Railway edge computing service caching and task unloading method and system

Abstract

The invention discloses a railway edge computing service caching and task unloading method and system, and the method comprises the steps: building an ACSCTO network model based on a pre-built railway system optimization model, in the ACSCTO network model, a service caching and task unloading network generates a service caching decision, generates a task unloading decision based on the service caching decision, and loads the task unloading decision to a server; the bandwidth resource allocation network generates a bandwidth resource allocation decision based on the service cache and task unloading decision, and the computing resource allocation network generates a computing resource allocation decision based on the service cache, task unloading and bandwidth resource allocation decision, and trains an ACSCOO network model; and performing railway system service caching, task unloading and resource allocation in real time by using the trained ACSCTO network model. According to the ACSCTO network model provided by the invention, service cache and task unloading can be adaptively and dynamically adjusted according to different task requirements, and computing resources and network bandwidth are allocated, so that the resource utilization rate is improved, the task processing efficiency is maximized, and the processing delay is reduced.

Description

Technical Field

[0001] The present invention belongs to the technical field of task offloading, and in particular relates to a railway edge computing service caching and task offloading method and system. Background Art

[0002] Mobile edge computing (MEC) significantly reduces the latency in data transmission by extending computing resources to locations close to railway vehicles (such as edge servers deployed along the way). MEC allows real-time interaction and task processing in railway systems, thereby minimizing network latency and improving task execution efficiency, especially in terms of time and energy consumption. Typically, computing tasks for railway vehicles are offloaded to edge servers for processing. Since tasks such as train monitoring and passenger information services not only require processing user-specific data, but also rely on large databases (such as passenger information databases or vehicle status databases), the transmission of these large-scale data may significantly increase communication latency, thereby affecting the overall performance of MEC. In addition, the resources of edge servers are limited compared to cloud centers. In railway scenarios, high-speed trains and concurrent task processing on multiple trains further intensify competition for bandwidth and computing resources. Tasks are highly heterogeneous in terms of data size, computing resource requirements, and latency constraints, which places higher requirements on the optimal utilization of resources.

[0003] Existing technologies often focus on task offloading and service caching, without comprehensively considering the multi-dimensional optimization of the entire system. Although task offloading can reduce the processing delay of a single task, it may cause the edge server resource load to be too large, affecting the performance of the overall system due to ignoring the coordinated allocation of bandwidth and computing resources. Furthermore, existing technologies usually focus on reducing task delays or improving resource utilization, and often ignore fairness between vehicles, especially when multiple railway vehicle tasks need to be processed. This may cause the computing tasks of some vehicles to be processed first, while the tasks of other vehicles are delayed, which cannot meet the requirements of the railway transportation system for real-time and fairness. The resources of the edge server are limited, and how to fairly allocate these resources while ensuring system performance is a difficult point. In summary, existing technologies are often only applicable to scenarios with light loads. Due to the characteristics of high mobility, task heterogeneity, real-time requirements and resource competition of railway vehicles, the performance of existing technologies will drop sharply when facing high-density task requests in the railway system. Summary of the invention

[0004] The purpose of the present invention is to address the above-mentioned problems existing in the prior art and to provide a railway edge computing service caching and task offloading method and system that can adaptively adjust service caching and task offloading, fairly allocate computing resources and network bandwidth, thereby improving the resource utilization of edge servers and local railway vehicles, reducing processing delays and improving task completion rates, and meeting the requirements of tasks of different types and complexities in railway scenarios.

[0005] To achieve the above objectives, the technical solution of the present invention is as follows:

[0006] In a first aspect, the present invention provides a railway edge computing service caching and task offloading method, the method comprising:

[0007] Step 1: Construct a railway system optimization model with the goal of maximizing average satisfaction;

[0008] Step 2: construct an ACSCTO network model, which includes a service cache and task offloading network, a bandwidth resource allocation network, and a computing resource allocation network, and is used to first generate a service cache decision, and then generate a task offloading decision based on the generated service cache decision. The bandwidth resource allocation network is constructed based on the bandwidth resource allocation model deconstructed from the railway system optimization model, and is used to generate bandwidth resource allocation decisions based on service cache decisions and task offloading decisions. The computing resource allocation network is constructed based on the computing resource allocation model deconstructed from the railway system optimization model, and is used to generate computing resource allocation decisions based on service cache decisions, task offloading decisions, and bandwidth resource allocation decisions.

[0009] Step 3: Train the ACSCTO network model;

[0010] Step 4: Use the trained ACSCTO network model to generate service cache decisions, task offloading decisions, and resource allocation decisions for the railway system. The resource allocation decisions include bandwidth resource allocation and computing resource allocation.

[0011] The training process of the ACSCTO network model includes:

[0012] S1, the service cache and task offloading network includes a service cache network and a task offloading network, the service cache network includes an Actor network and a Critic network, the Actor network and the Critic network have the same structure and share the same input data, set the time slot The input data is , Included in time slot The task information, service information and cache status of the previous time slot, the task information includes the number of services required for the task, the size of the task input data, the computing resources required for the task, and the channel gain between the task and the edge server. The service information includes the cache status and storage requirements of the service on the edge server; the output of the Actor network is transformed through the softmax layer to obtain a probability distribution vector , the probability distribution vector A single element in represents the edge server in the time slot Cache service The probability of the probability distribution vector Perform non-repeating sampling to obtain cache decisions ; Then enter the data , caching decision Combining to form input data , input data Input to the task offloading network, which outputs multiple offloading decisions ;

[0013] S2. Cache decision , uninstall decision The bandwidth resource allocation module outputs the bandwidth resource allocation decision and caches the decision. , uninstall decision , bandwidth resource allocation decision is input into the computing resource allocation network, and the computing resource allocation network outputs the computing resource allocation decision; based on the cache decision obtained , uninstall decision ,Bandwidth resource allocation decision,Computational resource allocation decision,Calculation average satisfaction;

[0014] Select the uninstall decision with the highest average satisfaction and name it uninstall decision , update the network parameters of the task offloading network according to the following formula :

[0015] ;

[0016] In the above formula, represents the learning rate of the task offloading network; Represents the network parameters relative to the task offloading The gradient of

[0017] S3. Calculate the offloading decision The satisfaction of a single task in the game is named reward , calculate the satisfaction of a single task when all tasks are executed locally and name it reward , the uninstall decision is calculated according to the following formula Uninstall rewards for a single task : ;

[0018] S4, combine to get a quaternion ,in For time slot Input data; convert the four tuple Input to the Critic network, the output size of the Critic network is Vector , represents the number of edge server cache services, the vector The elements in are , Represents the input data Select Service The predicted value of

[0019] S5. Utilization Network parameters of the Critic network And the network parameters of the Actor network Update; the update formula is:

[0020] ;

[0021] ;

[0022] ;

[0023] In the above formula, Represents the learning rate of the Critic network; Represents the parameters relative to the Critic network The gradient of Indicates that in a given Next, according to the Actor network parameters The probability distribution of the selected action; Represents input data The target value; Represents the discount factor.

[0024] The objective function of the railway system optimization model is:

[0025] ;

[0026] ;

[0027] ;

[0028] ;

[0029] ; ;

[0030] ; ;

[0031] In the above formula, Indicates average satisfaction; represents a computing resource vector; Represents the service cache vector; represents the bandwidth allocation vector, represents the task offloading vector; Representation Task satisfaction; Indicates the scheduling period; Indicates the number of user devices; Representation Task The maximum tolerable delay of Representation Task The uninstall decision Representation Task Execute locally, Representation Task Offloaded to edge servers for execution; Indicates the task assigned computing resources; Indicates the time allocated to a task in a single time slot bandwidth resources; represents the satisfaction correction parameter; Indicates the overall completion time of the unloading task; Representation Task Total processing time executed locally; Indicates service The cache decision when Indicates service Cache at the edge, when Indicates service Not cached; Representation Task The transmission delay of wireless communication from local devices to edge servers; Indicates service The time required to download from the cloud to the edge server; Representation Task Execution delays; , Respectively represent tasks In time slot , Time Slot The cache status of the required service; Indicates the transmission rate between the edge server and the cloud; Indicates the uplink data rate; Indicates time slot Task The wireless channel gain between the local device and the edge server; Indicates the transmit power of the local device; represents the noise power spectral density; Indicates the task The input data size; Indicates service Storage requirements; Indicates the computing resources required for the task; Indicates the task In time slot Request Service indicator variable of Indicates the average processing frequency of the user equipment;

[0032] The constraints of the railway system optimization model include:

[0033] ;

[0034] ; ;

[0035] ; ;

[0036] ; ;

[0037] ;

[0038] ; ;

[0039] ; ;

[0040] In the above formula, Indicates time slot Total energy consumption of edge servers; Indicates time slot Task The energy consumed by execution on edge servers; represents the maximum energy consumption allowed by the edge server; Indicates time slot Task The energy consumed by transmission on edge servers; Represents the total computing power of the edge server; Indicates the maximum cache capacity of the edge server; Indicates the number of services cached in the edge server; Indicates the total bandwidth of the edge server; Indicates service Average static power consumption; Represents the effective switch capacitance coefficient associated with the edge server.

[0041] The objective function of the bandwidth resource allocation model is: ;

[0042] ;

[0043] In the above formula, Indicates the task satisfaction after deconstruction;

[0044] The constraints of the bandwidth resource allocation model are: ;

[0045] ;

[0046] ;

[0047] ; ;

[0048] ;

[0049] ;

[0050] In the above formula, represents the Lambert W function; Represents a computing task The minimum bandwidth resources required to meet the maximum tolerable delay condition; Indicates that the calculation is not cached in the service and needs to be downloaded from the cloud. The maximum bandwidth resources required; , Respectively represent the tasks assigned to them in a single time slot The maximum and minimum bandwidth resources.

[0051] The objective function of the computing resource allocation model includes: ;

[0052] ;

[0053] The constraints of the computing resource allocation model include:

[0054] Minimum resource constraints:

[0055] ;

[0056] In the above formula, Representation Task In time slot Minimum computing resources required;

[0057] Energy threshold constraint: ;

[0058] Computing resource upper limit constraints: ;

[0059] Introducing the Lagrangian function into the computing resource allocation model, the computing resource allocation model is converted into a convex problem for easy solution:

[0060] ;

[0061] In the above formula, is the Lagrangian function; , , They represent the Lagrange multipliers related to the minimum resource constraint, energy threshold constraint, and computing resource upper limit constraint respectively.

[0062] In a second aspect, the present invention provides a railway edge computing service caching and task offloading system, the system comprising:

[0063] Railway system optimization model building module, used to build a railway system optimization model with the goal of maximizing average satisfaction;

[0064] An ACSCTO network model construction module is used to construct an ACSCTO network model, wherein the ACSCTO network model includes a service cache and task offloading network, a bandwidth resource allocation network, and a computing resource allocation network. The bandwidth resource allocation network is constructed based on a bandwidth resource allocation model deconstructed from a railway system optimization model, and is used to first generate a service cache decision, and then generate a task offloading decision based on the generated service cache decision, and is used to generate a bandwidth resource allocation decision based on the service cache decision and the task offloading decision; the computing resource allocation network is constructed based on a computing resource allocation model deconstructed from a railway system optimization model, and generates a computing resource allocation decision based on the service cache decision, the task offloading decision, and the bandwidth resource allocation decision;

[0065] ACSCTO network model training module, used to train the ACSCTO network model;

[0066] The actual operation module is used to use the trained ACSCTO network model to obtain service cache decisions, task offloading decisions, and resource allocation decisions of the railway system. The resource allocation decisions include bandwidth resource allocation and computing resource allocation.

[0067] The ACSCTO network model training module trains the ACSCTO model according to the following steps:

[0068] S1, the service cache and task offloading network includes a service cache network and a task offloading network, the service cache network includes an Actor network and a Critic network, the Actor network and the Critic network have the same structure and share the same input data, set the time slot The input data is , Included in time slot The task information, service information and cache status of the previous time slot, the task information includes the number of services required for the task, the size of the task input data, the computing resources required for the task, and the channel gain between the task and the edge server. The service information includes the cache status and storage requirements of the service on the edge server; the output of the Actor network is transformed through the softmax layer to obtain a probability distribution vector , the probability distribution vector A single element in represents the edge server in the time slot Cache service The probability of the probability distribution vector Perform non-repeating sampling to obtain cache decisions ; Then enter the data , caching decision Combining to form input data , input data Input to the task offloading network, which outputs multiple offloading decisions ;

[0069] S2. Cache decision , uninstall decision The bandwidth resource allocation module outputs the bandwidth resource allocation decision and caches the decision. , uninstall decision , bandwidth resource allocation decision is input into the computing resource allocation network, and the computing resource allocation network outputs the computing resource allocation decision; based on the cache decision obtained , uninstall decision ,Bandwidth resource allocation decision,Computational resource allocation decision,Calculation average satisfaction;

[0070] Select the uninstall decision with the highest average satisfaction and name it uninstall decision , update the network parameters of the task offloading network according to the following formula :

[0071] ;

[0072] In the above formula, represents the learning rate of the task offloading network; Represents the network parameters relative to the task offloading The gradient of

[0073] S3. Calculate the offloading decision The satisfaction of a single task in the game is named reward , calculate the satisfaction of a single task when all tasks are executed locally and name it reward , the uninstall decision is calculated according to the following formula Uninstall rewards for a single task : ;

[0074] S4, combine to get a quaternion ,in For time slot Input data; convert the four tuple Input to the Critic network, the output size of the Critic network is Vector , represents the number of edge server cache services, the vector The elements in are , Represents the input data Select Service The predicted value of

[0075] S5. Utilization Network parameters of the Critic network And the network parameters of the Actor network Update; the update formula is:

[0076] ;

[0077] ;

[0078] ;

[0079] In the above formula, Represents the learning rate of the Critic network; Represents the parameters relative to the Critic network The gradient of Indicates that in a given Next, according to the Actor network parameters The probability distribution of the selected action; Represents input data The target value; Represents the discount factor.

[0080] The objective function of the railway system optimization model is:

[0081] ;

[0082] ;

[0083] ;

[0084] ;

[0085] ; ;

[0086] ; ;

[0087] In the above formula, Indicates average satisfaction; represents a computing resource vector; Represents the service cache vector; represents the bandwidth allocation vector, represents the task offloading vector; Representation Task satisfaction; Indicates the scheduling period; Indicates the number of user devices; Representation Task The maximum tolerable delay of Representation Task The uninstall decision Representation Task Execute locally, Representation Task Offloaded to edge servers for execution; Indicates the task assigned computing resources; Indicates the time allocated to a task in a single time slot bandwidth resources; represents the satisfaction correction parameter; Indicates the overall completion time of the unloading task; Representation Task Total processing time executed locally; Indicates service The cache decision when Indicates service Cache at the edge, when Indicates service Not cached; Representation Task The transmission delay of wireless communication from local devices to edge servers; Indicates service The time required to download from the cloud to the edge server; Representation Task Execution delays; , Respectively represent tasks In time slot , Time Slot The cache status of the required service; Indicates the transmission rate between the edge server and the cloud; Indicates the uplink data rate; Indicates time slot Task The wireless channel gain between the local device and the edge server; Indicates the transmit power of the local device; represents the noise power spectral density; Representation Task The input data size; Indicates service Storage requirements; Indicates the computing resources required for the task; Representation Task In time slot Request Service indicator variable of Indicates the average processing frequency of the user equipment;

[0088] The constraints of the railway system optimization model include:

[0089] ;

[0090] ; ;

[0091] ; ;

[0092] ; ;

[0093] ;

[0094] ; ;

[0095] ; ;

[0096] In the above formula, Indicates time slot Total energy consumption of edge servers; Indicates time slot Task The energy consumed by execution on edge servers; represents the maximum energy consumption allowed by the edge server; Indicates time slot Task The energy consumed by transmission on edge servers; Represents the total computing power of the edge server; Indicates the maximum cache capacity of the edge server; Indicates the number of services cached in the edge server; Indicates the total bandwidth of the edge server; Indicates service Average static power consumption; Represents the effective switch capacitance coefficient associated with the edge server.

[0097] The objective function of the bandwidth resource allocation model is:

[0098] ;

[0099] ;

[0100] In the above formula, Indicates the task satisfaction after deconstruction;

[0101] The constraints of the bandwidth resource allocation model are:

[0102] ;

[0103] ;

[0104] ;

[0105] ;

[0106] ;

[0107] ;

[0108] ;

[0109] In the above formula, represents the Lambert W function; Represents a computing task The minimum bandwidth resources required to meet the maximum tolerable delay condition; Indicates that the calculation is not cached in the service and needs to be downloaded from the cloud. The maximum bandwidth resources required; , Respectively represent the tasks assigned to them in a single time slot The maximum and minimum bandwidth resources.

[0110] The objective function of the computing resource allocation model includes: ;

[0111] ;

[0112] The constraints of the computing resource allocation model include:

[0113] Minimum resource constraints: ;

[0114] In the above formula, Representation Task In time slot Minimum computing resources required;

[0115] Energy threshold constraint: ;

[0116] Computing resource upper limit constraints: ;

[0117] Introducing the Lagrangian function into the computing resource allocation model, the computing resource allocation model is converted into a convex problem for easy solution:

[0118] ;

[0119] In the above formula, is the Lagrangian function; , , They represent the Lagrange multipliers related to the minimum resource constraint, energy threshold constraint, and computing resource upper limit constraint respectively.

[0120] Compared with the prior art, the present invention has the following beneficial effects:

[0121] 1. The railway edge computing service caching and task offloading method described in the present invention constructs an ACSCTO model based on the Actor-Critic reinforcement learning algorithm. The ACSCTO model can adaptively and dynamically adjust service caching and task offloading according to the needs of different tasks, allocate computing resources and network bandwidth, improve the resource utilization of edge servers and local railway vehicles, maximize task processing efficiency, reduce processing delays and improve task completion rates, especially in response to different types and complex task requirements in railway scenarios, such as real-time passenger information updates, train status monitoring and intelligent scheduling, which can ensure the effective processing of various tasks, reduce delays and improve overall operating efficiency.

[0122] 2. In the railway edge computing service caching and task offloading method described in the present invention, the concave characteristics of the logarithmic function are used when designing the satisfaction function to ensure the fairness of resource allocation between trains, prevent some trains from occupying a large amount of computing or communication resources for a long time, and help maintain the overall stability of the railway system. BRIEF DESCRIPTION OF THE DRAWINGS

[0123] Figure 1 The present invention is a flowchart of the method.

[0124] Figure 2 The structure diagram of the system of the present invention is shown in FIG.

[0125] Figure 3 This is a network architecture diagram of the ACSCTO network model described in the present invention.

[0126] Figure 4 Comparison chart of passenger satisfaction, average response time, task completion rate, fairness among vehicles, average overtime, energy consumption, average download delay, and average number of unloading for different algorithms.

[0127] Figure 5 This is a comparison chart of the convergence performance of different algorithms. DETAILED DESCRIPTION

[0128] The present invention is further described in detail below in conjunction with specific implementations and drawings.

[0129] Embodiment 1:

[0130] See also Figure 1 , a railway edge computing service caching and task offloading method, follows the following steps in sequence:

[0131] Step 1: Construct a railway system optimization model with the goal of maximizing average satisfaction;

[0132] The railway system includes the cloud, an edge server (ES), and multiple train local devices (UDs), each of which generates at most one computing task in a single time slot. , computing tasks is indivisible and can be described as ,in Indicates the input data size of the task, represents the computing resources required for the task, Indicates the maximum tolerable delay of a task; the computing tasks generated by local devices can be uploaded to the cloud. The edge server has cache and computing functions. The edge server is located next to the base station and can offload the tasks of local devices from the cloud to the edge server through the wireless network. The edge server determines which services should be cached and which tasks should be offloaded;

[0133] The railway system optimization model is constructed with the goal of maximizing the average satisfaction of tasks. The maximum tolerable delay of tasks is taken into account when designing the average satisfaction, and the concave characteristics of the logarithmic function are used to achieve fair distribution of computing resources among local devices. The objective function of the railway system optimization model is finally obtained as follows:

[0134] ;

[0135] ;

[0136] ;

[0137] ;

[0138] ; ;

[0139] ; ;

[0140] In the above formula, Indicates average satisfaction; is a bandwidth allocation vector, which represents the spectrum resources allocated to each task in a single time slot; is the computing resource vector, which represents the computing resources allocated for the tasks offloaded to the edge server in a single time slot; Represents the service cache vector; represents the task offloading vector; Indicates the task satisfaction; Indicates the scheduling period; Indicates the number of user devices; Indicates the task The maximum tolerable delay of Indicates the task The uninstall decision Indicates the task Execute locally, Representation Task Offloaded to edge servers for execution; Indicates the task assigned computing resources; Indicates the time allocated to a task in a single time slot bandwidth resources; represents the satisfaction correction parameter; Indicates the overall completion time of the unloading task; Representation Task Total processing time performed on the local device; Indicates service The cache decision when Indicates service Cache at the edge, when Indicates service Not cached; Representation Task The transmission delay of wireless communication from local devices to edge servers; Indicates service The time required to download from the cloud to the edge server; Representation Task Execution delays; , Respectively represent tasks In time slot , Time Slot The cache status of the required service; Indicates the transmission rate between the edge server and the cloud; Indicates the uplink data rate; Indicates time slot Task The wireless channel gain between the local device and the edge server; Indicates the transmit power of the local device; represents the noise power spectral density; Representation Task The input data size; Indicates service Storage requirements; Indicates the computing resources required for the task; Representation Task In time slot Request Service indicator variable of Indicates the average processing frequency of the user equipment;

[0141] The constraints of the railway system optimization model include:

[0142] ;

[0143] ; ;

[0144] ; ;

[0145] ; ;

[0146] ;

[0147] ; ;

[0148] ; ;

[0149] In the above formula, Indicates time slot Total energy consumption of edge servers; Indicates time slot Task The energy consumed by execution on edge servers; represents the maximum energy consumption allowed by the edge server; Indicates time slot Task The energy consumed by transmission on edge servers; Represents the total computing power of the edge server; Indicates the maximum cache capacity of the edge server; Indicates the number of services cached in the edge server; Indicates the total bandwidth of the edge server; Indicates service Average static power consumption; represents the effective switch capacitance coefficient associated with the edge server;

[0150] In order to simplify the railway system optimization model, it is divided into two sub-problems: service cache and task offloading problem and resource allocation problem (MRA); the MRA problem can also be decomposed into two sub-problems: bandwidth resource allocation problem (BRA) and computing resource allocation problem (CRA). The BRA problem and the CRA problem can be solved by alternating iterations until the problem converges; for the cache decision, offloading decision and bandwidth resource allocation decision for a given single time slot obtained above, the bandwidth resource allocation model is constructed as follows:

[0151] The objective function of the bandwidth resource allocation model is:

[0152] ;

[0153] ;

[0154] In the above formula, Representation Task satisfaction after deconstruction;

[0155] The constraints of the bandwidth resource allocation model are: ;

[0156] ;

[0157] ;

[0158] ;

[0159] ;

[0160] ;

[0161] ;

[0162] In the above formula, represents the Lambert W function; Represents a computing task The minimum bandwidth resources required to meet the maximum tolerable delay condition; Indicates that the calculation is not cached in the service and needs to be downloaded from the cloud. The maximum bandwidth resources required; , Respectively represent the tasks assigned to them in a single time slot Maximum and minimum bandwidth resources;

[0163] The Lagrangian function is introduced into the bandwidth resource allocation model; the introduced Lagrangian function can be expressed as:

[0164] ;

[0165] In the above formula, The Lagrangian function introduced into the bandwidth resource allocation model; is a Lagrange multiplier vector, each multiplier corresponds to a constraint in the bandwidth resource allocation model;

[0166] Calculate according to the following formula And use binary update , until convergence;

[0167] ;

[0168] ; ; ;

[0169] For the cache decision, offload decision, and bandwidth allocation decision obtained above, the computing resource allocation model is constructed as follows:

[0170] The objective functions of the computing resource allocation model include:

[0171] ;

[0172] ;

[0173] The constraints of the computing resource allocation model include:

[0174] The minimum resource constraint is used to ensure that the computing resources allocated to each task are no less than the minimum required resources:

[0175] ;

[0176] In the above formula, Indicates the task In time slot Minimum computing resources required;

[0177] Energy threshold constraint is used to ensure that the energy consumption of the entire network does not exceed the predetermined threshold:

[0178] ;

[0179] The computing resource upper limit constraint is used to ensure that the total amount of computing resources allocated to all tasks does not exceed the total computing capacity of the edge server:

[0180] ;

[0181] Since the computing resource allocation model has linear and convex problems, the Lagrangian function is introduced into the computing resource allocation model to convert the computing resource allocation model into a convex problem for easy solution; the Lagrangian function can be expressed as:

[0182] ;

[0183] In the above formula, is the Lagrangian function; , , They represent the Lagrange multipliers related to the minimum resource constraint, energy threshold constraint, and computing resource upper limit constraint respectively;

[0184] After determining the Lagrange multipliers, we can calculate by solving the following function :

[0185] ;

[0186] Iteratively update the Lagrange multiplier until it satisfies ,in , Respectively Iteration, Iterations are assigned to tasks of computing resources, for Tolerance; the iterative update formula of the Lagrange multiplier is as follows:

[0187] ;

[0188] ;

[0189] ;

[0190] In the above formula, , , Both represent positive gradient step length; Indicates the current iteration number; , Respectively represent Iteration, Iterated Lagrange multiplier ; , Respectively represent Iteration, Iterated Lagrange multiplier ; , Respectively represent Iteration, Iterated Lagrange multiplier ;

[0191] By continuously updating the Lagrange multiplier to approach the optimal solution, in the process of approaching the optimal solution, the computing resource allocation decision can be dynamically adjusted to adapt to changes in user needs and network conditions, thereby achieving dynamic computing resource allocation;

[0192] Step 2: Construct ACSCTO network model, such as Figure 3As shown, the ACSCTO network model includes a service cache and task offloading network, a bandwidth resource allocation network, and a computing resource allocation network, which are used to first generate a service cache decision, and then generate a task offloading decision based on the generated service cache decision. The bandwidth resource allocation network is constructed based on the bandwidth resource allocation model deconstructed from the railway system optimization model, and is used to generate bandwidth resource allocation decisions based on the service cache decision and the task offloading decision. The computing resource allocation network is constructed based on the computing resource allocation model deconstructed from the railway system optimization model, and is used to generate computing resource allocation decisions based on the service cache decision, the task offloading decision, and the bandwidth resource allocation decision.

[0193] Step 3: Train the ACSCTO network model; Figure 3 As shown in Figure 1, the training process of the ACSCTO network model includes the following steps:

[0194] S1. The ACSCTO network model makes service cache and task offloading decisions based on the Actor-Critic framework; the service cache and task offloading network includes a service cache network and a task offloading network, the service cache network includes an Actor network and a Critic network, the Actor network has the same structure as the Critic network and shares the same input data, and the time slot The input data is , Included in time slot The task information, service information and cache status of the previous time slot, the task information includes the number of services required for the task, the size of the task input data, the computing resources required for the task, and the channel gain between the task and the edge server. The service information includes the cache status, energy consumption, and storage requirements of the service on the edge server; the output of the Actor network is transformed through the softmax layer to obtain a probability distribution vector , the probability distribution vector A single element in represents the edge server in the time slot Cache service The probability of each probability value is between 0 and 1, and the sum of the probability distribution is equal to 1; for the probability distribution vector Perform non-repeating sampling to obtain cache decisions ; Then enter the data , caching decision Combining to form input data , input data Input to the task offloading network, which outputs multiple offloading decisions ;

[0195] S2. Cache decision , uninstall decision The bandwidth resource allocation module outputs the bandwidth resource allocation decision and caches the decision. , uninstall decision , bandwidth resource allocation decision is input into the computing resource allocation network, and the computing resource allocation network outputs the computing resource allocation decision; based on the cache decision obtained , uninstall decision ,Bandwidth resource allocation decision,Computational resource allocation decision,Calculation average satisfaction;

[0196] For each time slot, the offloading decision with the highest average satisfaction is selected and named as offloading decision , will uninstall the decision And the corresponding input data Combine into two-tuples Stored in a limited buffer When the buffer When the memory reaches full capacity, the newly generated two-tuple Will replace the buffer The oldest tuple in; randomly from the buffer Select a batch of binary groups from the task and update the network parameters of the task offloading network according to the following formula :

[0197] ;

[0198] In the above formula, represents the learning rate of the task offloading network; Represents the network parameters relative to the task offloading The gradient of

[0199] S3. Calculate the offloading decision The satisfaction of a single task in the game is named reward , calculate the satisfaction of a single task when all tasks are executed locally and name it reward , the uninstall decision is calculated according to the following formula Uninstall rewards for a single task :

[0200] ;

[0201] S4, generate a 4-tuple for each time slot ,in For time slot Input data; store the newly generated quadruple at each time step in a memory with limited capacity When the memory When the memory reaches full capacity, the newly generated quadruplet replaces the memory The oldest quad in the random order from memory A batch of quadruple is extracted from the input to the Critic network, and the Critic network generates a state prediction value for each service in the current time slot. The output size of the Critic network is Vector , Represents the number of edge server cache services, vector The elements in are , Represents the input data Select Service The predicted value of In Should be understood as service ;

[0202] S5. Utilization Network parameters of the Critic network And the network parameters of the Actor network Update; the update formula is:

[0203] ;

[0204] ;

[0205] ;

[0206] In the above formula, Represents the learning rate of the Critic network; Represents the parameters relative to the Critic network The gradient of Indicates that in a given Next, according to the Actor network parameters The probability distribution of the selected action; Represents input data The target value; represents the discount factor;

[0207] Adopting the back-propagation algorithm and the offloading reward derived from the task offloading network Optimize the network parameters of the Actor network and the Critic network. Network parameters generally refer to weights and biases. By iteratively adjusting network parameters, the performance of the Actor network and the Critic network are gradually enhanced, thereby making more accurate decisions during the service caching process.

[0208] The specific training process of the ACSCTO network model is as follows:

[0209]

[0210] Step 4: Use the trained ACSCTO network model to generate service cache decisions, task offloading decisions, and resource allocation decisions for the railway system. The resource allocation decisions include bandwidth resource allocation and computing resource allocation.

[0211] In order to determine the optimal scheduling decision for each time slot, the ACSCTO model described in the present invention, on the one hand, jointly optimizes service caching and task offloading to achieve overall performance improvement. By iteratively updating the service caching network and the task offloading network, and using the rewards derived from the task offloading model to optimize the service caching network, the coordination and consistency of caching decisions and offloading decisions can be ensured, thereby maximizing satisfaction. On the other hand, based on the caching decisions and offloading decisions obtained from the service caching network and the task offloading network, the resource allocation network is used to determine the optimal resource allocation strategy, ultimately enabling the ACSCTO model to dynamically adapt to the environment and more effectively solve complex tasks and service requirements.

[0212] Performance Verification:

[0213] In order to verify the effectiveness of the method described in the present invention, a railway model simulation model is constructed, and the ACSCTO algorithm described in the present invention and other baseline algorithms are used to perform simulation calculations on the railway model simulation model. The simulation calculations are performed for 10,000 time slots, and the performance of each algorithm is compared based on the simulation results and the pre-constructed evaluation index system.

[0214] In the railway model simulation model, the edge server is located at the railway communication base station, covering an area of ​​100 km × 100 km. 30 trains are evenly distributed in this area. In order to meet the high real-time requirements of the trains, the edge server's storage capacity is set to 50GB, the computing power is 80GHz, and the total network capacity is 60GHz to ensure that it can handle task requests issued by multiple trains at the same time. The railway model simulation model provides ten train services, covering key tasks such as train status monitoring, real-time passenger information updates, and train scheduling. Each service occupies storage resources ranging from about 2GB to 8GB, and the energy consumption required for each service cache is about 3 to 7 units. The input data size of the computing task of each railway vehicle is about 5MB to 20MB, the computing resources required for the task are about 2GHz to 10GHz, and the task execution period is 1 second to 5 seconds.

[0215] Other baseline algorithms include: AC uniform allocation method (AC-ea), AC single value method (AC-sr), AC without decomposition method (AC-wd), popularity cache method (PC), and local calculation method (LP); AC uniform allocation method adopts the same cache and offloading strategy as the ACSCTO algorithm, but uniformly allocates bandwidth and computing resources (computing resources, namely CPU frequency) to the offloaded tasks; the difference between AC single value method and ACSCTO algorithm is that the state estimation value of the entire offloading decision is directly calculated using the A2C framework, instead of calculating the state estimation value based on a single service; the difference between AC without decomposition method and ACSCTO algorithm is that the cache decision and offloading decision are directly determined after obtaining the state estimation value of a single service; popularity cache method makes cache decisions based on the popularity of the service, and the popularity of the service is defined according to the frequency of service requests, and prioritizes tasks according to the urgency of the task (namely, the task delay limit), and tasks with stricter deadlines will be assigned higher priorities; local calculation method is that all tasks are executed locally.

[0216] The evaluation index system includes: convergence performance, passenger satisfaction, average response time, task completion rate, fairness between vehicles, average overtime, energy consumption, average download delay, and average number of unloads. The calculation results of each evaluation index are as follows:

[0217] (1) The convergence curve obtained by simulation is as follows Figure 5 As shown. Figure 5 It can be seen that these algorithms show rapid convergence in the initial stage when the time frame is less than 5000, and then gradually slow down and approach convergence at about 20000 time frames. Although these algorithms show similar convergence speed, there are significant differences in their final rewards. The ACSCTO algorithm proposed in the present invention shows the highest final reward, which verifies the effectiveness of the method proposed in the present invention. The final reward of the AC uniform allocation method exceeds that of the AC single value method, which highlights the advantages of computing state estimates based on a single service in optimizing task scheduling and resource allocation.

[0218] (2) Calculate passenger satisfaction, average response time, task completion rate, fairness between vehicles, average overtime, energy consumption, average download delay, and average number of unloads. Figure 4 As shown in (a)-(h). Figure 4As can be seen from Figure (a), the passenger satisfaction of all algorithms is negative. Among the five baseline algorithms, the AC uniform allocation method shows higher user satisfaction. Compared with the AC single value method, the AC uniform allocation method allocates less time for service updates, but has a higher task offloading rate to the edge server. This shows that the AC uniform allocation method has the ability to optimize service cache decisions, thereby reducing service download delays, improving task offloading decisions, and improving user satisfaction. However, the passenger satisfaction of the ACSCTO algorithm proposed in the present invention exceeds that of the AC uniform allocation method, which has the best performance among all baseline algorithms, because the ACSCTO algorithm not only inherits all the advantages of the AC uniform allocation method, but also integrates the Lagrangian function convex optimizer for resource allocation, which improves the upper limit of the overall algorithm. Therefore, ACSCTO shows rapid adaptation to vehicle tasks and provides greater flexibility in resource allocation. When faced with an urgent offloading task, the ACSCTO model will allocate more bandwidth and computing resources to it to prevent the delay from exceeding the time limit, thereby avoiding penalties due to delayed task completion. Therefore, as Figure 4 As shown in Figures (b)-(e), the ACSCTO algorithm has the shortest average response time, the lowest average overtime, the highest task completion rate and fairness among vehicles. At the same time, the energy consumption of the ACSCTO algorithm is still comparable to that of the AC uniform allocation method, which means that the ACSCTO algorithm can achieve preferential utilization of limited resources. The non-decomposition AC method does not separate service caching from task offloading, but uses the A2C framework to determine the optimal caching and offloading decisions. Therefore, it cannot understand the correspondence between services and tasks, lacks consistency between service caching and task offloading, and results in a significant reduction in the final reward. The popularity cache method is a heuristic algorithm that aims to cache frequently requested services on edge servers and preferentially offload more urgent tasks to edge servers. However, it lacks versatility and rapid adaptability. It not only shows the second worst performance after the local computing method in terms of passenger satisfaction, average response time and task completion rate, but also has the worst fairness among users and shows significant fluctuations in the average number of offloads.

[0219] Embodiment 2:

[0220] See also Figure 2, a railway edge computing service caching and task offloading system, the system includes a railway system optimization model construction module, an ACSCTO network model construction module, an ACSCTO network model training module, and an actual operation module, the railway system optimization model construction module is used to construct a railway system optimization model with the goal of maximizing average satisfaction; the objective function and constraint conditions of the railway system optimization model are as shown in step 1 of Example 1; the ACSCTO network model construction module is used to construct an ACSCTO network model, the ACSCTO network model includes a service cache and task offloading network, a bandwidth resource allocation network, and a computing resource allocation network, the bandwidth resource allocation network is constructed based on the bandwidth resource allocation model deconstructed from the railway system optimization model, and is used to first generate a service cache decision, and then generate a service cache decision based on the generated service cache decision. The service caching decision generates a task offloading decision; the computing resource allocation network is constructed based on the computing resource allocation model deconstructed from the railway system optimization model, and the computing resource allocation decision is generated based on the service caching decision, the task offloading decision, and the bandwidth resource allocation decision; the objective function and constraint conditions of the bandwidth resource allocation model, the objective function and constraint conditions of the computing resource allocation model are all as shown in step one of Example 1; the ACSCTO network model training module is used to train the ACSCTO network model; the specific training steps of the ACSCTO network model are shown in step three of Example 1; the actual operation module is used to use the trained ACSCTO network model to obtain the service caching decision, task offloading decision, and resource allocation decision of the railway system, and the resource allocation decision includes bandwidth resource allocation and computing resource allocation.

Claims

1. A railway edge computing service caching and task offloading method, characterized in that: The method comprises: Step 1: Construct a railway system optimization model with the goal of maximizing average satisfaction; Step 2: construct an ACSCTO network model, which includes a service cache and task offloading network, a bandwidth resource allocation network, and a computing resource allocation network, and is used to first generate a service cache decision, and then generate a task offloading decision based on the generated service cache decision. The bandwidth resource allocation network is constructed based on the bandwidth resource allocation model deconstructed from the railway system optimization model, and is used to generate bandwidth resource allocation decisions based on service cache decisions and task offloading decisions. The computing resource allocation network is constructed based on the computing resource allocation model deconstructed from the railway system optimization model, and is used to generate computing resource allocation decisions based on service cache decisions, task offloading decisions, and bandwidth resource allocation decisions. Step 3: Train the ACSCTO network model; Step 4: Use the trained ACSCTO network model to generate service cache decisions, task offloading decisions, and resource allocation decisions for the railway system. The resource allocation decisions include bandwidth resource allocation and computing resource allocation.

2. A railway edge computing service caching and task offloading method according to claim 1, characterized in that: The training process of the ACSCTO network model includes: S1, the service cache and task offloading network includes a service cache network and a task offloading network, the service cache network includes an Actor network and a Critic network, the Actor network and the Critic network have the same structure and share the same input data, set the time slot The input data is , Included in time slot The task information, service information and cache status of the previous time slot, the task information includes the number of services required for the task, the size of the task input data, the computing resources required for the task, and the channel gain between the task and the edge server. The service information includes the cache status and storage requirements of the service on the edge server; the output of the Actor network is transformed through the softmax layer to obtain a probability distribution vector , the probability distribution vector A single element in represents the edge server in the time slot Cache service The probability of the probability distribution vector Perform non-repeating sampling to obtain cache decisions ; Then enter the data , caching decision Combining to form input data , input data Input to the task offloading network, which outputs multiple offloading decisions ; S2. Cache decision , uninstall decision The bandwidth resource allocation module outputs the bandwidth resource allocation decision and caches the decision. , uninstall decision , bandwidth resource allocation decision is input into the computing resource allocation network, and the computing resource allocation network outputs the computing resource allocation decision; based on the cache decision obtained , uninstall decision ,Bandwidth resource allocation decision,Computational resource allocation decision,Calculation average satisfaction; Select the uninstall decision with the highest average satisfaction and name it uninstall decision , update the network parameters of the task offloading network according to the following formula : ； In the above formula, represents the learning rate of the task offloading network; Represents the network parameters relative to the task offloading The gradient of S3. Calculate the offloading decision The satisfaction of a single task in the game is named reward , calculate the satisfaction of a single task when all tasks are executed locally and name it reward , the uninstall decision is calculated according to the following formula Uninstall rewards for a single task : ； S4, combine to get a quaternion ,in For time slot Input data; the four-tuple Input to the Critic network, the output size of the Critic network is Vector , represents the number of edge server cache services, the vector The elements in are , Represents the input data Select Service The predicted value of S5. Utilization Network parameters of the Critic network And the network parameters of the Actor network Update; the update formula is: ； ； ； In the above formula, Represents the learning rate of the Critic network; Represents the parameters relative to the Critic network The gradient of Indicates that in a given Next, according to the Actor network parameters The probability distribution of the selected action; Represents input data The target value; Represents the discount factor.

3. A railway edge computing service caching and task offloading method according to claim 2, characterized in that: The objective function of the railway system optimization model is: ； ； ； ； ； ； ； ； In the above formula, Indicates average satisfaction; represents a computing resource vector; Represents the service cache vector; represents the bandwidth allocation vector, represents the task offloading vector; Indicates the task satisfaction; Indicates the scheduling period; Indicates the number of user devices; Indicates the task The maximum tolerable delay; Indicates the task The uninstall decision Indicates the task Execute locally, Indicates the task Offloaded to edge servers for execution; Indicates the assignment to a task computing resources; Indicates the time allocated to a task in a single time slot bandwidth resources; represents the satisfaction correction parameter; Indicates the overall completion time of the unloading task; Indicates the task Total processing time executed locally; Indicates service The cache decision when Indicates service Cache at the edge, when Indicates service Not cached; Indicates the task The transmission delay of wireless communication from local devices to edge servers; Indicates service The time required to download from the cloud to the edge server; Indicates the task Execution delays; , Respectively represent tasks In time slot , Time Slot The cache status of the required service; Indicates the transmission rate between the edge server and the cloud; Indicates the uplink data rate; Indicates time slot Task The wireless channel gain between the local device and the edge server; Indicates the transmit power of the local device; represents the noise power spectral density; Indicates the task The input data size; Indicates service Storage requirements; Indicates the computing resources required for the task; Indicates the task In time slot Request Service indicator variable of Indicates the average processing frequency of the user equipment; The constraints of the railway system optimization model include: ； ； ； ； ； ； ； ； ； ； ； ； In the above formula, Indicates time slot Total energy consumption of edge servers; Indicates time slot Task The energy consumed by execution on edge servers; represents the maximum energy consumption allowed by the edge server; Indicates time slot Task The energy consumed by transmission on edge servers; Represents the total computing power of the edge server; Indicates the maximum cache capacity of the edge server; Indicates the number of services cached in the edge server; Indicates the total bandwidth of the edge server; Indicates service Average static power consumption; Represents the effective switch capacitance coefficient associated with the edge server.

4. A railway edge computing service caching and task offloading method according to claim 3, characterized in that: The objective function of the bandwidth resource allocation model is: ； ； In the above formula, Indicates the task satisfaction after deconstruction; The constraints of the bandwidth resource allocation model are: ； ； ； ； ； ； ； In the above formula, represents the Lambert W function; Represents a computing task The minimum bandwidth resources required to meet the maximum tolerable delay condition; Indicates that the calculation is not cached in the service and needs to be downloaded from the cloud. The maximum bandwidth resources required; , Respectively represent the tasks assigned to each time slot The maximum and minimum bandwidth resources.

5. A railway edge computing service caching and task offloading method according to claim 3 or 4, characterized in that: The objective function of the computing resource allocation model includes: ； ； The constraints of the computing resource allocation model include: Minimum resource constraints: ； In the above formula, Indicates the task In time slot Minimum computing resources required; Energy threshold constraint: ； Computing resource upper limit constraints: ； Introducing the Lagrangian function into the computing resource allocation model, the computing resource allocation model is converted into a convex problem for easy solution: ； In the above formula, is the Lagrangian function; , , They represent the Lagrange multipliers related to the minimum resource constraint, energy threshold constraint, and computing resource upper limit constraint respectively.

6. A railway edge computing service caching and task offloading system, characterized by: The system comprises: Railway system optimization model building module, used to build a railway system optimization model with the goal of maximizing average satisfaction; An ACSCTO network model construction module is used to construct an ACSCTO network model, wherein the ACSCTO network model includes a service cache and task offloading network, a bandwidth resource allocation network, and a computing resource allocation network. The bandwidth resource allocation network is constructed based on a bandwidth resource allocation model deconstructed from a railway system optimization model, and is used to first generate a service cache decision, and then generate a task offloading decision based on the generated service cache decision, and is used to generate a bandwidth resource allocation decision based on the service cache decision and the task offloading decision; the computing resource allocation network is constructed based on a computing resource allocation model deconstructed from a railway system optimization model, and generates a computing resource allocation decision based on the service cache decision, the task offloading decision, and the bandwidth resource allocation decision; ACSCTO network model training module, used to train the ACSCTO network model; The actual operation module is used to use the trained ACSCTO network model to obtain service cache decisions, task offloading decisions, and resource allocation decisions of the railway system. The resource allocation decisions include bandwidth resource allocation and computing resource allocation.

7. A railway edge computing service caching and task offloading system according to claim 6, characterized in that: The ACSCTO network model training module trains the ACSCTO network model according to the following steps: S1, the service cache and task offloading network includes a service cache network and a task offloading network, the service cache network includes an Actor network and a Critic network, the Actor network and the Critic network have the same structure and share the same input data, set the time slot The input data is , Included in time slot The task information, service information and cache status of the previous time slot, the task information includes the number of services required for the task, the size of the task input data, the computing resources required for the task, and the channel gain between the task and the edge server. The service information includes the cache status and storage requirements of the service on the edge server; the output of the Actor network is transformed through the softmax layer to obtain a probability distribution vector , the probability distribution vector A single element in represents the edge server in the time slot Cache service The probability of the probability distribution vector Perform non-repeating sampling to obtain cache decisions ; Then enter the data , caching decision Combining to form input data , input data Input to the task offloading network, which outputs multiple offloading decisions ; S2. Cache decision , uninstall decision The bandwidth resource allocation module outputs the bandwidth resource allocation decision and caches the decision. , uninstall decision , bandwidth resource allocation decision is input into the computing resource allocation network, and the computing resource allocation network outputs the computing resource allocation decision; based on the cache decision obtained , uninstall decision ,Bandwidth resource allocation decision,Computational resource allocation decision,Calculation average satisfaction; Select the uninstall decision with the highest average satisfaction and name it uninstall decision , update the network parameters of the task offloading network according to the following formula : ； In the above formula, represents the learning rate of the task offloading network; Represents the network parameters relative to the task offloading The gradient of S3. Calculate the offloading decision The satisfaction of a single task in the game is named reward , calculate the satisfaction of a single task when all tasks are executed locally and name it reward , the uninstall decision is calculated according to the following formula Uninstall rewards for a single task : ； S4, combine to get a quaternion ,in For time slot Input data; convert the four tuple Input to the Critic network, the output size of the Critic network is Vector , represents the number of edge server cache services, the vector The elements in are , Represents the input data Select Service The predicted value of S5. Utilization Network parameters of the Critic network And the network parameters of the Actor network Update; the update formula is: ； ； ； In the above formula, Represents the learning rate of the Critic network; Represents the parameters relative to the Critic network The gradient of Indicates that in a given Next, according to the Actor network parameters The probability distribution of the selected action; Represents input data The target value; Represents the discount factor.

8. A railway edge computing service caching and task offloading system according to claim 7, characterized in that: The objective function of the railway system optimization model is: ； ； ； ； ； ； ； ； In the above formula, Indicates average satisfaction; represents a computing resource vector; Represents the service cache vector; represents the bandwidth allocation vector, represents the task offloading vector; Indicates the task satisfaction; Indicates the scheduling period; Indicates the number of user devices; Indicates the task The maximum tolerable delay of Indicates the task The uninstall decision Indicates the task Execute locally, Indicates the task Offloaded to edge servers for execution; Indicates the task assigned computing resources; Indicates the time allocated to a task in a single time slot bandwidth resources; represents the satisfaction correction parameter; Indicates the overall completion time of the unloading task; Indicates the task Total processing time executed locally; Indicates service The cache decision when Indicates service Cache at the edge, when Indicates service Not cached; Indicates the task The transmission delay of wireless communication from local devices to edge servers; Indicates service The time required to download from the cloud to the edge server; Indicates the task Execution delays; , Respectively represent tasks In time slot , Time Slot The cache status of the required service; Indicates the transmission rate between the edge server and the cloud; Indicates the uplink data rate; Indicates time slot Task The wireless channel gain between the local device and the edge server; Indicates the transmit power of the local device; represents the noise power spectral density; Indicates the task The input data size; Indicates service Storage requirements; Indicates the computing resources required for the task; Indicates the task In time slot Request Service indicator variable of Indicates the average processing frequency of the user equipment; The constraints of the railway system optimization model include: ； ； ； ； ； ； ； ； ； ； ； ； In the above formula, Indicates time slot Total energy consumption of edge servers; Indicates time slot Task The energy consumed by execution on edge servers; represents the maximum energy consumption allowed by the edge server; Indicates time slot Task The energy consumed by transmission on edge servers; Represents the total computing power of the edge server; Indicates the maximum cache capacity of the edge server; Indicates the number of services cached in the edge server; Indicates the total bandwidth of the edge server; Indicates service Average static power consumption; Represents the effective switch capacitance coefficient associated with the edge server.

9. A railway edge computing service caching and task offloading system according to claim 8, characterized in that: The objective function of the bandwidth resource allocation model is: ； ； In the above formula, Indicates the task satisfaction after deconstruction; The constraints of the bandwidth resource allocation model are: ； ； ； ； ； ； ； In the above formula, represents the Lambert W function; Represents a computing task The minimum bandwidth resources required to meet the maximum tolerable delay condition; Indicates that the calculation is not cached in the service and needs to be downloaded from the cloud. The maximum bandwidth resources required; , Respectively represent the tasks assigned to them in a single time slot The maximum and minimum bandwidth resources.

10. A railway edge computing service caching and task offloading system according to claim 8 or 9, characterized in that: The objective function of the computing resource allocation model includes: ； ； The constraints of the computing resource allocation model include: Minimum resource constraints: ； In the above formula, Indicates the task In time slot Minimum computing resources required; Energy threshold constraint: ； Computing resource upper limit constraints: ； Introducing the Lagrangian function into the computing resource allocation model, the computing resource allocation model is converted into a convex problem for easy solution: ； In the above formula, is the Lagrangian function; , , They represent the Lagrange multipliers related to the minimum resource constraint, energy threshold constraint, and computing resource upper limit constraint respectively.

Images