Routing selection method, device, electronic device and storage medium

Abstract

The present invention provides a routing selection method, device, electronic device and storage medium. The method comprises: for any destination, obtaining the predicted total traffic volume and the predicted maximum traffic volume per unit time from the traffic input place to the target destination within the predicted time period; based on the dynamic programming algorithm, according to the predicted total traffic volume, the tariff model of each transit business, the link capacity requirement of the target destination and the link capacity limit of each transit business to the target destination, determining the transit path with the lowest total tariff from the traffic input place to the target destination; based on the transit path with the lowest total tariff corresponding to each destination, determining the total link capacity requirement of each transit business; according to the comparison result between the total link capacity requirement of each transit business and the link capacity of each transit business, adjusting the transit path with the lowest total tariff corresponding to each destination, and obtaining the optimal transit path corresponding to each destination. The cost and call quality can be made to reach a controllable and optimal balance.

Description

Technical Field

[0001] The present invention relates to the field of communication technology, and in particular to a routing selection method, device, electronic equipment and storage medium. Background Art

[0002] Routing selection refers to how the switch responsible for transferring outbound calls selects physical lines to transmit international long-distance calls after they are generated. In the process of routing calls in multiple regions, there is the problem of selecting multiple transit providers. Because there are multiple billing structures, commitments, and penalty mechanisms in the agreements between operators and transit providers, and there are also overlapping times between the agreements, it is not possible to simply select transit providers based on the principle of the lowest unit price.

[0003] Existing routing selection schemes can be roughly divided into two types: the first is to manually select and allocate routing schemes based on line conditions when there are few regions and available transit providers. This approach wastes human resources, cannot control the overall routing cost, and cannot guarantee the call loss rate; the second is an intelligent routing selection method based on reinforcement learning, which focuses on issues such as routing protocols and path bandwidth utilization when allocating communication routes, and cannot allocate the entire routing system at a macro level to achieve a balance between cost and call quality. Summary of the invention

[0004] The present invention provides a routing selection method, device, electronic equipment and storage medium, which are used to solve the technical problem that the existing routing selection scheme cannot achieve a balance between cost and call quality.

[0005] In a first aspect, the present invention provides a routing selection method, comprising:

[0006] For any destination, obtain the predicted total traffic volume from the traffic input point to the target destination within the predicted time period and the predicted maximum traffic volume per unit time;

[0007] Based on a dynamic programming algorithm, according to the predicted total traffic volume, the tariff model of each transit provider, the link capacity requirement of the target destination, and the link capacity limit of each transit provider to the target destination, a transit path with the lowest total tariff from the traffic input location to the target destination is determined; wherein the link capacity requirement of the target destination is obtained based on the predicted maximum traffic volume per unit time;

[0008] Determine the total link capacity requirement of each transit provider based on the transit path with the lowest total tariff corresponding to each destination;

[0009] According to the comparison result between the total link capacity requirements of each transit operator and the link capacity of each transit operator, the transit path with the lowest total fee corresponding to each destination is adjusted to obtain the optimal transit path corresponding to each destination.

[0010] In one embodiment, the step of adjusting the transit path with the lowest total fee corresponding to each destination according to the comparison result between the total link capacity requirements of each transit operator and the link capacity of each transit operator to obtain the optimal transit path corresponding to each destination includes:

[0011] If it is determined that the total link capacity requirements of each transit operator are less than or equal to the link capacity of each transit operator, then the transit path with the lowest total fee corresponding to each destination is determined as the optimal transit path corresponding to each destination; or,

[0012] If it is determined that the total link capacity demand of the first transit operator is greater than the link capacity of the first transit operator, then determine the first path corresponding to the first transit operator that needs to transfer link capacity, and increase the link capacity that needs to be transferred on the first path transferred by the second transit operator, traverse all transit operators until the total link capacity demand of each transit operator is less than or equal to the link capacity of each transit operator, and obtain the optimal transit path corresponding to each destination;

[0013] Among them, the second transit operator is the transit operator with the lowest fee for transiting the link capacity that needs to be transferred on the first path among the transit operators whose total link capacity demand is less than the link capacity, and the link capacity that needs to be transferred is the link capacity difference between the total link capacity demand of the first transit operator and the link capacity of the first transit operator.

[0014] In one embodiment, determining the first path corresponding to the first transit operator and requiring link capacity transfer includes:

[0015] For any path corresponding to the first transit operator, respectively obtaining the tariff difference of the link capacity that needs to be transferred on the transit target path between the first transit operator and each transit operator whose total link capacity demand is less than the link capacity;

[0016] The path corresponding to the minimum value of each tariff difference is determined as the first path corresponding to the first transit operator that requires link capacity transfer.

[0017] In one embodiment, the step of obtaining the predicted total traffic volume and the predicted maximum traffic volume per unit time from the traffic input location to the target destination within the predicted time period includes:

[0018] Based on the ARMA model or LSTM model, the predicted total traffic volume from the traffic input point to the target destination within the predicted time period and the predicted maximum traffic volume per unit time are obtained;

[0019] The ARMA model or the LSTM model is obtained after training based on historical data samples of the total amount of traffic and the maximum amount of traffic per unit time with certain training labels, and the training labels are the true values ​​of the total amount of traffic and the maximum amount of traffic per unit time in the time period predicted by the training samples.

[0020] In one embodiment, the method of obtaining the predicted total traffic volume and the predicted maximum traffic volume per unit time from the traffic input location to the target destination within the predicted time period based on the ARMA model or the LSTM model includes:

[0021] If it is determined that the historical data series of the total traffic volume and the maximum traffic volume per unit time are stationary time series, then based on the ARMA model, the predicted total traffic volume and the predicted maximum traffic volume per unit time from the traffic input point to the target destination within the time period to be predicted are obtained; or,

[0022] If it is determined that the historical data series of the total traffic volume and the maximum traffic volume per unit time are non-stationary time series, the predicted total traffic volume and the predicted maximum traffic volume per unit time from the traffic input point to the target destination within the predicted time period are obtained based on the LSTM model.

[0023] In one embodiment, the ARMA model is an ARMA model that estimates model parameters based on the L-BFGS algorithm.

[0024] In one embodiment, the link capacity requirement of the target destination is obtained based on the predicted maximum traffic volume per unit time, including:

[0025] Based on the Erlang formula and the limited call loss rate, the predicted maximum traffic volume per unit time is converted into the link capacity requirement of the target destination.

[0026] In a second aspect, the present invention provides a routing selection device, comprising:

[0027] An acquisition module, for acquiring, for any destination, the predicted total traffic volume from the traffic input point to the target destination within the predicted time period and the predicted maximum traffic volume per unit time;

[0028] A first determination module is used to determine, based on a dynamic programming algorithm, a transit path with the lowest total tariff from the traffic input location to the target destination according to the predicted total traffic volume, the tariff model of each transit operator, the link capacity requirement of the target destination, and the link capacity limit of each transit operator to the target destination; wherein the link capacity requirement of the target destination is obtained based on the predicted maximum traffic volume per unit time;

[0029] A second determination module is used to determine the total link capacity requirement of each transit provider based on the transit path with the lowest total tariff corresponding to each destination;

[0030] The adjustment module is used to adjust the transit path with the lowest total fee corresponding to each destination according to the comparison result between the total link capacity requirements of each transit operator and the link capacity of each transit operator, so as to obtain the optimal transit path corresponding to each destination.

[0031] In a third aspect, the present invention provides an electronic device, comprising a processor and a memory storing a computer program, wherein the processor implements the steps of the routing selection method described in the first aspect when executing the program.

[0032] In a fourth aspect, the present invention provides a non-transitory computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of the routing selection method described in the first aspect.

[0033] The routing selection method, device, electronic device and storage medium provided by the present invention transfer link capacity after determining that the total link capacity demand of a transit operator is greater than the link capacity of the transit operator, so that the link capacity of each transit operator can finally meet the link capacity requirements of each destination, obtain the optimal transit path corresponding to each destination, realize the selection of the transit operator with the best price and the optimal allocation of the links of each transit operator, thereby achieving the guarantee of call quality while controlling costs, so that the two can reach a controllable balance and optimality. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0035] Figure 1 It is a flow chart of the routing selection method provided by the present invention;

[0036] Figure 2 It is a schematic diagram of the implementation of the routing selection method provided by the present invention;

[0037] Figure 3 It is a schematic diagram of the prediction process of the total amount of traffic and the maximum traffic per unit time provided by the present invention;

[0038] Figure 4 It is a structural schematic diagram of the routing selection device provided by the present invention;

[0039] Figure 5 It is a structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0040] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0041] Among the current methods of route selection, the method of selecting routes based on manual experience often relies on historical call volume and expert experience to select call links for call traffic through manual intervention. The overall process is greatly affected by manual intervention, and the routing cost cannot be finely controlled. In addition, the systemic risk is relatively large. Manual intervention cannot guarantee the correctness and timeliness of route allocation. If delays or wrong operations lead to production accidents, operators will suffer losses. However, the intelligent routing method based on reinforcement learning cannot allocate the entire routing system in a macroscopic way. The routing method based on reinforcement learning establishes the network topology between destinations and designs a decision model for routing. This method is only a routing method that selects the shortest path in the network topology of multiple destinations. In the scenario of multiple transit providers and multiple destinations, due to many restrictions and a large system, the algorithm has a large amount of calculation and the network topology occupies a lot of resources. In addition, the method based on reinforcement learning does not systematically consider the impact of historical call volume on the system, and does not analyze and predict the traffic volume data. The shortest path solution it solves does not consider the call quality. Excessive call quality loss leads to a decline in customer perceived user experience, which will also cause losses to operators.

[0042] From the results, neither the manual intervention method nor the routing method based on reinforcement learning can achieve a balanced optimal solution in terms of cost and call quality.

[0043] In order to solve the above problems existing in the prior art, the present invention provides a link allocation solution from traffic prediction to link allocation, which is a full-process and systematic solution. By predicting the maximum traffic volume per unit time of different destinations, the link capacity requirements of each destination are analyzed, and the predicted total traffic volume of different destinations, the tariff model of each transit operator, the link capacity limit of each transit operator to different destinations, and the link capacity constraint of each transit operator itself are combined. While ensuring the call quality, the decoupled link planning problem is solved to improve the allocation efficiency more efficiently. The present invention has the following main features: First, the program has a faster response speed and high accuracy, and reduces or does not require manual intervention in link selection; second, the present invention solves the planning problem of solving the optimal link combination in a decoupled scenario for multiple transit operators; third, the present invention can accurately determine the routing traffic allocation ratio of transit operators in each direction, which can make the ratio subdivided, the system planning more accurate, and the mode more efficient, thereby controlling the cost while ensuring the call quality, and both achieve a controllable balance and optimality.

[0044] Figure 1 A schematic diagram of the flow of the routing selection method provided by the present invention is shown in FIG. Figure 1 As shown, the method comprises the following steps:

[0045] Step 100: For any destination, obtain the predicted total traffic volume from the traffic input point to the target destination within the predicted time period and the predicted maximum traffic volume per unit time;

[0046] Specifically, in the embodiment of the present invention, in order to achieve optimal link selection and traffic allocation for multiple transit providers and multiple destinations in international long-distance calls, when planning the optimal transit route, it is first necessary to predict the total amount of traffic and the maximum traffic per unit time from the traffic input point to any destination in a certain time period in the future.

[0047] The call input point is the starting point of the planned route, and the destination is the end point of the planned route. There are multiple transit providers between the call input point and each destination, and the transit providers provide call transit services from the call input point to different destinations.

[0048] The time period to be predicted can be set according to actual needs, for example, it can be the next month, the next week, or the next two weeks, etc.

[0049] The total traffic volume refers to the total traffic volume in a certain period of time; the maximum traffic volume per unit time refers to the maximum value of the traffic volume per unit time in a certain period of time. For example, the unit time can be hours, minutes, etc.

[0050] Step 101: Based on a dynamic programming algorithm, according to the predicted total traffic volume, the tariff model of each transit operator, the link capacity requirement of the target destination, and the link capacity limit from each transit operator to the target destination, determine the transit path with the lowest total tariff from the traffic input location to the target destination; wherein the link capacity requirement of the target destination is obtained based on the predicted maximum traffic volume per unit time;

[0051] Specifically, the embodiment of the present invention adopts the idea of ​​global decoupling based on the greedy algorithm, decoupling the planning problem of solving multiple transit providers and multiple destinations into the problem of allocating multiple transit providers to a single destination. If the transit path corresponding to each destination is optimal, the global optimality is achieved. Therefore, after determining the predicted total traffic volume and the predicted maximum traffic volume per unit time corresponding to each destination, the route selection for a single destination can be performed.

[0052] For any destination, a transit path with the lowest total rate from the traffic input point to the destination can be determined based on a dynamic programming algorithm (such as the Viterbi algorithm) according to the predicted total traffic volume of the destination, the rate model of each transit operator, the link capacity requirement of the destination, and the link capacity limit of each transit operator to the destination. The transit path can meet the link capacity requirement of the destination, and the link capacity undertaken by each transit operator in the transit path does not exceed the link capacity limit of each transit operator to the destination.

[0053] Among them, link capacity refers to the number of channels in the link. Links from different transit operators to different destinations have an upper limit on the number of channels, that is, the link capacity limit. The call loss rate, link capacity and unit time traffic can be converted to each other through the Irish (call loss) formula.

[0054] Optionally, after predicting the maximum traffic volume per unit time of a destination, the link capacity required for the destination under the condition of limiting the call loss rate can be obtained based on the limited call loss rate and the Irish formula, that is, the link capacity requirement of the destination. If the link capacity from a single transit operator to the destination cannot meet the link capacity requirement of the destination, the traffic volume from the traffic input point to the destination through one or more transit operators can be increased. The limited call loss rate can be flexibly set according to actual needs and is not limited here.

[0055] It can be understood that each transit operator may have multiple tariff models for different destinations and different traffic volumes, such as the Flat model, the Tier model or the Interval model, etc. Therefore, the tariff for each transit operator to transfer the predicted total traffic volume to the destination can be obtained based on the predicted total traffic volume corresponding to a certain destination and the tariff model of each transit operator, and the transit operator with the lowest tariff is selected to transfer the traffic volume from the traffic input point to the destination. When the transit operator with the lowest tariff cannot meet the link capacity requirement of the destination, for example, the link capacity from the transit operator a with the lowest tariff to the destination is A1, and the link capacity requirement of the destination is M1 (M1>A1), then the remaining link capacity (M1-A1) needs to be transferred by adding more transit operators.

[0056] An optional method for selecting a transit operator is: calculate the ratio of (M1-A1) to M1, multiply the ratio by the predicted total traffic volume corresponding to the destination to obtain the remaining allocated traffic volume, calculate the charges for transiting the remaining allocated traffic volume to the destination by other transit operators except a, and select the transit operator with the lowest charge to transit the remaining link capacity (M1-A1). If the link capacity requirement of the destination still cannot be met, the same selection method can be used to add more transit operators until the sum of the link capacities of multiple transit operators can meet the link capacity requirement of the destination.

[0057] Step 102: Determine the total link capacity requirement of each transit provider based on the transit path with the lowest total tariff corresponding to each destination;

[0058] Specifically, after determining the transit path with the lowest total charge corresponding to each destination, the link capacity required for each transit operator in the transit path with the lowest total charge corresponding to each destination can be obtained, and the total link capacity requirement of each transit operator can be obtained by summing up these link capacities.

[0059] For example, Figure 2 The schematic diagram of the implementation of the routing selection method provided by the present invention is as follows: Figure 2 As shown, the link capacities that transit operator b needs to transfer in the transit paths with the lowest total charges corresponding to destinations 1, 2 and 3 are B1, B2 and B3 respectively, so the total link capacity requirement of transit operator b is B1+B2+B3.

[0060] Step 103: According to the comparison result between the total link capacity requirements of each transit operator and the link capacity of each transit operator, the transit path with the lowest total fee corresponding to each destination is adjusted to obtain the optimal transit path corresponding to each destination.

[0061] Specifically, since the link capacity of each transit operator, that is, the link capacity of the link from the traffic input location to each transit operator, is also limited, then for a certain transit operator, if the total link capacity demand of the transit operator is greater than the link capacity of the transit operator, the link capacity that the transit operator can actually transfer cannot meet the total link capacity demand of the transit operator, and thus cannot meet the link capacity requirements of each destination. Figure 2 As shown, for transit operator b, if B1+B2+B3 is greater than the link capacity B of transit operator b, the link capacity that transit operator b can actually transfer is B, resulting in at least one of the link capacity requirements M1, M2 and M3 of destinations 1, 2 and 3 being unable to be achieved, which in turn causes the system call quality to be unable to be guaranteed.

[0062] Therefore, after determining the total link capacity requirements of each transit operator, the transit path with the lowest total fee corresponding to each destination can be adjusted based on the comparison results between the total link capacity requirements of each transit operator and the link capacity of each transit operator, and finally the optimal transit path corresponding to each destination that takes into account both cost and call quality can be obtained.

[0063] The routing selection method provided in the embodiment of the present invention converts the link planning problem of multiple transit providers with multiple destinations into the selection problem of multiple transit providers with a single destination by means of global decoupling, and uses a dynamic programming algorithm to determine the transit path with the lowest total fee corresponding to each destination, and adjusts the transit path with the lowest total fee corresponding to each destination according to the comparison result between the total link capacity demand of each transit provider and the link capacity of each transit provider, so as to obtain the optimal transit path corresponding to each destination, realize the selection of the transit provider with the best price and the optimal allocation of the links of each transit provider, thereby realizing the guarantee of call quality while controlling the cost, so that the two can reach a controllable balance and optimality.

[0064] Optionally, adjusting the transit path with the lowest total fee corresponding to each destination according to the comparison result between the total link capacity requirements of each transit operator and the link capacity of each transit operator to obtain the optimal transit path corresponding to each destination includes:

[0065] If it is determined that the total link capacity requirements of each transit operator are less than or equal to the link capacity of each transit operator, then the transit path with the lowest total fee corresponding to each destination is determined as the optimal transit path corresponding to each destination; or,

[0066] If it is determined that the total link capacity demand of the first transit operator is greater than the link capacity of the first transit operator, the first path corresponding to the first transit operator that needs to transfer the link capacity is determined, and the link capacity that needs to be transferred on the first path transited by the second transit operator is increased, and all transit operators are traversed until the total link capacity demand of each transit operator is less than or equal to the link capacity of each transit operator, and the optimal transit path corresponding to each destination is obtained;

[0067] Among them, the second transit operator is the transit operator with the lowest link capacity tariff for the link capacity that needs to be transferred on the first transit path among the transit operators whose total link capacity demand is less than the link capacity, and the link capacity that needs to be transferred is the link capacity difference between the total link capacity demand of the first transit operator and the link capacity of the first transit operator.

[0068] Specifically, in an embodiment of the present invention, after determining the total link capacity requirements of each transit operator, the total link capacity requirements of each transit operator and the link capacity of each transit operator can be compared to determine how to adjust the transit path with the lowest total fee corresponding to each destination.

[0069] For example, if it is determined that the total link capacity requirements of each transit operator are less than or equal to the link capacity of each transit operator, it means that the link capacity of each transit operator can meet the link capacity requirements of each destination, and the transit path with the lowest total fee corresponding to each destination can be determined as the optimal transit path corresponding to each destination.

[0070] If it is determined that the total link capacity demand of a transit operator is greater than the link capacity of the transit operator, that is, if it is determined that the total link capacity demand of the first transit operator is greater than the link capacity of the first transit operator, then link capacity transfer is required, the first path corresponding to the first transit operator that requires link capacity transfer is determined, and the link capacity that needs to be transferred on the first path transited by the second transit operator is increased.

[0071] For example, Figure 2 As shown, the link capacities that transit business b needs to transfer in the transit paths with the lowest total charges corresponding to destinations 1, 2 and 3 are B1, B2 and B3 respectively. If B1+B2+B3 is greater than the link capacity B of transit business b, then it is necessary to select one path from the three paths corresponding to destinations 1, 2 and 3 of transit business b. For example, the path corresponding to destination 1 is determined as the path that requires link capacity transfer, and other transit businesses with insufficient capacity (i.e., transit businesses with total link capacity requirements less than link capacity) such as transit business c are added to transfer part of the link capacity from the traffic input location to destination 1, that is, to increase part of the link capacity transferred from transit business b to destination 1 before transit business c comes to transfer. This part of the link capacity is the link capacity B1+B2+B3-B that needs to be transferred. This transit business c is the transit business with the lowest charge required to transfer this part of the link capacity to destination 1 among the transit businesses with insufficient capacity.

[0072] If after adding transit operator c to transfer the above-mentioned part of the link capacity from the input point of traffic to destination 1, the total link capacity demand of transit operator c is greater than the link capacity, then the link capacity of transit operator c can be transferred in the same way as transit operator b mentioned above, and so on, traversing all transit operators until the total link capacity demand of each transit operator is less than or equal to the link capacity of each transit operator, and the transit path with the lowest total charge corresponding to each destination after adjustment is determined as the optimal transit path corresponding to each destination.

[0073] The routing selection method provided by the embodiment of the present invention transfers the link capacity after determining that the total link capacity demand of a transit operator is greater than the link capacity of the transit operator, so that the link capacity of each transit operator can finally meet the link capacity requirement of each destination, and the optimal transit path corresponding to each destination is obtained, so as to achieve the selection of the transit operator with the best price and the optimal allocation of the links of each transit operator, thereby achieving the guarantee of call quality while controlling costs, so that the two can reach a controllable balance and optimality.

[0074] Optionally, the determining a first path corresponding to the first transit operator and requiring link capacity transfer includes:

[0075] For any path corresponding to the first transit operator, respectively obtain the tariff difference of the link capacity that needs to be transferred on the transit target path between the first transit operator and each transit operator whose total link capacity demand is less than the link capacity;

[0076] The path corresponding to the minimum value of each tariff difference is determined as the first path corresponding to the first transit operator and requiring link capacity transfer.

[0077] Specifically, the first path corresponding to the first transit operator and requiring link capacity transfer may be determined in the following manner:

[0078] For any path corresponding to the first transit operator, the tariff difference of the link capacity that needs to be transferred on the transit target path between the first transit operator and each transit operator whose total link capacity demand is less than the link capacity is obtained respectively, and then all the tariff differences are compared, and the path corresponding to the minimum value is determined as the first path corresponding to the first transit operator that needs link capacity transfer, and the transit operator corresponding to the minimum value is the above-mentioned second transit operator.

[0079] For example, Figure 2 As shown in the figure, the link capacities that transit business b needs to transfer in the transit paths with the lowest total charges corresponding to destinations 1, 2 and 3 are B1, B2 and B3 respectively. If B1+B2+B3 is greater than the link capacity B of transit business b, it is necessary to select one path from the three paths corresponding to destinations 1, 2 and 3 of transit business b to transfer the link capacity. At this time, transit businesses such as c, d and e with insufficient capacity are searched, and the charges for the link capacities that transit business b, c, d and e need to transfer to destination 1 are b1, c1, d1 and e1 respectively, the charges for the link capacities that transit business b, c, d and e need to transfer to destination 2 are b2, c2, d2 and e2 respectively, and the charges for the link capacities that transit business b, c, d and e need to transfer to destination 3 are b3, c3, d3 and e1 respectively. 3. Take the path corresponding to min{(b1-c1),(b1-d1),(b1-e1),(b2-c2),(b2-d2),(b2-e2),(b3-c3),(b3-d3),(b3-e3)} as the path corresponding to transit operator b that requires link capacity transfer. If min{(b1-c1),(b1-d1),(b1-e1),(b2-c2),(b2-d2),(b2-e2),(b3-c3),(b3-d3),(b3-e3)} is (b1-c1), it means that the path corresponding to destination 1 of transit operator b is the path that requires link capacity transfer, and transit operator c is the transit operator with the lowest fee for transferring the link capacity to destination 1 among the transit operators with insufficient capacity.

[0080] Optionally, obtaining the predicted total traffic volume and predicted maximum traffic volume per unit time from the traffic input location to the target destination within the time period to be predicted includes:

[0081] Based on the ARMA model or LSTM model, the predicted total traffic volume from the traffic input point to the target destination within the predicted time period and the predicted maximum traffic volume per unit time are obtained;

[0082] Among them, the ARMA model or LSTM model is obtained after training based on historical data samples of total traffic volume and maximum traffic volume per unit time with certain training labels, and the training labels are the true values ​​of the total traffic volume and maximum traffic volume per unit time in the time period predicted by the training samples.

[0083] Specifically, in the embodiment of the present invention, an ARMA model (Autoregressive Moving Average model) or an LSTM (Long Short-Term Memory) model can be used to predict the total traffic volume and the maximum traffic volume per unit time. Among them, the LSTM used to predict the total traffic volume and the maximum traffic volume per unit time has been trained based on the historical data samples of the total traffic volume and the maximum traffic volume per unit time with a certain training label, and the training label is the true value of the total traffic volume and the maximum traffic volume per unit time in the time period predicted by the training sample, so that accurate prediction of the total traffic volume and the maximum traffic volume per unit time can be achieved.

[0084] Optionally, the obtaining of the predicted total traffic volume and the predicted maximum traffic volume per unit time from the traffic input location to the target destination within the time period to be predicted based on the ARMA model or the LSTM model includes:

[0085] If it is determined that the historical data series of the total traffic volume and the maximum traffic volume per unit time are stationary time series, then based on the ARMA model, the predicted total traffic volume and the predicted maximum traffic volume per unit time from the traffic input point to the target destination within the time period to be predicted are obtained; or,

[0086] If it is determined that the historical data series of the total traffic volume and the maximum traffic volume per unit time are non-stationary time series, the predicted total traffic volume and the predicted maximum traffic volume per unit time from the traffic input point to the target destination within the predicted time period are obtained based on the LSTM model.

[0087] Specifically, in an embodiment of the present invention, before predicting the total traffic volume and the maximum traffic volume per unit time, it is possible to determine whether the historical data series of the total traffic volume and the maximum traffic volume per unit time corresponding to each destination is a stationary time series after data cleaning. If it is a stationary time series, an ARMA model with better accuracy and fitting effect can be selected; if it is a non-stationary time series, LSTM can be selected as the prediction model for fitting, which ensures the prediction accuracy while optimizing the resources occupied and time consumed by the algorithm.

[0088] Optionally, whether the data series is a stationary time series can be determined based on an ADF (Augmented Dickey-Fuller) unit root test.

[0089] Optionally, the above-mentioned ARMA model can be an ARMA model that estimates model parameters based on the L-BFGS (Limited-memory BFGS) algorithm. The traditional ARMA model mainly uses the Newton method for parameter estimation. Although the Newton method has a fast convergence speed, each iteration requires the calculation of the inverse matrix of the Hessian Matrix, which is complex to calculate, and sometimes the Hessian matrix of the objective function cannot remain positive, which makes the Newton method invalid. By adopting the L-BSFG algorithm for parameter estimation, the unit matrix is ​​gradually approximated to the Hessian matrix, and calculations can be performed in limited memory, which simplifies the complexity of the algorithm and can avoid the failure of parameter estimation.

[0090] Figure 3 The present invention provides a flow chart for predicting the total volume of traffic and the maximum volume of traffic per unit time, as shown in FIG. Figure 3 As shown in the figure, an independent traffic forecasting model is built at a single destination, and its stability is determined after data cleaning. When the data series is a stationary time series, the ARMA model based on the L-BFGS algorithm is used for forecasting. When the data series is a non-stationary time series, the LSTM deep learning model is used for forecasting, thereby solving the problem of occupying more resources and large amount of calculation in multiple LSTM scenarios. Under the premise of high prediction accuracy, the calculation efficiency is faster.

[0091] The routing selection device provided by the present invention is described below. The routing selection device described below and the routing selection method described above can be referred to each other.

[0092] Figure 4 A schematic diagram of the structure of the routing selection device provided by the present invention, such as Figure 4 As shown, the device comprises:

[0093] The acquisition module 400 is used to acquire, for any destination, the predicted total traffic volume from the traffic input point to the target destination within the predicted time period and the predicted maximum traffic volume per unit time;

[0094] The first determination module 410 is used to determine the transit path with the lowest total tariff from the traffic input location to the target destination based on the predicted total traffic volume, the tariff model of each transit operator, the link capacity requirement of the target destination, and the link capacity limit of each transit operator to the target destination based on a dynamic programming algorithm; wherein the link capacity requirement of the target destination is obtained based on the predicted maximum traffic volume per unit time;

[0095] A second determination module 420 is used to determine the total link capacity requirement of each transit provider based on the transit path with the lowest total tariff corresponding to each destination;

[0096] The adjustment module 430 is used to adjust the transit path with the lowest total fee corresponding to each destination according to the comparison result between the total link capacity requirements of each transit operator and the link capacity of each transit operator, so as to obtain the optimal transit path corresponding to each destination.

[0097] Optionally, the adjustment module 430 is used to: if it is determined that the total link capacity requirements of each transit operator are less than or equal to the link capacity of each transit operator, then determine the transit path with the lowest total fee corresponding to each destination as the optimal transit path corresponding to each destination; or, if it is determined that the total link capacity requirement of the first transit operator is greater than the link capacity of the first transit operator, then determine the first path corresponding to the first transit operator that requires link capacity transfer, and increase the link capacity that needs to be transferred on the first path of transit by the second transit operator, traverse all transit operators until the total link capacity requirements of each transit operator are less than or equal to the link capacity of each transit operator, and obtain the optimal transit path corresponding to each destination; wherein the second transit operator is a transit operator with the lowest fee for link capacity that needs to be transferred on the first transit path among the transit operators whose total link capacity requirements are less than the link capacity, and the link capacity that needs to be transferred is the link capacity difference between the total link capacity requirement of the first transit operator and the link capacity of the first transit operator.

[0098] Optionally, determining the first path corresponding to the first transit operator that requires link capacity transfer includes: for any path corresponding to the first transit operator, respectively obtaining the tariff difference of the link capacity that needs to be transferred on the transit target path between the first transit operator and each transit operator whose total link capacity demand is less than the link capacity; and determining the path corresponding to the minimum value of each tariff difference as the first path corresponding to the first transit operator that requires link capacity transfer.

[0099] Optionally, the acquisition module 400 is used to: obtain the predicted total traffic volume and the predicted maximum traffic volume per unit time from the traffic input point to the target destination within the predicted time period based on an ARMA model or an LSTM model; wherein the ARMA model or the LSTM model is obtained after training based on historical data samples of the total traffic volume and the maximum traffic volume per unit time with certain training labels, and the training labels are the true values ​​of the total traffic volume and the maximum traffic volume per unit time within the time period predicted by the training samples.

[0100] Optionally, the method of obtaining the predicted total traffic volume and the predicted maximum traffic volume per unit time from the traffic input point to the target destination within the time period to be predicted based on the ARMA model or the LSTM model includes: if the historical data series of the total traffic volume and the maximum traffic volume per unit time is determined to be a stationary time series, then based on the ARMA model, obtaining the predicted total traffic volume and the predicted maximum traffic volume per unit time from the traffic input point to the target destination within the time period to be predicted; or, if the historical data series of the total traffic volume and the maximum traffic volume per unit time is determined to be a non-stationary time series, then based on the LSTM model, obtaining the predicted total traffic volume and the predicted maximum traffic volume per unit time from the traffic input point to the target destination within the time period to be predicted.

[0101] Optionally, the ARMA model is an ARMA model that estimates model parameters based on an L-BFGS algorithm.

[0102] Optionally, the link capacity requirement of the target destination is obtained based on the predicted maximum traffic volume per unit time, including: based on the Erlang formula and the limited call loss rate, converting the predicted maximum traffic volume per unit time into the link capacity requirement of the target destination.

[0103] It should be noted here that the above-mentioned device provided by the present invention can implement all the method steps implemented by the above-mentioned method embodiment, and can achieve the same technical effect. The parts and beneficial effects that are the same as the method embodiment in this embodiment will not be described in detail here.

[0104] Figure 5 A schematic diagram of the structure of an electronic device provided by the present invention, such as Figure 5As shown, the electronic device may include: a processor (processor) 510, a communication interface (Communication Interface) 520, a memory (memory) 530 and a communication bus 540, wherein the processor 510, the communication interface 520, and the memory 530 communicate with each other through the communication bus 540. The processor 510 can call the computer program in the memory 530 to execute the steps of any of the routing selection methods provided in the above embodiments, for example, including: for any destination, obtaining the predicted total traffic volume and the predicted maximum traffic volume per unit time from the traffic input location to the target destination within the predicted time period; based on the dynamic programming algorithm, according to the predicted total traffic volume, the tariff model of each transit operator, the link capacity requirement of the target destination, and the link capacity limit of each transit operator to the target destination, determine the transit path with the lowest total tariff from the traffic input location to the target destination; wherein the link capacity requirement of the target destination is obtained based on the predicted maximum traffic volume per unit time; based on the transit path with the lowest total tariff corresponding to each destination, determine the total link capacity requirement of each transit operator; according to the comparison result between the total link capacity requirement of each transit operator and the link capacity of each transit operator, adjust the transit path with the lowest total tariff corresponding to each destination to obtain the optimal transit path corresponding to each destination.

[0105] In addition, the logic instructions in the above-mentioned memory 530 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0106] On the other hand, the present invention also provides a computer program product, which includes a computer program stored on a non-transitory computer-readable storage medium, and the computer program includes program instructions. When the program instructions are executed by a computer, the computer can execute the steps of any of the routing selection methods provided in the above embodiments, for example, including: for any destination, obtaining the predicted total traffic volume and the predicted maximum traffic volume per unit time from the traffic input point to the target destination within the predicted time period; based on a dynamic programming algorithm, according to the predicted total traffic volume, the tariff model of each transit operator, the link capacity requirement of the target destination, and the link capacity limit of each transit operator to the target destination, determine the transit path with the lowest total tariff from the traffic input point to the target destination; wherein the link capacity requirement of the target destination is obtained based on the predicted maximum traffic volume per unit time; based on the transit path with the lowest total tariff corresponding to each destination, determine the total link capacity requirement of each transit operator; according to the comparison result between the total link capacity requirement of each transit operator and the link capacity of each transit operator, adjust the transit path with the lowest total tariff corresponding to each destination to obtain the optimal transit path corresponding to each destination.

[0107] On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of any of the routing selection methods provided in the above-mentioned embodiments, for example, including: for any destination, obtaining the predicted total traffic volume and the predicted maximum traffic volume per unit time from the traffic input point to the target destination within the predicted time period; based on a dynamic programming algorithm, determining the transit path with the lowest total tariff from the traffic input point to the target destination according to the predicted total traffic volume, the tariff model of each transit operator, the link capacity requirement of the target destination, and the link capacity limit of each transit operator to the target destination; wherein the link capacity requirement of the target destination is obtained based on the predicted maximum traffic volume per unit time; determining the total link capacity requirement of each transit operator based on the transit path with the lowest total tariff corresponding to each destination; adjusting the transit path with the lowest total tariff corresponding to each destination according to the comparison result between the total link capacity requirement of each transit operator and the link capacity of each transit operator, and obtaining the optimal transit path corresponding to each destination.

[0108] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0109] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0110] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A routing selection method, characterized in that: include: For any destination, obtain the predicted total traffic volume from the traffic input point to the target destination within the predicted time period and the predicted maximum traffic volume per unit time; Based on a dynamic programming algorithm, according to the predicted total traffic volume, the tariff model of each transit provider, the link capacity requirement of the target destination, and the link capacity limit of each transit provider to the target destination, a transit path with the lowest total tariff from the traffic input location to the target destination is determined; wherein the link capacity requirement of the target destination is obtained based on the predicted maximum traffic volume per unit time; Determine the total link capacity requirement of each transit provider based on the transit path with the lowest total tariff corresponding to each destination; According to the comparison result between the total link capacity requirements of each transit operator and the link capacity of each transit operator, the transit path with the lowest total fee corresponding to each destination is adjusted to obtain the optimal transit path corresponding to each destination; The step of adjusting the transit path with the lowest total fee corresponding to each destination according to the comparison result between the total link capacity requirements of each transit operator and the link capacity of each transit operator to obtain the optimal transit path corresponding to each destination includes: If it is determined that the total link capacity requirements of each transit operator are less than or equal to the link capacity of each transit operator, then the transit path with the lowest total fee corresponding to each destination is determined as the optimal transit path corresponding to each destination; or, If it is determined that the total link capacity demand of the first transit operator is greater than the link capacity of the first transit operator, then determine the first path corresponding to the first transit operator that needs to transfer link capacity, and increase the link capacity that needs to be transferred on the first path transferred by the second transit operator, traverse all transit operators until the total link capacity demand of each transit operator is less than or equal to the link capacity of each transit operator, and obtain the optimal transit path corresponding to each destination; Among them, the second transit operator is the transit operator with the lowest fee for transiting the link capacity that needs to be transferred on the first path among the transit operators whose total link capacity demand is less than the link capacity, and the link capacity that needs to be transferred is the link capacity difference between the total link capacity demand of the first transit operator and the link capacity of the first transit operator.

2. The routing selection method according to claim 1, characterized in that: The determining a first path corresponding to the first transit operator and requiring link capacity transfer includes: For any path corresponding to the first transit operator, respectively obtaining the tariff difference of the link capacity that needs to be transferred on the transit target path between the first transit operator and each transit operator whose total link capacity demand is less than the link capacity; The path corresponding to the minimum value of each tariff difference is determined as the first path corresponding to the first transit operator that requires link capacity transfer.

3. The routing selection method according to claim 1, characterized in that: The step of obtaining the predicted total traffic volume from the traffic input location to the target destination within the predicted time period and the predicted maximum traffic volume per unit time includes: Based on the ARMA model or LSTM model, the predicted total traffic volume from the traffic input point to the target destination within the predicted time period and the predicted maximum traffic volume per unit time are obtained; The ARMA model or the LSTM model is obtained after training based on historical data samples of the total amount of traffic and the maximum amount of traffic per unit time with certain training labels, and the training labels are the true values ​​of the total amount of traffic and the maximum amount of traffic per unit time in the time period predicted by the training samples.

4. The routing selection method according to claim 3, characterized in that: The method of obtaining the predicted total traffic volume and the predicted maximum traffic volume per unit time from the traffic input location to the target destination within the predicted time period based on the ARMA model or the LSTM model includes: If it is determined that the historical data series of the total traffic volume and the maximum traffic volume per unit time are stationary time series, then based on the ARMA model, the predicted total traffic volume and the predicted maximum traffic volume per unit time from the traffic input point to the target destination within the time period to be predicted are obtained; or, If it is determined that the historical data series of the total traffic volume and the maximum traffic volume per unit time are non-stationary time series, the predicted total traffic volume and the predicted maximum traffic volume per unit time from the traffic input point to the target destination within the predicted time period are obtained based on the LSTM model.

5. The routing selection method according to claim 3 or 4, characterized in that: The ARMA model is an ARMA model that estimates model parameters based on the L-BFGS algorithm.

6. The routing selection method according to claim 1, characterized in that: The link capacity requirement of the target destination is obtained based on the predicted maximum traffic volume per unit time, including: Based on the Erlang formula and the limited call loss rate, the predicted maximum traffic volume per unit time is converted into the link capacity requirement of the target destination.

7. A routing selection device, characterized in that: include: An acquisition module, for acquiring, for any destination, the predicted total traffic volume from the traffic input point to the target destination within the predicted time period and the predicted maximum traffic volume per unit time; A first determination module is used to determine, based on a dynamic programming algorithm, a transit path with the lowest total tariff from the traffic input location to the target destination according to the predicted total traffic volume, the tariff model of each transit operator, the link capacity requirement of the target destination, and the link capacity limit of each transit operator to the target destination; wherein the link capacity requirement of the target destination is obtained based on the predicted maximum traffic volume per unit time; A second determination module is used to determine the total link capacity requirement of each transit provider based on the transit path with the lowest total tariff corresponding to each destination; An adjustment module, configured to adjust the transit path with the lowest total fee corresponding to each destination according to the comparison result between the total link capacity requirements of each transit operator and the link capacity of each transit operator, so as to obtain the optimal transit path corresponding to each destination; The step of adjusting the transit path with the lowest total fee corresponding to each destination according to the comparison result between the total link capacity requirements of each transit operator and the link capacity of each transit operator to obtain the optimal transit path corresponding to each destination includes: If it is determined that the total link capacity requirements of each transit operator are less than or equal to the link capacity of each transit operator, then the transit path with the lowest total fee corresponding to each destination is determined as the optimal transit path corresponding to each destination; or, If it is determined that the total link capacity demand of the first transit operator is greater than the link capacity of the first transit operator, then determine the first path corresponding to the first transit operator that needs to transfer link capacity, and increase the link capacity that needs to be transferred on the first path transferred by the second transit operator, traverse all transit operators until the total link capacity demand of each transit operator is less than or equal to the link capacity of each transit operator, and obtain the optimal transit path corresponding to each destination; Among them, the second transit operator is the transit operator with the lowest fee for transiting the link capacity that needs to be transferred on the first path among the transit operators whose total link capacity demand is less than the link capacity, and the link capacity that needs to be transferred is the link capacity difference between the total link capacity demand of the first transit operator and the link capacity of the first transit operator.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the steps of the routing selection method according to any one of claims 1 to 6 are implemented.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the routing method according to any one of claims 1 to 6 are implemented.

Images