from the first embodiment in the processing of the weight determination unit 123. Furthermore, the storage unit 15 according to the second embodiment further stores a model holding unit 154 that holds a trained model related to reinforcement learning.

[0055] When a feature vector representing the environment of the multi-layer network NW (such as the transmission mode, frequency band, and time slot of the existing wavelength path) is input, the trained model held by the model holding unit 154 is app , y app , x est , y est This is a model trained to output The weight determination unit 123 inputs the state of the multi-layer network NW when a multi-layer path demand occurs into the trained model, thereby determining x app , y app , x est , y est The weight determination unit 123 calculates the calculated x app , y app , x est , y est The weight of each edge is determined by performing the calculations of equations (1) and (2) using the above equations. The output of the trained model is x app , y app , x est , y est In this case, the weight determining unit 123 determines the probability density function of x app , y app , x est , y est For each of the edges, the peak value of the probability density function may be used to determine the weight of each edge.

[0056] Moreover, the optical path design device 10 according to the second embodiment further includes, in addition to the configuration of the first embodiment, a demand generation unit 16, a simulator 17, and an update unit 18. The demand generation unit 16, the simulator 17, and the update unit 18 function as agents of reinforcement learning. The simulator 17 simulates the behavior of the multi-layer network NW. The simulator 17 simulates the design of a path based on the path design information from the design unit 14, and simulates communication based on the designed path. The demand generator 16 generates a multi-layer path demand to be simulated by the simulator 17. The demand generator 16 determines, for example, the start node, end node, and required bit rate of the multi-layer path demand, as well as the generation timing, based on random numbers.

[0057] The update unit 18 calculates a reward for the path design based on the behavior of the simulator 17 after the path designed by the design unit 14 in response to the multi-layer path demand generated by the demand generation unit 16 is set, and updates the parameters of the model held by the model holding unit 154 based on the calculated reward. The reward design can be arbitrarily set by the operator; for example, when an objective function is designed to minimize the rejection rate of the multi-layer path demand, the reward can be set to +1 if the design is successful and to -1 if the design is rejected. The update unit 18 updates the parameter x so that the reward for the path design is maximized. app , y app , x est , y est Update.

[0058] As described above, according to the second embodiment, the optical path design device 10 determines edge weights using a reinforcement learning model. While an appropriate path design may change depending on the environment of the multi-layer network NW, the optical path design device 10 of the second embodiment can realize a path design that suits the environment by appropriately changing the edge weights depending on the environment.

[0059] In the second embodiment, the parameter x app , y app , x est , y est For example, in another embodiment, the trained model calculates a common x est , y est Instead, x for each transmission mode of the new pathi est , y i est The subscript i indicates the number of the transmission mode.

[0060] In still other embodiments, the trained model includes a parameter x app , y app , x est , y est In addition to or instead of the above, the trained model may be configured to output a transmission mode to be set. In this case, the trained model may be configured to output a reward value for each transmission mode, and the transmission mode determination unit 1221 may select the transmission mode with the highest reward.

[0061] In still other embodiments, the trained model includes a parameter x app , y app , x est , y est In addition to or instead of the above, the trained model may be configured to output a frequency to be set. In this case, the trained model may be configured to output a reward value for each frequency slot, and the frequency determiner 1222 may select a frequency band centered on the frequency slot with the highest reward.

[0062] In the second embodiment, the optical path design device 10 performs both learning of the model parameters and calculation of weights using the learned model, but this is not limited to this. For example, in another embodiment, the configuration for learning parameters (e.g., the demand generation unit 16, the simulator 17, and the update unit 18) may be executed by a separate learning device, and the optical path design device 10 may calculate weights using the learned model that is the learning result of the learning device.

[0063] Other Embodiments Although one embodiment has been described in detail above with reference to the drawings, the specific configuration is not limited to the above, and various design modifications are possible. That is, in other embodiments, the order of the above-described processes may be changed as appropriate. Furthermore, some processes may be executed in parallel. The optical path design device 10 according to the above-described embodiment may be configured by a single computer, or the configuration of the optical path design device 10 may be distributed across multiple computers, and the multiple computers may work together to function as the optical path design device 10.

[0064] Furthermore, although the node N according to the above-described embodiment determines the modulation scheme and symbol rate by selecting a transmission mode, this is not limiting. For example, the node N according to other embodiments may directly specify the modulation scheme and symbol rate.

[0065] <Computer Configuration> The optical path design device 10 includes a processor, a memory, an auxiliary storage device, etc., which are connected by a bus, and by executing an optical path design program, functions as a device including a demand acquisition unit 11, a graph generation unit 12, a search unit 13, a design unit 14, and a storage unit 15. Examples of the processor include a CPU (Central Processing Unit), a GPU (Graphic Processing Unit), and a microprocessor. The optical path design program may be recorded on a computer-readable recording medium. Examples of the computer-readable recording medium include a storage device such as a magnetic disk, a magneto-optical disk, an optical disk, or a semiconductor memory. The optical path design program may be transmitted via a telecommunications line. All or part of the functions of the optical path design program may be realized using a custom LSI (Large Scale Integrated Circuit) such as an ASIC (Application Specific Integrated Circuit) or a PLD (Programmable Logic Device). Examples of PLDs include PAL (Programmable Array Logic), GAL (Generic Array Logic), CPLD (Complex Programmable Logic Device), and FPGA (Field Programmable Gate Array). Such integrated circuits are also included in the scope of the processor. [Explanation of symbols]

[0066] 10... Optical path design device 11... Demand acquisition unit 12... Graph generation unit 121... Existing graph generation unit 122... New graph generation unit 1221... Transmission mode determination unit 1222... Frequency determination unit 123... Weight determination unit 13... Search unit 14... Design unit 15... Memory unit 151... Transmission mode storage unit 152... Topology storage unit 153... Path storage unit 154... Model storage unit 16... Demand generation unit 17... Simulator 18... Update unit N... Node NW... Multi-layer network

Claims

1. An optical path design device that designs paths based on path demand in an optical network having an optical communication path in which one or more time-divided wavelength paths are set, comprising: a demand acquisition unit that acquires the path demand indicating a start point, an end point, and a required communication capacity of the communication; a graph generating unit that generates an auxiliary graph that connects a plurality of nodes that constitute the optical network by existing edges, which are edges that indicate wavelength paths that can be assigned the path demand among wavelength paths that have already been set in the optical communication path, and new edges, which are edges that indicate candidates for wavelength paths that can be newly set, and in which the transmission method and frequency of the wavelength path and the weight of the edge are set in each edge; a search unit that searches the auxiliary graph for a route from a start point to an end point indicated by the path demand based on the weight; a design unit that sets a wavelength path in the optical communication path using the transmission method related to the new edge included in the searched route, and generates path design information for accommodating the path demand in the wavelength path using a frequency related to each edge included in the route; Equipped with The graph generation unit specifies the number of time slots required for each wavelength path already set in the optical communication path based on the required communication capacity related to the path demand and the transmission method of the wavelength path, and determines the wavelength path that can secure the required time slots as the existing edge. Optical path design device.

2. The graph generation unit sets the weights for each edge by reinforcement learning using a reward for the allocation of the path demands based on a state of the optical network simulated by a simulator that simulates the state of the optical network after the path demands are allocated. The optical path design device according to claim 1 .

3. The graph generation unit sets the weight of each edge based on the distance of the wavelength path.

3. The optical path design device according to claim 1.

4. The graph generation unit sets the transmission method of the new edge to a transmission method among a plurality of transmission methods in which the difference between the transmission distance and the distance of the wavelength path related to the new edge is smallest.

3. The optical path design device according to claim 1.

5. The graph generation unit sets the weights based on the transmission method or the frequency.

3. The optical path design device according to claim 1.

6. the graph generation unit provides the new edge for each transmission method or each frequency, The search unit searches for a route consisting of a combination of edges with the smallest weight. The optical path design device according to claim 5 .

7. An optical path design method for designing paths based on path demand in an optical network having an optical communication path in which one or more time-divided wavelength paths are set, comprising: a demand acquisition step of acquiring the path demand indicating a start point, an end point, and a required communication capacity of the communication; an auxiliary graph generation step of generating an auxiliary graph that connects a plurality of nodes constituting the optical network by existing edges, which are edges indicating wavelength paths to which the path demand can be allocated among wavelength paths already set in the optical communication path, and new edges, which are edges indicating candidates for wavelength paths that can be newly set, and in which the transmission method and frequency of the wavelength path and the weight of the edge are set; a search step of searching for a route from a start point to an end point indicated by the path demand in the auxiliary graph based on the weight; a design step of generating path design information, which sets wavelength paths in the optical communication path using the transmission method related to the new edge included in the searched route, and allocates the path demand to the wavelength paths using frequencies related to each edge included in the route; Including, In the auxiliary graph generation step, for each wavelength path already set in the optical communication path, the number of time slots required is identified based on the required communication capacity related to the path demand and the transmission method of the wavelength path, and the wavelength path that can secure the required time slots is determined as the existing edge. Optical path design method.

8. A program for causing a computer to function as the optical path design device according to claim 1 or 2.

Images