A pre-aided caching method based on vehicle location

Abstract

This invention relates to a pre-caching method based on vehicle location. The method includes: constructing a highly dynamic scenario with different cached content requirements based on heterogeneous network units, and obtaining parameters such as vehicle location; modeling base stations, roadside units, and vehicles as leaders and followers in a Stackelberg game; designing a utility function based on data caching rate and data caching cost calculated from vehicle location and other parameters; obtaining the optimal pre-caching alliance members by solving the established game model; calculating caching efficiency and node utility values ​​under this game model by jointly considering the timeliness, popularity, and data volume of cached content; constructing a multi-dimensional knapsack objective function constrained by the alliance members; and then using dynamic programming to backtrack and solve for the optimal pre-cached content combination and its deployment scheme. This invention can effectively reduce redundant transmission and energy consumption under complex business loads while compressing transmission latency to the theoretically optimal range, providing crucial underlying support for real-time vehicle services.

Description

Technical Field

[0001] This invention belongs to the field of heterogeneous vehicle networking caching technology, specifically relating to a pre-assisted caching method based on vehicle location. Background Technology

[0002] With the exponential growth of the Internet of Vehicles (IoV) network, caching service demands are becoming increasingly diversified and differentiated, exacerbating network heterogeneity and significantly reducing the stability of caching transmission links. In highly dynamic scenarios involving changes in road topology, vehicle movement states, and differentiated caching service demands, on the one hand, the caching content requested by vehicles (such as audio, environmental perception data, and video applications) changes rapidly in type, data volume, and timeliness requirements. Frequent core network data requests lead to a surge in cross-base station scheduling, resulting in long-term occupation of backhaul link resources. This not only increases the number of link switching attempts but also significantly increases end-to-end transmission latency, making it difficult to meet the response requirements of low-latency services. On the other hand, traditional caching methods generally employ static strategies based on historical request frequencies, making caching decisions based on the past popularity of content. This approach cannot adapt to rapid changes in business needs. Because the popularity of cached content changes dynamically with time and location, it easily causes a serious discrepancy between pre-cached content and actual requests, leading to a large amount of unnecessary redundant data transmission and storage. This not only causes a sharp increase in energy consumption but also severely impacts the transmission efficiency of business data. Therefore, ensuring low-power transmission while meeting stringent low-latency requirements has become a key issue to be addressed in current vehicle-to-everything (V2X) caching technology.

[0003] The document with application number "202310158049.9" discloses "A Low-Power Secure Caching Method and Medium for Vehicle Networking." This method significantly improves system power consumption by jointly considering caching schemes based on vehicle preferences, vehicle activity levels, content size, and power and bandwidth allocation. It also utilizes an improved elliptic curve cryptography algorithm and advanced encryption standard algorithms to securely place content, effectively protecting data security. However, during content placement and distribution, dynamic changes in channel states caused by vehicle movement and cross-base station interference lead to unstable data transmission rates after a vehicle sends a request, resulting in latency issues that affect the transmission efficiency of service data and the fulfillment of low-latency service requirements. The application with application number "202410630336.X" discloses "a method for selecting cache nodes in vehicle networking under complex scenarios". This method comprehensively considers multiple factors such as obstacle occlusion, vehicle-to-vehicle distance, and link duration, dynamically allocates file requests, achieves load balancing, and covers the entire network with the fewest cache nodes. It effectively improves the communication interruption and uneven node load caused by link occlusion and multiple requests per vehicle in traditional solutions. However, in highly dynamic scenarios, the node selection strategy based on instantaneous comprehensive weights may cause vehicles to frequently switch connection nodes due to environmental disturbances, resulting in transmission rate fluctuations and affecting user experience. At the same time, instantaneous decision-making is difficult to guarantee long-term load balancing and may instead bring new load unevenness and interruption risks. In addition, unstable links lead to a large number of data retransmissions, which, together with node overload, result in high data transmission energy consumption.

[0004] The common problem with the above documents is that in the dynamically changing vehicle-to-everything (V2X) environment, collaborative caching struggles to balance latency and energy consumption, thus failing to efficiently guarantee user experience. Summary of the Invention

[0005] This invention provides a pre-assisted caching method based on vehicle location to solve the problem in the prior art that collaborative caching is difficult to balance latency and energy consumption in a dynamically changing vehicle network environment.

[0006] To achieve the above objectives, the technical solution provided by this invention is: a pre-assisted caching method based on vehicle location, comprising the following steps:

[0007] Step 1: Construct a high-dynamic scenario with different cache content requirements based on heterogeneous network units, and obtain relevant network parameters, including the locations of vehicles, base stations, and roadside units;

[0008] Step 2: Model the base station, roadside unit, and vehicle as leaders and followers in the Stackelberg game model, respectively. Under this game framework, calculate the data caching rate and data caching cost of each network layer based on the relevant network parameters obtained in Step 1. Use these as policy variables to design corresponding utility functions for each participant. Finally, by solving the established Stackelberg game model, obtain the composition of the optimal pre-caching alliance members and maximize the total utility of the system.

[0009] Step 3: Based on the Stackelberg game model established in Step 2, calculate the caching efficiency by combining the timeliness, popularity and data volume of the cached content, and calculate the caching node utility value of each cache node in the optimal pre-caching alliance member obtained in Step 2. Using the total capacity of the alliance member as a constraint, design the objective function of the multidimensional knapsack problem, and then use the dynamic programming method to backtrack and solve it, thereby determining the optimal combination of pre-cached content and its deployment scheme in the optimal pre-caching alliance member.

[0010] Furthermore, in step two above, the data caching cost of each network layer includes payment costs and energy consumption costs:

[0011]

[0012]

[0013]

[0014]

[0015] in, and These represent the data caching costs of the base station and the auxiliary caching node layers, respectively. Indicates a secondary cache node. The cost of paying for the auxiliary cache node layer network of heterogeneous vehicle-to-everything (V2X) networks. The energy consumption cost of each layer of the network, Indicates base station For auxiliary cache nodes The unit price set for allocating cached tasks. Indicates base station Assigned to secondary cache nodes The amount of cached tasks, Indicates secondary cache node The unit energy consumption coefficient, , The corresponding weights.

[0016] Furthermore, in step two above, the utility function is designed based on network performance and data caching costs as follows, whereby network performance includes caching task volume, latency, and effective service duration:

[0017]

[0018]

[0019]

[0020]

[0021]

[0022] in, Let be the total utility function of the participants. Let be the utility function of the base station. The utility function for a selected roadside unit or a selected vehicle. , Select the roadside unit and the number of vehicles; The value of currently allocated cached tasks, For the latency of the base station layer; For the latency of each network layer, To select a roadside unit or a vehicle's effective service duration within the current heterogeneous network unit, For data caching rate, The weights are the corresponding to the time delay.

[0023] The optimization objective is:

[0024]

[0025]

[0026]

[0027]

[0028]

[0029]

[0030]

[0031]

[0032] in, Allocate the total cache task amount for the base station. This represents the total number of cached tasks. To allocate the amount of cached tasks to the roadside units, The amount of cached tasks allocated to vehicles. This represents the maximum total cache capacity acceptable to the roadside unit. The maximum acceptable total cache size for vehicles. To the maximum tolerable delay, This represents the total number of roadside units under the current heterogeneous network unit. This represents the total number of vehicles in the current heterogeneous network unit. This indicates the distance from the base station to the auxiliary buffer node. Let P1 be the communication range of the base station. In general, P1 is the optimization objective, which represents maximizing the total utility function of the pre-caching consortium after the Stackelberg game. C1, C2, and C3 respectively constrain the allocated cache task amount to not exceed the maximum load, C4 constrains the maximum latency, C5 and C6 constrain the number of pre-cached RSUs and pre-cached vehicles, and C7 constrains the communication effectiveness of pre-caching consortium members.

[0033] Furthermore, in step three above, the caching efficiency is calculated based on the timeliness, popularity, and data volume of the jointly cached content. as follows:

[0034]

[0035]

[0036]

[0037] in, Indicates the first The caching benefits of pre-cached content This indicates the popularity of cached content, and is determined by calculating the number of requests generated by a vehicle within a specified time interval to understand the popularity of each piece of content. Indicates the cached content at time The timeliness score is dynamically calculated using an exponential decay model. Indicates the penalty factor for the size of cached content. , This represents the corresponding weighting coefficient; This indicates the initial timeliness score. The attenuation coefficient controls the rate at which time-sensitivity decreases. Indicates the time step. This represents the lower limit of timeliness, ensuring that timeliness does not decay indefinitely. Indicates the first The size of the cached content data.

[0038] Furthermore, in step three above, the objective function... Represented as:

[0039]

[0040]

[0041] in, This indicates the total number of candidate pre-cached contents. This indicates the total number of nodes in the pre-cached consortium. Indicates the first The quality of service gain of pre-cached content is determined by its data volume. and expected average transmission rate Decide, Indicating the first in the pre-caching alliance The cache node utility value of each cache node. For binary decision variables, , , This represents the corresponding weighting coefficient.

[0042] The optimization objective is:

[0043]

[0044]

[0045]

[0046] Where C1 represents the constraint on the node's storage space, Indicates the first The storage space required for each cached content; C2 represents the uniqueness constraint on the content, ensuring that for any given content... and any two distinct nodes and This content can only be stored on one node at most.

[0047] Compared with the prior art, the beneficial effects of the present invention are:

[0048] (1) Based on the Stackelberg game framework, this invention designs a utility function that comprehensively considers data caching rate and data caching cost. By modeling roadside units and vehicles as flexibly accessible auxiliary caching nodes, a hierarchical collaborative pre-caching alliance mechanism is constructed. This mechanism can dynamically select heterogeneous nodes with high collaboration potential according to the real-time network status, optimize the allocation and utilization of caching node capacity resources, and thus improve adaptability and collaboration efficiency in dynamic environments. In addition, load balancing is achieved through distributed decision-making, effectively reducing the transmission pressure on the core network and providing a feasible path to solve the problem of unstable maintenance of node collaboration relationships in highly dynamic scenarios.

[0049] (2) This invention designs a cache benefit calculation method that integrates multiple indicators such as timeliness, popularity and data volume. It models the capacity constraint and cache content selection and deployment problem as a multidimensional knapsack problem, and solves the problem by backtracking based on dynamic programming method. This enables accurate screening and pre-deployment of cache content, completes cache decision before the vehicle enters the effective communication range, avoids peak pressure of real-time transmission, reduces energy consumption and ensures service continuity.

[0050] (3) This invention deeply integrates Stackelberg game modeling to construct a distributed pre-caching alliance with a multi-dimensional constraint knapsack algorithm, and constructs a closed-loop decision framework of "dynamic node selection - content optimization caching". This solves the core decision problems of "who to cooperate with", "what to cache", and "how to cache optimally". The high degree of matching between the two in terms of timing, constraints and objectives, and the total capacity of the node set output by the game as the input constraint of the multi-dimensional knapsack problem, enable the system to respond to real-time caching business changes, thereby effectively solving the end-to-end dynamic optimization problem and realizing the elastic deployment of collaborative caching content and the overall performance leap.

[0051] (4) The caching method of the present invention avoids the blindness of random caching and overcomes the response bottleneck and complexity of a single base station and traditional AI model in a dynamic vehicle network environment. Thus, under complex business loads, it can effectively reduce redundant transmission and energy consumption while compressing transmission latency to the theoretical optimal range, providing key underlying support for real-time vehicle services. Attached Figure Description

[0052] Figure 1 This is a flowchart of the present invention;

[0053] Figure 2 This is a schematic diagram illustrating an applicable scenario for an example of the present invention;

[0054] Figure 3 A comparison chart of latency performance for different caching methods;

[0055] Figure 4This is a comparison chart of the energy consumption performance of different caching methods. Detailed Implementation

[0056] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0057] like Figure 1 As shown, the design concept of this invention is to construct a three-tiered heterogeneous network unit system with cache competition relationships: "base station – roadside unit – vehicle". The key competitive indicators are cache offloading latency, storage energy consumption, and economic benefits, with transmission latency and transmission energy consumption as the core optimization objectives. A further design approach involves: first, obtaining network parameters for highly dynamic vehicle scenarios based on the heterogeneous network units; second, constructing a Stackelberg game model to calculate data caching rate and cost, constructing a utility function, and determining the optimal pre-caching alliance members; and finally, designing a multi-dimensional constrained knapsack algorithm, combined with dynamic programming backtracking to generate the optimal pre-cached content combination and its deployment scheme.

[0058] The present invention will be further described in detail below with reference to specific embodiments and accompanying drawings.

[0059] The present invention provides a pre-assisted caching method based on vehicle location, the specific steps of which are as follows:

[0060] Step 1: Construct a high-dynamic scenario based on heterogeneous network units to meet different caching content requirements, and obtain relevant network parameters, including the locations of vehicles, base stations, and roadside units. This specifically includes the following sub-steps:

[0061] 1.1 Establishment of heterogeneous network units:

[0062] In this invention, highly mobile vehicles with different cache content requirements in heterogeneous network units, such as... Figure 2 As shown in the diagram, this architecture of heterogeneous network units is constructed based on the coverage of base stations, illustrating the architecture, game hierarchy, and communication buffer links between the core network, base stations (BS), roadside units (RSU), and vehicles. Each heterogeneous network unit consists of one base station, roadside units and Each vehicle (each vehicle has a unique identification ID, such as...) Composed of, among which, , ; , Each member of the heterogeneous network unit can provide corresponding cached content for user vehicles within its communication coverage area. Roadside units and vehicles can act as pluggable auxiliary cache nodes, providing parallel multi-link content caching and transmission services to optimize the storage efficiency of cached content. Base stations and roadside units store the basic data of the cached content, ensuring the minimum requirements for obtaining the cached content, while vehicles store enhanced data of the cached content, improving the high quality of the cached content.

[0063] 1.2 Different cached content based on differentiated needs:

[0064] In the constructed heterogeneous network unit architecture, vehicles dynamically request diverse cached content during road travel, such as audio, environmental perception data, and video applications. These contents exhibit varying sensitivities to transmission latency and energy consumption due to differences in data type, timeliness requirements, data volume, and popularity. Latency-sensitive content (such as high-definition video streams, real-time monitoring data, and emergency alarms) typically involves large data volumes or high timeliness requirements, necessitating priority for low-latency transmission. Energy-sensitive content (such as sensor status reports and offline map updates) generally has smaller data volumes and less stringent timeliness requirements, prioritizing energy efficiency optimization during transmission. The transmission strategy for balanced optimization content (such as software incremental upgrade packages and collaborative perception shared data) requires dynamic decision-making based on real-time network conditions and content attributes to achieve joint optimization of latency and energy consumption goals, thereby matching the dynamic environment of multi-node collaborative caching in the heterogeneous units.

[0065] 1.3 Determine the relevant network parameters, including the location information of vehicles, base stations, and roadside units:

[0066] Base stations and roadside units, as fixed infrastructure in heterogeneous networks, have location information (such as the location coordinates of base stations). The location coordinates of the roadside unit are ,in and These represent the horizontal and vertical coordinates of the base station, respectively. and The horizontal and vertical coordinates of the roadside units (representing the horizontal and vertical coordinates respectively) can be obtained through geographic information systems or pre-deployed planning data. These locations serve as the benchmark for calculating the distance between nodes and assessing the channel status.

[0067] Discretize the time domain into time slots The vehicle's travel path is divided into continuous sub-regions. The vehicle obtains its location in each time slot via GPS. The starting and ending coordinates of the sub-region in which it is located (e.g. and This discretized position sequence accurately describes the vehicle's trajectory.

[0068] By integrating the above location information, a dynamic network topology is constructed in real time. In subsequent processing, this location data will serve as the basis for analyzing the quality of communication links between nodes, calculating latency and energy consumption, and ultimately making optimization decisions.

[0069] Step Two: The base station, roadside unit, and vehicle are modeled as leaders and followers in a Stackelberg game model. Within this game framework, the data caching rate and cost of each network layer are calculated based on parameters such as vehicle location. These are then used as policy variables to design corresponding utility functions for each participant. Finally, by solving the established Stackelberg game model, the optimal composition of the pre-caching alliance is obtained, maximizing the overall system utility. This includes the following sub-steps:

[0070] 2.1 Base stations, roadside units, and vehicles are modeled as leaders and followers in the Stackelberg game model, respectively. Base stations pay fees to followers (roadside units and vehicles) for auxiliary cache storage. Under this game framework, the data caching rate and data caching cost of each network layer are calculated based on parameters such as vehicle location.

[0071] The data caching rate of each network layer in a heterogeneous vehicular network, that is, the data caching rate (downlink) of the caching tasks allocated by the base station to the auxiliary caching nodes, is calculated as follows:

[0072] Assuming the vehicle is in a time slot Located in sub-region Its starting position is The termination position is The vehicle is in the sub-area. The position inside can be represented as The midpoint is obtained by calculating the starting and ending positions within this sub-region. (Base station) To roadside unit European distance and base station To the vehicle European distance It can be calculated using the following formula:

[0073]

[0074] in, Indicates a secondary cache node. , .

[0075] The data buffering rate of each network layer is based on Shannon's formula and takes into account the rapid channel changes caused by the high-speed movement of vehicles, Rayleigh channel gain, path loss, and noise. The calculation is as follows:

[0076]

[0077]

[0078] in, For sub-region indexing, This represents the equivalent channel gain of the downlink. This represents the channel gain caused by Rayleigh fading. For other integrated propagation loss factors, This is the path loss index. This indicates that the base station is assigned to the auxiliary buffer node. The data caching rate of the caching task This indicates the channel bandwidth of the base station. This indicates the base station's transmit power. Indicates secondary cache node The noise power of the receiver.

[0079] In heterogeneous vehicular networks, the data caching cost at each network layer includes payment costs (fees paid by base stations to auxiliary caching nodes for caching tasks or revenue generated by auxiliary caching nodes) and energy consumption costs.

[0080]

[0081]

[0082]

[0083]

[0084] in, and These represent the data caching costs of the base station and the auxiliary caching node layers, respectively. The cost of paying for the auxiliary cache node layer network of heterogeneous vehicle-to-everything (V2X) networks. The energy consumption cost of each layer of the network, Indicates base station For auxiliary cache nodes The unit price set for allocating cached tasks. Indicates base station Assigned to secondary cache nodes The amount of cached tasks, Indicates secondary cache node The unit energy consumption coefficient, , The corresponding weights.

[0085] 2.2 Using the data caching rate and data caching cost from step 2.1 as policy variables, design a corresponding utility function for each participant, and use the total utility function of all participants as the optimization objective. By solving the established Stackelberg game model, obtain the composition of the optimal pre-caching alliance members. This maximizes the overall utility of the system.

[0086] The utility function for each participant is designed based on network performance (caching task volume, latency, effective service duration) and data caching cost as follows:

[0087]

[0088]

[0089]

[0090]

[0091]

[0092] in, Let be the total utility function of the participants. Let be the utility function of the base station. The utility function for a selected roadside unit or a selected vehicle. , To select roadside units and the number of vehicles, The value of currently allocated cached tasks, For the latency of the base station layer, For the latency of each network layer, To select a roadside unit or a vehicle's effective service duration within the current heterogeneous network unit, The weights are the corresponding to the time delay.

[0093] The optimization objective is:

[0094]

[0095]

[0096]

[0097]

[0098]

[0099]

[0100]

[0101]

[0102] in, Allocate the total cache task amount for the base station. This represents the total number of cached tasks. To allocate the amount of cached tasks to the roadside units, The amount of cached tasks allocated to vehicles. This represents the maximum total cache capacity acceptable to the roadside unit. The maximum acceptable total cache size for vehicles. To the maximum tolerable delay, This represents the total number of roadside units under the current heterogeneous network unit. This represents the total number of vehicles in the current heterogeneous network unit. Let P1 be the communication range of the base station. In general, P1 is the optimization objective, which represents maximizing the total utility function of the pre-caching consortium after the Stackelberg game. C1, C2, and C3 respectively constrain the allocated cache task amount to not exceed the maximum load, C4 constrains the maximum latency, C5 and C6 constrain the number of pre-cached RSUs and pre-cached vehicles, and C7 constrains the communication effectiveness of pre-caching consortium members.

[0103] Step 3: Based on the Stackelberg game model established in Step 2, calculate the caching efficiency by considering the timeliness, popularity, and data volume of the cached content. Also, calculate the cache node utility value of each cache node in the optimal pre-caching consortium members from Step 2. Using the total capacity of the consortium members as a constraint, design the objective function for the multi-dimensional knapsack problem, and then use dynamic programming to backtrack and solve it, thereby determining the optimal combination of pre-cached content and its deployment scheme among the optimal pre-caching consortium members. Specifically, this includes the following sub-steps:

[0104] 3.1 Based on the Stackelberg game model established in step two, calculate the caching efficiency by combining the timeliness, popularity and data volume of different cached content in step one; at the same time, evaluate the deployment priority of different cache nodes in the optimal pre-caching alliance members and calculate the cache node utility value of each cache node.

[0105] caching benefits The design is as follows:

[0106]

[0107]

[0108]

[0109] in, Indicates the first The caching benefits of pre-cached content This indicates the popularity of cached content, and is determined by calculating the number of requests generated by a vehicle within a specified time interval to understand the popularity of each piece of content. Indicates the cached content at time The timeliness score is dynamically calculated using an exponential decay model. Indicates the penalty factor for the size of cached content. , This represents the corresponding weighting coefficient; This indicates the initial timeliness score. The attenuation coefficient controls the rate at which time-sensitivity decreases. Indicates the time step. This represents the lower limit of timeliness, ensuring that timeliness does not decay indefinitely. Indicates the first The size of the cached content data.

[0110] Furthermore, to measure the comprehensive value of different cache nodes as cache carriers among the members of the optimal pre-caching consortium, the first... One cache node ( ) cache node utility value as follows:

[0111]

[0112] in, Indicates the vehicle density within the coverage area. Indicates the maximum vehicle density. Indicates the cache capacity used by the node. Indicates the maximum cache capacity of the node. Represents a node Total number of content requests within the coverage area. , , These are the corresponding weighting coefficients.

[0113] 3.2 The total capacity constraint of the optimal pre-cached alliance members in step 2 and the caching efficiency and cache node utility in step 3.1 are incorporated into a unified optimization framework. The problem of pre-cached content selection and cache node deployment is modeled as a multidimensional knapsack problem in combinatorial optimization problems, and an objective function is designed.

[0114] Constructed target function Represented as:

[0115]

[0116]

[0117] in, This indicates the total number of candidate pre-cached contents. This indicates the total number of nodes in the pre-cached consortium. Indicates the first The quality of service gain of pre-cached content is determined by its data volume. and expected average transmission rate Decide, As a binary decision variable, it determines the content. Whether to cache on node middle, , , This represents the corresponding weighting coefficient.

[0118] The optimization objective is:

[0119]

[0120]

[0121]

[0122] Where C1 represents the constraint on the node's storage space, Indicates the first The storage space required for each cached content; C2 represents the uniqueness constraint on the content, ensuring that for any given content... and any two distinct nodes and This content can only be stored on one node at most.

[0123] 3.3 A dynamic programming algorithm is used to backtrack and solve the multidimensional knapsack problem in step 3.2 to determine the optimal combination of pre-cached content under different caching requirements and its deployment scheme among the optimal pre-caching alliance members. The specific steps are as follows:

[0124] Constructing a dynamic programming matrix Among them, row dimension Indicates the preceding One candidate cached content item, column dimension The storage capacity constraint of the cache node is represented by the matrix element value. The record is before consideration Item cache content, within capacity limits The optimal value obtained below.

[0125] For each cached content item Its comprehensive value is determined by the objective function term in step 3.2, and the state transition equation is expressed as:

[0126]

[0127] in, Indicates the first The overall benefit generated when a cached content item is deployed on the corresponding cache node, the overall benefit being comprised of cache efficiency. Service quality gain and cache node utility value To be determined jointly.

[0128] After obtaining the dynamic programming matrix, the optimal solution is obtained using a reverse backtracking method. Specifically, the initial positioning points are... Construct the initial optimal cache content set ;from arrive Reverse the backtracking operation: If Then it means the first Each content item is selected and deployed to the corresponding cache node, and then added to the collection. At the same time, update the remaining capacity. for Otherwise, it means the first If no content item is selected, continue judging the previous content item; when... or At that point, the backtracking process ends.

[0129] Through the above dynamic programming solution and reverse backtracking process, the optimal combination of pre-cached content that satisfies the cache node capacity constraint and content uniqueness constraint, as well as the deployment position of each pre-cached content in the optimal pre-cached alliance member, can be obtained, thus forming the final pre-cached content deployment scheme.

[0130] The following specific simulation examples will be provided to illustrate the invention in detail:

[0131] In the experiment, a road of 1 km in length is considered, representing the road length covered by a heterogeneous network unit. This heterogeneous network unit consists of 1 BS, n RSUs, and m vehicles. The vehicles travel at a constant speed. Specific parameter settings for the communication range, bandwidth, transmit power, and operating frequency of the base station, roadside units, and vehicles are shown in Table 1.

[0132]

[0133] In the experiment, the cache node performed a pre-caching process by using Stackelberg game to build the optimal pre-caching alliance and an improved knapsack algorithm to improve the caching method, and finally obtained a pre-aided caching method based on vehicle location required for specific application scenarios.

[0134] See Figure 3The performance comparison (latency) of caching methods demonstrates the latency performance of four caching methods under different vehicle densities. The horizontal axis represents the request ratio, and the vertical axis represents latency. From the overall average performance perspective, in low-density scenarios, the Knap caching method proposed in this invention reduces the average latency by approximately 31.44% and 29.19% compared to random caching (randn) and non-cooperative caching (onlybs), respectively; in medium-density scenarios, it reduces the latency by approximately 31.40% and 29.98%, respectively; and in high-density scenarios, it reduces the latency by approximately 30.42% and 27.99%, respectively. Compared to the Deep Reinforcement Learning (DRL) caching method, Knap caching performs similarly to it in most request ratio ranges, with the latency difference generally controlled within 3%, demonstrating that the pre-caching method provided by this invention has strong low transmission latency capabilities.

[0135] See Figure 4 The performance comparison (energy consumption) of four caching methods under different network densities is shown in the caching method performance comparison (energy consumption). The horizontal axis represents the request ratio, and the vertical axis represents energy consumption. From the overall average performance perspective, in low-density scenarios, the Knap caching method reduces average energy consumption by approximately 28.34% and 30.32% compared to randn caching and DRL caching, respectively; in medium-density scenarios, the reductions are approximately 27.99% and 30.34%; and in high-density scenarios, the reductions are approximately 28.73% and 30.81%. On the other hand, compared to Onlybs caching, the Knap caching method's energy consumption performance is generally similar. Therefore, the pre-caching method provided by this invention can achieve better energy efficiency control while ensuring low latency, and has better overall performance.

[0136] The above description is a specific illustration of the present invention, and not a limitation thereof. Those skilled in the art can make various equivalent technical solutions without departing from the scope of the present invention; therefore, all equivalent technical solutions should be included within the protection scope of the present invention.

Claims

1. A pre-assisted caching method based on vehicle location, characterized in that, Includes the following steps: Step 1: Construct a high-dynamic scenario with different cache content requirements based on heterogeneous network units, and obtain relevant network parameters, including the locations of vehicles, base stations, and roadside units; Step 2: Model the base station, roadside unit, and vehicle as leaders and followers in the Stackelberg game model, respectively. Under this game framework, calculate the data caching rate and data caching cost of each network layer based on the relevant network parameters obtained in Step 1, and use these as policy variables to design corresponding utility functions for each participant. Finally, by solving the established Stackelberg game model, obtain the composition of the optimal pre-caching alliance members. Step 3: Based on the Stackelberg game model established in Step 2, calculate the caching efficiency by combining the timeliness, popularity and data volume of the cached content, and calculate the caching node utility value of each cache node in the optimal pre-caching alliance member obtained in Step 2. Using the total capacity of the alliance member as a constraint, design the objective function of the multidimensional knapsack problem, and then use the dynamic programming method to backtrack and solve it, thereby determining the optimal combination of pre-cached content and its deployment scheme in the optimal pre-caching alliance member.

2. The pre-assisted caching method based on vehicle location according to claim 1, characterized in that: In step two, the data caching cost of each network layer includes payment cost and energy consumption cost: in, and These represent the data caching costs of the base station and the auxiliary caching node layers, respectively. Indicates a secondary cache node. The cost of paying for the auxiliary cache node layer network of heterogeneous vehicle-to-everything (V2X) networks. The energy consumption cost of each layer of the network, Indicates base station For auxiliary cache nodes The unit price set for allocating cached tasks. Indicates base station Assigned to secondary cache nodes The amount of cached tasks, Indicates secondary cache node The unit energy consumption coefficient, , The corresponding weights.

3. The pre-assisted caching method based on vehicle location according to claim 2, characterized in that: In step two, the utility function is designed based on network performance and data caching costs as follows, where network performance includes caching task volume, latency, and effective service duration: in, Let be the total utility function of the participants. Let be the utility function of the base station. The utility function for a selected roadside unit or a selected vehicle. , Select the roadside unit and the number of vehicles; The value of currently allocated cached tasks, For the latency of the base station layer; For the latency of each layer of the network, To select a roadside unit or a vehicle's effective service duration within the current heterogeneous network unit, For data caching rate, The weights are the corresponding time delays. The optimization objective is: in, Allocate the total cache task amount for the base station. This represents the total number of cached tasks. To allocate the amount of cached tasks to the roadside units, The amount of cached tasks allocated to vehicles. This represents the maximum total cache capacity acceptable to the roadside unit. The maximum acceptable total cache size for vehicles. To the maximum tolerable delay, This represents the total number of roadside units under the current heterogeneous network unit. This represents the total number of vehicles in the current heterogeneous network unit. This indicates the distance from the base station to the auxiliary buffer node. Let P1 be the communication range of the base station. In general, P1 is the optimization objective, which represents maximizing the total utility function of the pre-caching consortium after the Stackelberg game. C1, C2, and C3 respectively constrain the allocated cache task amount to not exceed the maximum load, C4 constrains the maximum latency, C5 and C6 constrain the number of pre-cached RSUs and pre-cached vehicles, and C7 constrains the communication effectiveness of pre-caching consortium members.

4. The pre-assisted caching method based on vehicle location according to claim 3, characterized in that: In step three, the caching efficiency is calculated based on the timeliness, popularity, and data volume of the jointly cached content. as follows: in, Indicates the first The caching benefits of pre-cached content This indicates the popularity of cached content, and is determined by calculating the number of requests generated by a vehicle within a specified time interval to understand the popularity of each piece of content. Indicates the cached content at time The timeliness score is dynamically calculated using an exponential decay model. Indicates the penalty factor for the size of cached content. , This represents the corresponding weighting coefficient; This indicates the initial timeliness score. The attenuation coefficient controls the rate at which time-sensitivity decreases. Indicates the time step. This represents the lower limit of timeliness, ensuring that timeliness does not decay indefinitely. Indicates the first The size of the cached content data.

5. The pre-assisted caching method based on vehicle location according to claim 4, characterized in that: In step three, the objective function Represented as: in, This indicates the total number of candidate pre-cached contents. This indicates the total number of cache nodes among the pre-caching alliance members. Indicates the first The quality of service gain of pre-cached content is determined by its data volume. and expected average transmission rate Decide, Indicating the first in the pre-caching alliance The cache node utility value of each cache node. For binary decision variables, , , This represents the corresponding weighting coefficient. The optimization objective is: Where C1 represents the constraint on the node's storage space, Indicates the first The storage space required for each cached content; C2 represents the uniqueness constraint on the content, ensuring that for any given content... and any two distinct nodes and This content can only be stored on one node at most.

Images