Multipath data transmission methods, electronic devices and storage media

Abstract

This application provides a multi-path data transmission method, electronic device, and storage medium. It acquires communication parameters for each path in a multi-path transmission scheme and determines the replication probability of the data packet to be transmitted based on these parameters and a preset multi-armed slot machine model. This allows for the transmission of part or all of the data in the data packet across multiple paths according to the replication probability, thus saving transmission latency. In the multi-path transmission scheme, all data in the data packet is transmitted through the path with the minimum transmission latency. In each path other than the path with the minimum transmission latency, a portion of the data in the data packet is replicated according to the replication probability and transmitted. In other words, all data in the data packet is transmitted through the path with the minimum transmission latency, while the other paths replicate the portion of the data corresponding to the replication probability. This reduces the retransmission rate, decreases transmission latency, and increases the transmission rate.

Description

Technical Field

[0001] This application relates to the field of communication technology, and in particular to a multipath data transmission method, electronic device, and storage medium. Background Technology

[0002] Multipath QUIC (Quick UDP Internet Connections) has advantages in bandwidth aggregation and fault tolerance, but in disaster scenarios, the high heterogeneity and rapid time-varying nature of the links can induce out-of-order buffer expansion and head-of-line congestion, resulting in decreased throughput and increased latency jitter, causing unstable data transmission and leading to a high data packet loss rate. In order to improve data transmission performance, a more stable data transmission method is needed. Summary of the Invention

[0003] In view of this, the purpose of this application is to propose a multipath data transmission method, electronic device and storage medium to solve the problem of high data packet loss rate caused by multipath data transmission in the prior art.

[0004] To achieve the above objectives, this application provides a multipath data transmission method, comprising: Retrieve communication parameters for each path in multiple paths; Based on the communication parameters and the preset multi-armed slot machine model, the probability of copying and transmitting the data packet to be transmitted is determined; In multiple paths, all data in the data packet to be transmitted is transmitted through the path with the least transmission delay. In each path other than the path with the least transmission delay, a portion of the data in the data packet to be transmitted is transmitted according to the replication transmission probability.

[0005] Optionally, the path with the shortest transmission delay among multiple paths transmits all data in the data packet to be transmitted. In all paths other than the path with the shortest transmission delay, a portion of the data in the data packet to be transmitted is transmitted according to the replication transmission probability. Following this, the process further includes: Calculate the reward value of the replication and transmission probability based on the set reward function; The model parameters of the multi-armed slot machine model are updated based on the reward value.

[0006] Optionally, determining the replication probability of the data packet to be transmitted based on the communication parameters and a preset multi-armed slot machine model includes: The communication parameters are preprocessed, and the preprocessed communication parameters are input into the preset multi-armed slot machine model. The replication transmission probability is output through the multi-armed slot machine model.

[0007] Optionally, the preprocessing of the communication parameters includes: The initial parameters are obtained by filling in missing values ​​for each parameter in the communication parameters. The initial parameters are normalized to obtain normalized parameters; All normalized parameters are concatenated to obtain the preprocessed communication parameters.

[0008] Optionally, the method for constructing the reward function includes: The reward function is constructed with the objectives of minimizing transmission latency and maximizing network throughput, and is described by the following formula: ; in, For transmission throughput; This is the average delay; Packet loss rate; This represents the penalty coefficient for packet loss on overall transmission performance. This indicates the relative weights of throughput and latency in the overall utility. , .

[0009] Optionally, the model parameters include: statistics of the performed actions and posterior distribution parameters; The step of updating the model parameters of the multi-armed slot machine model based on the reward value includes: Based on the reward value, the statistics of the executed action are updated to obtain the updated statistics; Based on the updated statistics, the posterior distribution parameters are updated to obtain the updated posterior distribution parameters.

[0010] Optionally, the statistics of the executed action include the cumulative action reward vector and the cumulative action context matrix; Based on the reward value, the statistics of the executed action are updated to obtain the updated statistics, including: The action cumulative context matrix is ​​incrementally updated using the following formula: ; in, For action Action cumulative context matrix, For the current context vector, This is the transpose of the context vector; Indicates assignment / incremental update; The cumulative reward vector for the action is incrementally updated using the following formula: ; in, For action The cumulative reward vector for actions, with dimension [missing information]. , As a reward value, This is the current context vector.

[0011] Optionally, the posterior distribution parameters include: posterior mean and posterior covariance matrix; The step of updating the posterior distribution parameters based on the updated statistics to obtain the updated posterior distribution parameters includes: The posterior mean is updated using the following formula: ; in, This is the updated posterior mean vector. Accumulate the inverse of the context matrix for the action. action The cumulative reward vector for actions; The posterior covariance matrix is ​​updated using the following formula: ; in, This is the updated posterior covariance matrix; For action Action cumulative context matrix, This represents finding the inverse of a matrix.

[0012] Based on the same inventive concept, this application also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable by the processor, wherein the processor implements the method described above when executing the computer program.

[0013] Based on the same inventive concept, this application also provides a non-transitory computer-readable storage medium that stores computer instructions for causing a computer to perform the method described above.

[0014] As described above, the multi-path data transmission method, electronic device, and storage medium provided in this application acquire communication parameters for each path in multiple paths. A communication reference represents the communication state of each path. Based on the communication parameters and a preset multi-armed slot machine model, the replication transmission probability of the data packet to be transmitted is determined, that is, the replication transmission probability of the data packet to be transmitted in each path is determined based on the communication state. Therefore, partial and complete data of the data packet to be transmitted can be transmitted in multiple paths according to the replication transmission probability, saving transmission latency. In multiple paths, all data in the data packet to be transmitted is transmitted through the path with the minimum transmission latency. In each path other than the path with the minimum transmission latency, a portion of the data packet to be transmitted is copied according to the replication transmission probability and transmitted. That is, all data of the data packet to be transmitted is transmitted through the path with the minimum transmission latency, while the other paths copy the portion of the data packet corresponding to the replication transmission probability. This reduces the retransmission rate, reduces transmission latency, and increases the transmission rate. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in this application or related technologies, the drawings used in the description of the embodiments or related technologies will be briefly introduced below. Obviously, the drawings described below are only embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 This is a schematic flowchart of a multipath data transmission method according to an embodiment of this application; Figure 2 This is a schematic diagram of the multi-armed slot machine model update method according to an embodiment of this application; Figure 3 This is a schematic diagram of a multipath data transmission device according to an embodiment of this application; Figure 4 This is a schematic diagram of the multipath data structure in an embodiment of this application; Figure 5 This is a schematic diagram of the hardware structure of an electronic device according to an embodiment of this application. Detailed Implementation

[0017] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with specific embodiments and the accompanying drawings.

[0018] It should be noted that, unless otherwise defined, the technical or scientific terms used in the embodiments of this application should have the ordinary meaning understood by one of ordinary skill in the art to which this application pertains. The terms "first," "second," and similar terms used in the embodiments of this application do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Terms such as "comprising" or "including" mean that the element or object preceding the word encompasses the elements or objects listed after the word and their equivalents, without excluding other elements or objects. Terms such as "connected" or "linked" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. Terms such as "upper," "lower," "left," and "right" are only used to indicate relative positional relationships; when the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0019] As described in the background section, in disaster scenarios such as earthquakes, floods, and forest fires, communication infrastructure often suffers damage, power outages, or congestion, making it impossible for a single path to reliably carry critical data from the field to the command center. To quickly restore coverage, emergency systems typically utilize residual ground base stations, low-altitude UAV relays, satellite links, and backpack device hotspots to form a temporary heterogeneous multipath network. In multipath transmission, traditional transmission methods such as minimum latency priority, polling, or fixed redundancy tend to rely on preset, static parameter configurations, making it difficult to balance efficiency and stability at different stages of transmission.

[0020] To solve the above technical problems, such as Figure 1 As shown, this application provides a multipath data transmission method, including the following steps: Step 102: Obtain the communication parameters of each path in the multiple paths.

[0021] In this step, communication parameters include round-trip time, effective throughput, packet loss rate, congestion window, and jitter parameters. The communication parameters for each path represent the communication status of data transmission along that path. By using the communication parameters for each path, the probability of the data to be transmitted being transmitted in each path can be determined.

[0022] Step 104: Determine the probability of copying and transmitting the data packet to be transmitted based on the communication parameters and the preset multi-armed slot machine model.

[0023] In this step, communication parameters are used as input to a preset multi-armed slot machine model. The model outputs the probability of data packet duplication based on the communication parameters of each path. This can be understood as determining the probability of data packet duplication based on the communication status of each path. When the communication status of a transmission path is good, or when the communication status of all paths is good, the probability of data packet duplication is low. When the communication status of all transmission paths is poor, the probability of data packet duplication is increased to avoid packet loss. This increases the data transmission rate and reduces transmission latency. Step 106: In multiple paths, transmit all data in the data packet to be transmitted through the path with the least transmission delay, and in each path other than the path with the least transmission delay, transmit a portion of the data packet to be transmitted according to the replication transmission probability.

[0024] In this step, such as Figure 4 As shown, the exemplary multiple paths can be four paths. These four paths include a satellite transmission path, a drone transmission path, a base station transmission path, and a backpack device transmission path. One of the four paths is pre-set to have the minimum transmission latency. For example, the satellite transmission path is set to have the minimum transmission latency. All data in the data packet to be transmitted is transmitted through the path with the minimum transmission latency. In each path other than the path with the minimum transmission latency, a portion of the data in the data packet corresponding to the replication probability is replicated and transmitted. This reduces transmission latency, thereby reducing packet loss and increasing transmission rate.

[0025] Steps 102-106 obtain the communication parameters for each path in the multiple paths. The communication references represent the communication status of each path. Based on the communication parameters and a preset multi-armed slot machine model, the replication transmission probability of the data packet to be transmitted is determined, that is, the replication transmission probability of the data packet in each path is determined based on the communication status. Therefore, partial or complete data of the data packet to be transmitted can be transmitted in multiple paths according to the replication transmission probability, saving transmission latency. In the multiple paths, all data in the data packet to be transmitted is transmitted through the path with the minimum transmission latency. In each path other than the path with the minimum transmission latency, a portion of the data packet to be transmitted is copied according to the replication transmission probability and transmitted. In other words, all data of the data packet to be transmitted is transmitted through the path with the minimum transmission latency, while the other paths copy the portion of the data packet corresponding to the replication transmission probability. This reduces transmission latency, thereby reducing packet loss rate and increasing transmission rate.

[0026] In some embodiments, the path with the shortest transmission latency among multiple paths transmits all data in the data packet to be transmitted, and in all paths other than the path with the shortest transmission latency, a portion of the data in the data packet to be transmitted is transmitted according to the replication transmission probability. Then, the method further includes: Calculate the reward value of the replication and transmission probability based on the set reward function; The model parameters of the multi-armed slot machine model are updated based on the reward value.

[0027] Specifically, the reward function is constructed with the goal of minimizing transmission latency and maximizing network throughput. That is, the reward function is built based on communication parameters. After each path transmits the data packet to be transmitted, its communication parameters continuously change, and the reward function is continuously updated. Thus, the reward value calculated based on this reward function for the replication transmission probability is also continuously updated. Therefore, the model parameters of the multi-armed slot machine model are updated according to the reward value, enabling the multi-armed slot machine model to output a better replication transmission probability for the data packet to be transmitted in each path. In this embodiment, the update cycle of the multi-armed slot machine model is set to perform a model parameter update every 5 transmission delays. Inference and transmission still proceed as usual in each cycle, but the replication transmission probability is only updated using the multi-armed slot machine model (LinTS) parameters after accumulating a reward of 5 transmission delays. This allows the end-to-end impact (throughput, latency, packet loss) of the previous stage to be observed and averaged within a more complete time window. This filters out instantaneous noise and reduces the disturbance effect of instantaneous reward fluctuations on parameter estimation. Updating the model every 5 transmission delays achieves slower updates and avoids frequent fluctuations. Updating for every data packet would result in an excessively high learning frequency, causing the residual rate to fluctuate frequently within a very short time frame, leading to high-frequency jitter. By using 5 transmission delays as the timer, a fast response at the transmission delay level is maintained (the reward is updated every time, allowing for small adjustments), while keeping model updates at a lower frequency to avoid excessively frequent switching and copying transmissions. This achieves a more controllable engineering compromise between stability and agility.

[0028] Specifically, the method for constructing the reward function includes: The reward function is constructed with the objectives of minimizing transmission latency and maximizing network throughput, and is described by the following formula: ; in, For transmission throughput; This is the average delay; Packet loss rate; This represents the penalty coefficient for packet loss on overall transmission performance. This indicates the relative weights of throughput and latency in the overall utility. , The reward function is constructed with the goal of minimizing transmission latency and maximizing network throughput, aiming to achieve a controllable balance between efficient transmission and low latency.

[0029] In some embodiments, determining the replication probability of the data packet to be transmitted based on the communication parameters and a preset multi-armed slot machine model includes: The communication parameters are preprocessed, and the preprocessed communication parameters are input into the preset multi-armed slot machine model. The replication transmission probability is output through the multi-armed slot machine model.

[0030] Specifically, since the multi-armed slot machine model uses context vector format as input, the communication parameters are preprocessed to conform to the input format of the preset multi-armed slot machine model. The preprocessed communication parameters are then input into the preset multi-armed slot machine model, which outputs the replication transmission probability based on these parameters. In other words, the preset multi-armed slot machine model outputs the replication transmission probability of the data packet to be transmitted in each path based on the communication status of each path, thereby achieving minimum latency, maximum throughput, and minimum packet loss rate when the data packet is transmitted across all paths, thus improving data transmission performance.

[0031] In some embodiments, the preprocessing of the communication parameters includes: Missing values ​​are filled in for each parameter in the communication parameters to obtain the initial parameters.

[0032] Specifically, communication parameters include transmission time delay (RTT), effective throughput (Thru), packet loss rate (Loss), congestion window (cwnd), and jitter. A recent valid value completion algorithm is used to fill in missing values ​​in the communication parameters, replacing missing values ​​with previous values ​​to complete the data. The completed communication parameters are then used as the initial parameters.

[0033] The initial parameters are normalized to obtain normalized parameters.

[0034] Specifically, to facilitate direct use in the LinTS (Large Armed Slots) model, core metrics of different natures, such as transmission delay (RTT), effective throughput (Thru), packet loss rate (Loss), congestion window (cwnd), and jitter, are normalized and their direction is kept consistent. Min-Max normalization is applied to all parameters to normalize them to a unified [0, 1] interval. Then, for negative metrics such as delay and packet loss rate (smaller values ​​indicate better performance), a subtraction transformation is used to convert them into positive metrics; for positive metrics such as throughput and available capacity (larger values ​​indicate better performance), the original direction is maintained. The formulas for direction unification and normalization are: ; ; ; ; ; in: This is the current round-trip delay measurement for path i (in milliseconds). The minimum delay observed; The maximum observed time delay; The RTT value after normalization and direction processing. and It is dynamically maintained and updated during system operation. The normalization method for other indicators is the same, all using their respective historical minimum / maximum ranges for mapping. For example... This represents the effective throughput of path i within the current measurement period, signifying the actual data transmission capacity of that path. and These represent the system's historical minimum and maximum throughput, respectively. This is the normalized throughput value; the larger the value, the stronger the transmission capacity of the path. The packet loss rate of path i measured in the current measurement period represents the transmission reliability of that path. and These represent the system's historical minimum and maximum packet loss rates, respectively. The loss value is the normalized value with orientation adjustment. This represents the congestion window measured for path i within the current measurement period, reflecting the immediate available capacity of that path. and These are the historical minimum and maximum congestion window values ​​observed by the system, respectively. This is the normalized congestion window value; a larger value indicates more available capacity for the path. The delay jitter measured for path i within the current measurement period characterizes the link stability. and These are the historical minimum and maximum jitter values ​​observed by the system, respectively. The jitter parameter value is normalized and directional processed.

[0035] All normalized parameters are concatenated to obtain the preprocessed communication parameters.

[0036] Specifically, transmission time delay (RTT), effective throughput (Thru), packet loss rate (Loss), congestion window (cwnd), and jitter parameter are concatenated into a fixed-length context vector. (Including constant term 1). For example, for the four paths (corresponding to ground base stations, drones, satellites, and backpack equipment in a disaster scenario), a fixed-length context vector is constructed. For each path i, the following five normalized features are extracted: RTT, Thru, Loss, Cwnd, and Jitter. An example of constructing a context vector for decision-making based on the normalized indicators is shown below:

[0037] Wherein: Items 1–20 are the normalized features of the four paths, each path containing the following five metrics: RTT, Thru, Loss, Cwnd, and Jitter; 1 is a constant term. This represents the number of currently active paths.

[0038] When the actual number of active paths In this case, all five normalized features (RTT, Thru, Loss, Cwnd, Jitter) of the empty path are filled with 0. This setting ensures a fixed input dimension and stable structure, facilitating online linear parameter learning by algorithms such as LinTS.

[0039] For example, the processing procedure for a multi-armed slot machine model is as follows: For example, LinTS action definition: Actions are pre-defined for the multi-armed slot machine model, and the action set is as follows: ; Here, "action" represents the level corresponding to the probability of copying data transmission, and each action in the action set corresponds to a probability of copying data transmission. This refers to the number of gears (gears), such as the default 5 gears. ,at this time It can be configured to more. That is, in the probability of replicating data transmission ( All data packets to be transmitted are copied along all paths) and there is no redundancy ( The system discretizes the selection space between the paths (where data packets to be transmitted are only transmitted on the path with the minimum transmission delay, and other paths are not copied, i.e., no transmission occurs). In each cycle, the system selects an action and learns and updates it based on the transmission performance reward it brings. In other words, the multi-armed machine selects a probability of replicating data transmission in each update cycle, and then transmits data according to that probability in that cycle. Specifically, this is based on the preprocessed communication parameters described above. As context input, at time ,make This indicates that the input parameters belong to all context vectors in the above embodiments.

[0040] LinTS Motion Selection: Each Motion Maintain two sets of statistics: Its initial value is set to .

[0041] in: For action (gear), The dimension of the context vector. For action The cumulative reward vector, with dimension . , It is a d-dimensional identity matrix; This is the regularization coefficient.

[0042] Set up a cold start and exploration mechanism, and in the initial stage, select each tier at least once in sequence to ensure that each arm of the multi-armed slot machine model receives the initial reward.

[0043] Within each period, the posterior parameters are first calculated: ;

[0044] in, The sampling scale (exploration intensity coefficient) controls the degree of exploration; For action The sampling parameter vector, It is the inverse of the cumulative context matrix; For action The posterior covariance matrix. If drastic network fluctuations are detected, the amplification can be dynamically increased. This enhances the exploration level and prevents premature convergence.

[0045] Then Thompson sampling and scoring calculation And calculate the prediction score. .

[0046] in It follows a normal distribution. The parameter vector obtained from sampling. For action The predicted score, Transpose of the context vector.

[0047] Then greedily choose actions. .

[0048] in At any moment The chosen action The parameters that maximize the objective function. The algorithm selects actions (different replication data transmission probabilities) from a discrete set of replication data transmission probabilities. The algorithm output is only the replication data transmission probability ρ. For each new packet, replication is performed according to the replication data transmission probability ρ, that is, the data packet to be transmitted is sent on all paths with the replication data transmission probability ρ. With a replication data transmission probability of (1-ρ), no replication is performed. In this case, the default scheduling method of MPQUIC is followed, and the data packet to be transmitted is only transmitted on the path with the lowest round-trip delay.

[0049] In some embodiments, such as Figure 2 As shown, the model parameters include: statistics of the performed actions and posterior distribution parameters; The step of updating the model parameters of the multi-armed slot machine model based on the reward value includes the following steps: Step 202: Update the statistics of the executed action according to the reward value to obtain the updated statistics.

[0050] In this step, the statistics of the executed action include the cumulative action reward vector and the cumulative action context matrix; Based on the reward value, the statistics of the executed action are updated to obtain the updated statistics, including: The action cumulative context matrix is ​​incrementally updated using the following formula: ; in, For action Action cumulative context matrix, For the current context vector, This is the transpose of the context vector; Indicates assignment / incremental update; The cumulative reward vector for the action is incrementally updated using the following formula: ; in, For action The cumulative reward vector for actions, with dimension [missing information]. , As a reward value, This is the current context vector.

[0051] Step 204: Update the posterior distribution parameters according to the updated statistics to obtain the updated posterior distribution parameters.

[0052] In this step, the posterior distribution parameters include: the posterior mean and the posterior covariance matrix; The step of updating the posterior distribution parameters based on the updated statistics to obtain the updated posterior distribution parameters includes: The posterior mean is updated using the following formula: ; in, This is the updated posterior mean vector. Accumulate the inverse of the context matrix for the action. action The cumulative reward vector for actions; The posterior covariance matrix is ​​updated using the following formula: ; in, This is the updated posterior covariance matrix; For action Action cumulative context matrix, This represents finding the inverse of a matrix.

[0053] By updating the aforementioned data, the multi-armed slot machine model can maintain stable throughput and reduce packet loss rate, retransmission rate, and startup latency in environments with high packet loss and strong time-varying characteristics. User-space deployment is rapid; path starvation is avoided through quota limits and minimum proportions, improving system robustness; parameter sets, weights, and periods can be flexibly configured according to task levels, facilitating engineering optimization and auditing. The data transmission method in this application is applicable to scenarios such as disaster emergency communication, disaster relief, and emergency coverage in maritime and remote areas. Based on MPQUIC user-space implementation, it can be quickly deployed on existing equipment, demonstrating good engineering feasibility and application prospects.

[0054] It should be noted that the method in this embodiment can be executed by a single device, such as a computer or server. The method can also be applied in a distributed scenario, where multiple devices cooperate to complete the task. In such a distributed scenario, one of these devices may execute only one or more steps of the method in this embodiment, and the multiple devices will interact with each other to complete the method described.

[0055] It should be noted that the above description describes some embodiments of this application. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recorded in the claims can be performed in a different order than that shown in the above embodiments and still achieve the desired result. Furthermore, the processes depicted in the drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0056] Based on the same inventive concept, corresponding to any of the above embodiments, this application also provides a multipath data transmission device.

[0057] refer to Figure 3 The multipath data transmission device includes: The acquisition module 302 is configured to acquire the communication parameters of each path in multiple paths; The determining module 304 is configured to determine the probability of copying and transmitting the data packet to be transmitted based on the communication parameters and the preset multi-armed slot machine model; The transmission module 306 is configured to transmit all data in the data packet to be transmitted through the path with the least transmission delay in multiple paths, and to transmit a portion of the data in the data packet to be transmitted in each path other than the path with the least transmission delay, according to the replication transmission probability.

[0058] In some embodiments, the path with the shortest transmission latency among multiple paths transmits all data in the data packet to be transmitted, and in all paths other than the path with the shortest transmission latency, a portion of the data in the data packet to be transmitted is transmitted according to the replication transmission probability. Then, the method further includes: The calculation module 308 is configured to calculate the reward value of the replication transmission probability according to a set reward function; The update module 310 is configured to update the model parameters of the multi-armed slot machine model based on the reward value.

[0059] Optionally, the determining module 304 is further configured to determine the replication transmission probability of the data packet to be transmitted based on the communication parameters and a preset multi-armed slot machine model, including: The communication parameters are preprocessed, and the preprocessed communication parameters are input into the preset multi-armed slot machine model. The replication transmission probability is output through the multi-armed slot machine model.

[0060] In some embodiments, the determining module 304 is further configured to preprocess the communication parameters, including: The initial parameters are obtained by filling in missing values ​​for each parameter in the communication parameters. The initial parameters are normalized to obtain normalized parameters; All normalized parameters are concatenated to obtain the preprocessed communication parameters.

[0061] In some embodiments, the calculation module 308 is further configured as a method for constructing the reward function, including: The reward function is constructed with the objectives of minimizing transmission latency and maximizing network throughput, and is described by the following formula: ; in, For transmission throughput; This is the average delay; Packet loss rate; This represents the penalty coefficient for packet loss on overall transmission performance. This indicates the relative weights of throughput and latency in the overall utility. , .

[0062] In some embodiments, the update module 310 is further configured such that the model parameters include: statistics of the performed action and posterior distribution parameters; The step of updating the model parameters of the multi-armed slot machine model based on the reward value includes: Based on the reward value, the statistics of the executed action are updated to obtain the updated statistics; Based on the updated statistics, the posterior distribution parameters are updated to obtain the updated posterior distribution parameters.

[0063] In some embodiments, the update module 310 is further configured such that the statistics of the executed action include the cumulative action reward vector and the cumulative action context matrix; Based on the reward value, the statistics of the executed action are updated to obtain the updated statistics, including: The action cumulative context matrix is ​​incrementally updated using the following formula: ; in, For action Action cumulative context matrix, For the current context vector, This is the transpose of the context vector; Indicates assignment / incremental update; The cumulative reward vector for the action is incrementally updated using the following formula: ; in, For action The cumulative reward vector for actions, with dimension [missing information]. , As a reward value, This is the current context vector.

[0064] In some embodiments, the updating module 310 is further configured such that the posterior distribution parameters include: the posterior mean and the posterior covariance matrix; The step of updating the posterior distribution parameters based on the updated statistics to obtain the updated posterior distribution parameters includes: The posterior mean is updated using the following formula: ; in, This is the updated posterior mean vector. Accumulate the inverse of the context matrix for the action. action The cumulative reward vector for actions; The posterior covariance matrix is ​​updated using the following formula: ; in, This is the updated posterior covariance matrix; For action Action cumulative context matrix, This represents finding the inverse of a matrix.

[0065] For ease of description, the above devices are described in terms of function, divided into various modules. Of course, in implementing this application, the functions of each module can be implemented in one or more software and / or hardware.

[0066] The apparatus of the above embodiments is used to implement a corresponding multipath data method in any of the foregoing embodiments, and has the beneficial effects of the corresponding method embodiments, which will not be repeated here.

[0067] Based on the same inventive concept, corresponding to the methods of any of the above embodiments, this application also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement a multipath data method as described in any of the above embodiments.

[0068] Figure 5This embodiment illustrates a more specific hardware structure of an electronic device. The device may include a processor 1010, a memory 1020, an input / output interface 1030, a communication interface 1040, and a bus 1050. The processor 1010, memory 1020, input / output interface 1030, and communication interface 1040 are interconnected internally via the bus 1050.

[0069] The processor 1010 can be implemented using a general-purpose CPU (Central Processing Unit), microprocessor, application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of this specification.

[0070] The memory 1020 can be implemented in the form of ROM (Read Only Memory), RAM (Random Access Memory), static storage device, dynamic storage device, etc. The memory 1020 can store the operating system and other applications. When the technical solutions provided in the embodiments of this specification are implemented by software or firmware, the relevant program code is stored in the memory 1020 and is called and executed by the processor 1010.

[0071] The input / output interface 1030 is used to connect input / output modules to realize information input and output. Input / output modules can be configured as components within the device (not shown in the figure) or externally connected to the device to provide corresponding functions. Input devices may include keyboards, mice, touchscreens, microphones, various sensors, etc., while output devices may include displays, speakers, vibrators, indicator lights, etc.

[0072] The communication interface 1040 is used to connect a communication module (not shown in the figure) to enable communication between this device and other devices. The communication module can communicate via wired means (such as USB, Ethernet cable, etc.) or wireless means (such as mobile network, WIFI, Bluetooth, etc.).

[0073] Bus 1050 includes a pathway for transmitting information between various components of the device, such as processor 1010, memory 1020, input / output interface 1030, and communication interface 1040.

[0074] It should be noted that although the above-described device only shows the processor 1010, memory 1020, input / output interface 1030, communication interface 1040, and bus 1050, in specific implementations, the device may also include other components necessary for normal operation. Furthermore, those skilled in the art will understand that the above-described device may only include the components necessary for implementing the embodiments of this specification, and not necessarily all the components shown in the figures.

[0075] The electronic devices described above are used to implement a corresponding multipath data method in any of the foregoing embodiments, and have the beneficial effects of the corresponding method embodiments, which will not be repeated here.

[0076] Based on the same inventive concept, corresponding to the methods of any of the above embodiments, this application also provides a non-transitory computer-readable storage medium storing computer instructions for causing the computer to execute a multipath data method as described in any of the above embodiments.

[0077] The computer-readable medium of this embodiment includes permanent and non-permanent, removable and non-removable media, and information storage can be implemented by any method or technology. Information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transfer medium that can be used to store information accessible by a computing device.

[0078] The computer instructions stored in the storage medium of the above embodiments are used to cause the computer to execute a multipath data method as described in any of the above embodiments, and have the beneficial effects of the corresponding method embodiments, which will not be repeated here.

[0079] It is understood that before using the technical solutions of the various embodiments in this application, users will be informed of the type, scope of use, and usage scenarios of the personal information involved in an appropriate manner, and user authorization will be obtained.

[0080] For example, upon receiving a user's active request, a prompt message is sent to the user to explicitly inform them that the requested operation will require the acquisition and use of the user's personal information. This allows the user to independently choose, based on the prompt message, whether to provide personal information to the software or hardware such as electronic devices, applications, servers, or storage media performing the operations described in this application.

[0081] As an optional but not limited implementation, in response to a user's active request, sending a prompt message to the user can be done via a pop-up window, where the prompt message can be presented in text format. Furthermore, the pop-up window can also include a selection control allowing the user to choose "agree" or "disagree" to provide personal information to the electronic device.

[0082] It is understood that the above notification and user authorization process is merely illustrative and does not limit the implementation of this application. Other methods that comply with relevant laws and regulations may also be applied to the implementation of this application.

[0083] Those skilled in the art should understand that the discussion of any of the above embodiments is merely exemplary and is not intended to imply that the scope of this application (including the claims) is limited to these examples; within the framework of this application, the technical features of the above embodiments or different embodiments can also be combined, the steps can be implemented in any order, and there are many other variations of different aspects of the embodiments of this application as described above, which are not provided in the details for the sake of brevity.

[0084] Additionally, to simplify the description and discussion, and to avoid obscuring the embodiments of this application, the well-known power / ground connections to integrated circuit (IC) chips and other components may or may not be shown in the provided drawings. Furthermore, the apparatus may be shown in block diagram form to avoid obscuring the embodiments of this application, and this also takes into account the fact that the details of the implementation of these block diagram apparatuses are highly dependent on the platform on which the embodiments of this application will be implemented (i.e., these details should be fully understood by those skilled in the art). While specific details (e.g., circuits) have been set forth to describe exemplary embodiments of this application, it will be apparent to those skilled in the art that the embodiments of this application can be implemented without these specific details or with variations thereof. Therefore, these descriptions should be considered illustrative rather than restrictive.

[0085] Although this application has been described in conjunction with specific embodiments thereof, many substitutions, modifications, and variations of these embodiments will be apparent to those skilled in the art from the foregoing description. For example, other memory architectures (e.g., dynamic RAM (DRAM)) may be used with the embodiments discussed.

[0086] The embodiments of this application are intended to cover all such substitutions, modifications, and variations that fall within the broad scope of the appended claims. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the embodiments of this application should be included within the protection scope of this application.

Claims

1. A multipath data transmission method, characterized in that, include: Retrieve communication parameters for each path in multiple paths; Based on the communication parameters and the preset multi-armed slot machine model, the probability of copying and transmitting the data packet to be transmitted is determined; In multiple paths, all data in the data packet to be transmitted is transmitted through the path with the least transmission delay. In each path other than the path with the least transmission delay, a portion of the data in the data packet to be transmitted is transmitted according to the replication transmission probability.

2. The method according to claim 1, characterized in that, The path with the shortest transmission delay among multiple paths transmits all data in the data packet to be transmitted. In all other paths besides the path with the shortest transmission delay, a portion of the data packet to be transmitted is transmitted according to the replication transmission probability. Then, the process further includes: Calculate the reward value of the replication and transmission probability based on the set reward function; The model parameters of the multi-armed slot machine model are updated based on the reward value.

3. The method according to claim 1, characterized in that, The step of determining the replication transmission probability of the data packet to be transmitted based on the communication parameters and the preset multi-armed slot machine model includes: The communication parameters are preprocessed, and the preprocessed communication parameters are input into the preset multi-armed slot machine model. The replication transmission probability is output through the multi-armed slot machine model.

4. The method according to claim 3, characterized in that, The preprocessing of the communication parameters includes: The initial parameters are obtained by filling in missing values ​​for each parameter in the communication parameters. The initial parameters are normalized to obtain normalized parameters; All normalized parameters are concatenated to obtain the preprocessed communication parameters.

5. The method according to claim 2, characterized in that, The method for constructing the reward function includes: The reward function is constructed with the objectives of minimizing transmission latency and maximizing network throughput, and is described by the following formula: ； in, For transmission throughput; This is the average delay; Packet loss rate; This represents the penalty coefficient for packet loss on overall transmission performance. This indicates the relative weights of throughput and latency in the overall utility. , .

6. The method according to claim 2, characterized in that, The model parameters include: statistics of the executed actions and posterior distribution parameters; The step of updating the model parameters of the multi-armed slot machine model based on the reward value includes: Based on the reward value, the statistics of the executed action are updated to obtain the updated statistics; Based on the updated statistics, the posterior distribution parameters are updated to obtain the updated posterior distribution parameters.

7. The method according to claim 6, characterized in that, The statistics of the executed actions include the cumulative action reward vector and the cumulative action context matrix; Based on the reward value, the statistics of the executed action are updated to obtain the updated statistics, including: The action cumulative context matrix is ​​incrementally updated using the following formula: ； in, For action Action cumulative context matrix, For the current context vector, This is the transpose of the context vector; Indicates assignment / incremental update; The cumulative reward vector for the action is incrementally updated using the following formula: ； in, For action The cumulative reward vector for actions, with dimension [missing information]. , As a reward value, This is the current context vector.

8. The method according to claim 7, characterized in that, The posterior distribution parameters include: the posterior mean and the posterior covariance matrix; The step of updating the posterior distribution parameters based on the updated statistics to obtain the updated posterior distribution parameters includes: The posterior mean is updated using the following formula: ； in, This is the updated posterior mean vector. Accumulate the inverse of the context matrix for the action. action The cumulative reward vector for actions; The posterior covariance matrix is ​​updated using the following formula: ； in, This is the updated posterior covariance matrix; For action Action cumulative context matrix, This represents finding the inverse of a matrix.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the method as described in any one of claims 1 to 8.

10. A non-transitory computer-readable storage medium storing computer instructions, characterized in that, The computer instructions are used to cause the computer to perform the method described in any one of claims 1 to 8.

Images