CN115102974A

CN115102974A - Cooperative content caching method based on bilateral matching game

Info

Publication number: CN115102974A
Application number: CN202111488348.6A
Authority: CN
Inventors: 田淑娟; 邹松; 朱江; 杨云芳; 黄凌翔; 刘新杰
Original assignee: Xiangtan University
Current assignee: Xiangtan University
Priority date: 2021-12-08
Filing date: 2021-12-08
Publication date: 2022-09-23

Abstract

The invention provides a collaborative content caching method based on bilateral matching game. The method comprises the following steps: firstly, initializing a correlation value between vehicles according to a preference list of the requesting vehicles and a content request parameter; then, calculating a weighted value and network overhead of the vehicle according to a formula, and dynamically adjusting a correlation value list of the vehicle through a bilateral matching game algorithm; and finally, introducing a reinforced learning model to obtain an optimal content caching strategy. The cloud-side cooperative cache framework under the vehicle-mounted environment is constructed, cooperative cache of multiple devices is effectively utilized, and the cache hit rate is improved; and a bilateral matching game algorithm based on time delay perception is established by combining the high-speed dynamics of the vehicle, and the dynamic association between the equipment and the request vehicle is realized, so that a cache decision is made, and the content transmission time delay and the network overhead are reduced.

Description

Cooperative content caching method based on bilateral matching game

Technical Field

The invention relates to the technical field of edge caching and the field of game theory learning, in particular to a content caching decision method in Internet of vehicles.

Background

Under the concept of the internet of things, the internet of vehicles gradually appears in the field of vision of people, and the interconnection of vehicles and the interconnection of vehicles and everything are realized through the modern signal technology, so that the technologies of vehicle monitoring, intelligent path planning, automatic driving of vehicles and the like can be realized. For multi-input and multi-output-based multi-user and multi-task scenes needing to realize functions of real-time response, location awareness, context awareness and the like, the longer distance between the cloud server and the user equipment often brings larger communication delay, so that the performance bottleneck of the outsourcing system is formed. Meanwhile, for a vehicle moving at a high speed, since a wireless link has limited capacity and low stability, transmitting content from a cloud or a core network to the vehicle may cause high delay and low quality of service. In addition, the edge cache is used as an extension of a cloud center network, and still has performance optimization problems in some cloud networks. Therefore, an efficient collaborative caching mechanism is designed, real-time requests of vehicles are met, and the improvement of the satisfaction degree of vehicle-mounted users is still difficult and serious.

Disclosure of Invention

Aiming at the limitation of the existing content caching method, the invention provides a content collaborative caching algorithm research method based on bilateral matching game. The method can effectively improve the transmission efficiency of the edge server to the cache content task and effectively improve the service quality of the system.

The invention provides a cooperative content caching method based on bilateral matching game, which mainly comprises the following steps: an initialization stage, an association optimization stage and a cache decision optimization stage. The initialization phase comprises initializing the association values between the vehicles according to the preference list and the content parameters of the requesting vehicle; the association optimization stage comprises the steps of respectively calculating network time delay generated by requesting the vehicle to request contents of different devices by using a formula according to the high-speed mobility of the vehicle, and dynamically updating an association value list of the vehicle according to a bilateral matching game algorithm; the caching decision optimizing stage comprises the step of inputting the popularity of the content into a reinforced learning model to obtain the optimal content caching decision.

The invention has the following advantages:

1. the invention can effectively improve the transmission efficiency of the edge server to the cache content task and effectively improve the service quality of the system;

2. the invention can meet the requirement of online pre-caching of the vehicle in the process of high-speed motion.

Drawings

FIG. 1 is a flow chart of the present invention;

FIG. 2 is a schematic diagram of the content request phase of the present invention;

FIG. 3 is a schematic illustration of the vehicle association optimization phase of the present invention;

FIG. 4 is a diagram illustrating the cache decision optimization stage of the present invention.

Detailed description of the invention

The invention designs a cooperative content caching method based on bilateral matching game, which is combined with the graph 1, the graph 2, the graph 3 and the graph 4, and the specific implementation method is as follows:

1. step one-initialization phase

1.1. Requesting vehicle RV _n Sending out content request, generating content request parameter

1.2. Calculated transmission delay resulting from requesting vehicle to obtain content from RSU

1.3.UE _n According to

And

to determine whether to cache the vehicle CV _k Whether or not to associate with an RSU _m And associating the transmission content.

2. Step two-Association optimization phase

2.1. According to the formula

Calculate request vehicle RV _n According to the calculation result, the RV is calculated according to the preference values of different application types _n Sorting in descending order to list preference list A _n ；

2.2. Requesting vehicle RV _n Buffer vehicle CV _k Issuing a content request according to a formula

Calculating content transmission delay, and determining CV according to the calculation result _k Sorting in descending order to list preference list B _n ；

2.3. According to preference list A _n And B _n Vehicle prioritization of requesting vehicles and cached vehicles CV _k Performing dynamic RV _n Is associated with, if

Requesting vehicle and RSU _m Correlation, to buffer vehicle CV _k And removing the preference list and updating the preference list.

3. Step three-cache decision optimization phase

3.1. Due to the high-speed mobility of the vehicles, the cache vehicle CV is continuously updated by adopting a bilateral matching game algorithm according to time delay _k A preference list and associated values;

3.2. during the time period t, the content request number gamma will be predicted _S ＝{γ ₁ ，γ ₂ ，...，γ _q ，...，γ _S Is input into the Zipf model

In (1), output content request probability C _S ＝{C ₁ ，C ₂ ，...，C _q ，...，C _S }；

3.3. Inputting the content request probability into a reinforcement learning model according to the association value list of RVs and CVs, and outputting an optimal caching decision vector X _S，K ＝{x _1，1 ，x _2，2 ，...，x _q，k ，...，x _S，K }。

Claims

1. A cooperative content caching algorithm research method based on bilateral matching game is provided, an optimal content pre-caching strategy is realized, and the method at least comprises the following steps:

s1, constructing a vehicle-mounted cloud system, wherein the system comprises a base station, an RSU, a cache vehicle and vehicle user sides which are sequentially connected with a cloud end, and a plurality of vehicle user sides can request the base station, the RSU and the cache vehicle for transmitting contents;

s2, initializing and associating the user vehicle with different edge nodes by using the preference list and the content request parameters of the request vehicle;

s3, according to the vehicle cloud system network model, the vehicle user sends a content request to nearby cache vehicles, RSUs and base stations;

s4, optimizing the correlation value by using a bilateral matching game algorithm based on time delay perception according to the real-time dynamics of the vehicle user;

and S5, inputting the optimized association values and the popularity analysis of the contents into a reinforcement learning model, and outputting an optimal caching decision.

2. The method for researching cooperative content caching algorithm based on bilateral matching game as claimed in claim 1, wherein said S2 further comprises the following steps:

calculating the global weight value of the request vehicle by utilizing different content application types requested by vehicle users, arranging and adding the weight value into a weight list of the request vehicle in a descending manner, respectively calculating the content transmission time delay of the request vehicle, the cache vehicle and the RSU, and associating the request vehicle with the cache vehicle if the time delay generated by transmitting the content to the cache vehicle is less than the time delay generated by transmitting the content from the RSU; otherwise, the requesting vehicle is associated with the nearest RSU.

3. The method for researching cooperative content caching algorithm based on bilateral matching game as claimed in claim 1, wherein between S3 and S4 at least the following steps are included:

3.1. calculate requested vehicle RV _n For preference value of requested content type, RV is calculated according to calculation result _n Sorting in descending order to list preference lists;

3.2. requesting vehicle RV _n Buffer vehicle CV _k Sending out content request, calculating content transmission delay, and determining CV according to the calculation result _k Sorting in descending order to list preference lists;

3.3. vehicle prioritization of the list of preferences for the requesting vehicle Rv _n And buffer vehicle CV _k And performing dynamic association, updating the preference list, and optimizing the association value according to the preference lists of the request vehicle and the cache vehicle.

4. The method for researching cooperative content caching algorithm based on bilateral matching game as claimed in claim 1, wherein said S5 further comprises at least the following steps:

4.1. due to the high-speed mobility of the vehicles, the cache vehicle CV is continuously updated by adopting a bilateral matching game algorithm according to time delay _k A preference list and associated values;

4.2. after the vehicles are associated, the sequencing index of the content request number is input into a Zipf model, and the content request probability C is output _s ；

4.3. According to RV _n And CV _k The association value list inputs the content request probability into the reinforcement learning model and outputs the optimal caching decision vector.