A ris-assisted d2d caching network task offloading modeling method

Abstract

The application discloses a kind of RIS auxiliary D2D cache network task unloading modeling methods, belong to the technical field of queue modeling, reconfigurable intelligent surface (RIS) technology is combined with D2D cache network, content unloading includes two methods of self-unloading and D2D unloading, considering that auxiliary cache technology combines the double characteristics of user terminal cache and D2D communication, i.e., content unloading includes two methods of self-unloading and D2D unloading, content transmission needs to satisfy the double constraints of receiver signal-to-interference ratio and D2D distance.A stochastic service model is established based on queuing theory, combining the performance advantages of RIS and the advantages of D2D cache network, a queuing model is established.An M / G / 1 queuing model is established for the user's cache queue, and the transition of its queue state is analyzed.A Markov chain containing queue length state and transmission state is then constructed, and the closed-form expressions for successful unloading probability, average queue length and average waiting time of the cache network are derived.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of queue modeling, and particularly relates to a RIS assisted D2D cache network task offloading modeling method. BACKGROUND

[0002] With the commercialization of 5G and the development of 6G, accompanied by the advent and development of the era of self-media, mobile Internet traffic has shown an explosive growth trend. Device-to-device (D2D) communication is well known as one of the key technologies of 5G. In D2D communication, users who are physically close to each other can communicate directly without forwarding through a base station (BS). Device-to-device (D2D) communication has significant advantages such as short distance and low energy consumption. By utilizing D2D communication for caching on mobile user devices, this method combines the advantages of caching and D2D communication technology, reduces the heavy load on the backhaul link, and improves the throughput of the cellular network.

[0003] D2D communication effectively improves the spectral efficiency of the communication network by reusing the cellular spectrum resources, but also introduces new interference to the cellular network. Due to the sharing of the same frequency band, there is serious co-channel interference between the D2D communication link and the cellular link. A software-defined metasurface called reconfigurable intelligent surface (RIS) is proposed to solve this problem, which can be used to effectively eliminate D2D interference and meet stringent data rates. This can be achieved by configuring the RIS to control the reflection, refraction, and scattering of electromagnetic waves hitting the surface. Combining RIS with D2D can further improve the performance of future mobile communication systems. Many documents have conducted in-depth research on resource allocation in RIS assisted D2D communication systems.

[0004] Reconfigurable intelligent surface (RIS) can effectively control the phase of the signal with low-cost reflective elements, and it has great practical significance to apply this technology to D2D communication networks that are highly dependent on the propagation environment. However, it is very challenging to integrate RIS into D2D caching communication, and research on this topic has not been fully developed. In addition to the direct path, RIS also brings a reflected path, which not only provides diversity gain, but also enhances the interference between users. Modeling the traffic of D2D caching networks in this complex interference scenario is important but difficult. How to model and analyze D2D networks in dynamic environments and achieve resource optimization, the existing scheme cannot be directly applied. SUMMARY

[0005] In view of the above technical problems, the purpose of the present application is to provide a RIS-assisted D2D cache network task offloading modeling method, taking into account the influence of dynamic interference on the successful transmission probability of the cache network, the complementary cumulative distribution function of the user SINR is derived, and it is used as the service probability of the queue, a two-dimensional Geo / G / 1 Markov chain is established for the user requesting the file, and the transfer state of successfully transmitting the file through the D2D link and the cellular base station is analyzed by using the quasi-birth-and-death process, and the expressions of the key performance parameters such as system delay and successful offloading probability are obtained.

[0006] In order to achieve the above purpose, the technical scheme adopted by the present application is as follows: 1. A RIS-assisted D2D cache network task offloading modeling method, comprising the following steps:

[0007] Step 1: For the cache network scene of multiple D2D and multiple cellular users, reconfigurable intelligent surface is applied to D2D cache network, by analyzing the direct link interference and reflection link interference in RIS-assisted D2D network, the expression of SINR is derived, and the base station transmission success probability is obtained;

[0008] Step 2: For the communication system with all mobile users having cache capability, combining the double constraints of content hit and satisfying the SINR condition required for successful transmission of content, a file acquisition protocol is given, that is, content offloading includes self-offloading and D2D offloading;

[0009] Step 3: M / G / 1 queuing model is established at the request user end, considering the cache hit rate of the file and the actual interference, the change of the queue state is analyzed;

[0010] Step 4: Based on the generality of service time, the queues for storing D2D transmission files and base station transmission files are established, and two-dimensional discrete-time Markov chains of the two states are constructed;

[0011] Step 5: The quasi-birth-and-death process method is used to solve the steady-state probability distribution of the Markov chain, and the expressions of the key performance parameters such as successful offloading probability, average queue length and average waiting time of the cache network are obtained.

[0012] Compared with the prior art, the present application has the following beneficial effects:

[0013] (1) The present application applies reconfigurable intelligent surface to D2D cache communication system, analyzes the direct link interference and reflection link interference in RIS-assisted D2D communication system, and derives the expression of SINR, which represents the signal-to-interference-plus-noise ratio.

[0014] (2) The application is directed to a D2D-assisted mobile caching network, and the D2D caching technology combines the double characteristics of user terminal caching and D2D communication, that is, content offloading includes two offloading modes of self-offloading and D2D offloading, and content transmission needs to meet the double constraints of the receiving end signal-to-interference noise ratio and D2D distance. A user caching Geo / G / 1 queue model is established to analyze the transmission process of the requested file.

[0015] (3) The application constructs the queue state of the file as a Markov chain, analyzes the file queue state in different transmission modes by using a quasi-birth-and-death process, and obtains expression formulas of key performance parameters such as the successful offloading probability. BRIEF DESCRIPTION OF DRAWINGS

[0016] Figure 1 It is a schematic diagram of the file receiving model in the embodiment.

[0017] Figure 2 It is a state transition diagram of the file queuing model in the cache in the embodiment.

[0018] Figure 3 It is a curve graph of the queue length with the request arrival rate in the embodiment. Queue length and variation curve.

[0019] Figure 4 It is a curve graph of the average delay with the request arrival rate in the embodiment.

[0020] Figure 5 It is a curve graph of the self-offloading probability with the request arrival rate in the embodiment. Queue length and variation curve. DETAILED DESCRIPTION

[0021] In order to better understand the technical scheme of the application, the following will be further explained and described in combination with the drawings and specific embodiments.

[0022] The embodiment considers a distributed RIS-assisted D2D communication caching network, a caching network scene of multiple D2D and multiple cellular users, assumes that the channel resources of a cellular user (CU) are shared by at most one pair of D2D links, in order not to lose generality, the embodiment simplifies the model, and focuses on the single-cell scene, as shown in Figure 1 , which contains a base station, 1 RIS, CU and S (S≤K) pairs of D2D users, the cellular user connected with the base station is represented as CU, , the D2D pair contains a D2D transmitting end DT, and a D2D receiving end DR, The base station, D2D users, and CU are all assumed to be equipped with a single antenna. The RIS contains M reflecting elements. The reflection coefficient of each reflecting element in the RIS is adjusted through an intelligent controller.

[0023] Furthermore, this embodiment considers the downlink communication link scenario. The sub-channels allocated by the base station to the CU are mutually orthogonal, resulting in no interference between CUs. The channel resources of a CU can be shared by at most one pair of D2D links. For cellular users, they can receive not only direct signals from the D2D transmitter and the base station, but also reflected link signals after RIS reflection. The base station, upon receiving the superimposed signals, treats the signal from the D2D transmitter as interference. Similarly, the D2D receiver can receive direct link signals from the D2D transmitter and the base station, as well as reflected link signals after RIS reflection. The D2D receiver treats the signal from the base station as interference.

[0024] D2D user transmit power is expressed as: The receiving end DRs demodulates the signal received from DTs. In DRs, D represents D2D user, R represents receiver, and s represents the s-th user. In DTs, D represents D2D user, T represents transmitter, and s represents the s-th user. The received SINR value is then used. Represented as:

[0025] (1)

[0026] in, and They represent DT respectively To DR and BS to DR direct link channel, and They represent BS and DT respectively. The signal transmission power; BS stands for Base Station. and They represent from DT The reflection channels from the BS to the IRS and from the BS to the IRS, where IRS stands for Smart Reflector. Indicates from IRS to DR The reflection channel; assuming the CSI of all channels is known, CSI stands for Channel State Information. The phase shift matrix of RIS is represented as... , Represents a diagonal matrix. express Complex matrix, For complex fields, It is the m-th diagonal element of a diagonal matrix. This represents the modulation coefficient of the signal amplitude by the m-th unit. It is in complex exponential form, corresponding to the modulation of the signal phase by the m-th unit. It is the phase offset. The imaginary unit is used. Based on the continuous phase-shift reflection coefficient model, then... , ; CU With the The link multiplexing factor between D2D user pairs, where CU stands for Cellular User, refers to cellular users. Reuse CU When resources are available, =1, otherwise =0; Indicates the receiving end DR The power of the additive white Gaussian noise (AWGN) received at the location.

[0027] For cellular users, they can receive communication signals from base stations, but they will also receive interference signals. The SINR of the demodulated signal from the base station can be expressed as:

[0028] (2)

[0029] in, and They represent BS to CU respectively and DT To CU direct link channel, and These represent the routes from BS to IRS and from DT, respectively. Reflection channel to IRS Indicates from IRS to CU The reflection channel.

[0030] This embodiment considers heterogeneous cellular networks with D2D communication, where the base station locations follow an independent and identically distributed PPP model. The density is Cellular users and D2D users respectively follow PPP and The densities are respectively and They are independent and identically distributed. Assuming all users have caching capabilities, if the distance between the requesting user and the user storing the target file is less than [a certain value]... D2D communication link can be established between users to transmit files to the requesting user.

[0031] The multimedia file set is where is the total number of files, is the file set , and the dimension of the file set is , all multimedia files are set to the same size , the probability of a user requesting any one of the files obeys the popularity distribution of the file , it is assumed that obeys the Zipf distribution, which is expressed as:

[0032] (3)

[0033] where reflects the skewness of the popularity function, and a higher indicates greater file reusability. The physical meaning of is the probability that the requested file is , and the first few files with smaller values have greater popularity, which also corresponds to most requests.

[0034] Define as the probability set of each file stored by the user, where represents the probability that the user stores the file , and , since the user's cache capacity is limited, then must satisfy:

[0035] (4)

[0036] In the D2D caching network, the end user can be both a content cache and a content requester. When the requesting user has content request requirements, it can obtain the required content through three ways: local caching, caching other users within the effective range of the cache D2D, and through the backhaul link from the core network by the cellular base station, and the successful transmission of the content needs to meet the dual constraints of content hit and SINR condition. The file acquisition protocol of the embodiment includes the following three parts:

[0037] Self-request cache hit probability: since each user randomly caches a part of all files, the probability of finding the requested file in its own cache. The embodiment defaults to the self-unloading mode, and as long as the local cache hits, the target content can be obtained without error,

[0038] (5)

[0039] D2D cache hit probability: the probability that the requested file is not cached in its own memory, but in the memory of a device within a certain D2D communication distance R. If at least one user has the requested content, the user will establish D2D transmission with the nearest user and the cached content. File Probability of being hit by D2D users is expressed as:

[0040] (6)

[0041] When the received signal-to-interference ratio of the requesting user is not less than the decoding threshold, the content transmission is considered successful, and the probability of successfully transmitting the file through D2D communication is represented. Considering the cache hit and the success of content transmission, in the D2D offloading method, when the content meets the requirements, the location proximity, and the received signal-to-interference ratio is greater than the decoding threshold, the D2D offloading is considered successful, and its probability can be expressed as:

[0042] (7)

[0043] where represents the SINR of the requesting user end is not less than the minimum threshold for successful decoding information The probability is expressed as

[0044] (8)

[0045] where represents the probability of occurrence of event .

[0046] Base station transmission success probability: when the requested file cannot be found through the above two methods, assuming that the requesting user can always obtain the content from the core network through the backhaul link from the cellular base station, the success probability of obtaining content from the base station is the link SINR greater than the decoding threshold, which is expressed as:

[0047] (9)

[0048] The queuing modeling and performance analysis are described in detail as follows:

[0049] This embodiment establishes an M / G / 1 queuing model at the requesting user end, where M represents a Poisson arrival process, and G represents a service sending process subject to a general distribution, while considering the cache hit rate and actual interference of the file, and analyzing the change of its queue state. Then, based on the general service time, a two-dimensional Markov chain including the queue length state and the transmission state is constructed. Assuming that the request arrival follows a Bernoulli distribution with parameter , and the total capacity of the user's cache is A joint request-caching strategy is analyzed.

[0050] Let the queue storing the files transmitted by D2D be called and the queue storing the files transmitted by the base station be called . According to the aforementioned caching strategy, if a user requests a file and does not have it randomly stored, the user needs to obtain it from surrounding users in a D2D manner or send it by the base station. In this embodiment, a queuing model is established at the request user end, and the changes in the queue length of the file transmitted by the D2D link and the queue length of the file transmitted by the base station are discussed respectively. The state set of the two-dimensional Markov chain is defined as .

[0051] As Figure 2 is the state transition diagram, the behavior of the queue is evaluated by a quasi-birth-and-death process (QBD). The state of the classical birth-and-death process is decomposed into sub-states . This structure complies with the requirements of depicting the evolution of the process under the conditions of multiple levels, multiple phases, and variable parameters. In this embodiment, the states of the queue length are rearranged, where and represent the queue length of the file transmitted by the D2D link and the queue length of the file transmitted by the base station respectively, and the queue length is regarded as a one-dimensional process . Define , where is named as the level, is the phase.

[0052] Therefore, the corresponding probability transition matrix is a cell matrix, which is expressed as:

[0053] (10)

[0054] Each element in the cell matrix is also a matrix. The transition between the levels is described by the non-zero matrix , the sub-matrix on the diagonal represents the event set that the number of (levels) does not change in one time slot, while the number of (phases) increases by 1, decreases by 1, or remains unchanged, i.e. . The sub-matrix below the diagonal represents the event set that the number of (levels) decreases by 1 after one service period, while the number of (phases) remains unchanged, i.e. . represents Sub-matrix contains all events of adding a file.

[0055] The expression of all elements in the sub-matrix is given below. First, for the boundary condition , , , and remains unchanged during a time slot, while changes contain the following three cases:

[0056] Case 1: remains unchanged, i.e. , indicating that the user does not initiate a request with a probability of , or generates a request, and the stored file can meet it, in addition, the event that the file fails to be successfully transmitted from the core network through the D2D link or backhaul link will also cause .

[0057] Case 2: increases a file, i.e. , indicating that the request is generated and cannot be satisfied by the self-cached file, and the request user still cannot find a matching content cache user in the effective D2D communication range, and finally the event set of successful transmission and saving by the base station.

[0058] Case 3: decreases a file, i.e. , indicating the event of deleting the request after being satisfied.

[0059] (11)

[0060] where

[0061] (12)

[0062] (13)

[0063] (14)

[0064] For , indicates the change of state when .The difference from is that because in , when a request is generated, in addition to checking whether the user's pre-stored file meets the request, it is also necessary to check whether the file stored in the D2D queue and the BS queue has the required file.

[0065] (15)

[0066] (16)

[0067] (17)

[0068] (18)

[0069] For the boundary conditions of the sub-matrix ,

[0070] (19)

[0071] Wherein:

[0072] (20)

[0073] (21)

[0074] Contains Reduce all events of a file, this request only occurs when the request, itself pre-stored can not meet, The existing file can meet.

[0075] (22)

[0076] Wherein:

[0077] (23)

[0078] In the same way, Indicates . The sub-matrix Contains Increase all events of a file, which occurs when the request cannot be met by the pre-stored file, there is no file in the two queues, and the user requests and transmits successfully to the surrounding users.

[0079] (24)

[0080] Wherein:

[0081] (25)

[0082] Substitute formulas (11)-(23) into formula (10) to obtain the transition probability matrix of the queuing model built in this embodiment.

[0083] The steady-state distribution of this quasi-birth-death process is then solved. Since the queuing system reaches stability after a long period of operation, it is assumed that… It is the transition probability matrix The column matrix of the steady-state probability distribution, in order to utilize the block matrix The rule structure, in this embodiment will Based on the level of the pseudo-birth and death process, it is divided into:

[0084] (26)

[0085] in Given normalization and probability, utilizing... The steady-state distribution can be obtained by solving the following set of matrix equations

[27] :

[0086] (27)

[0087] The following embodiment defines several important performance parameters. First, considering the uncertainty of the content requested by a typical user, the probability of successful uninstallation can be expressed as:

[0088] (28)

[0089] and The average team leader is expressed as follows:

[0090] (29)

[0091] (30)

[0092] Given a steady-state probability distribution, and Average throughput is defined as and Their respective expressions can be derived as:

[0093] (31)

[0094] (32)

[0095] Little's Law can be used to evaluate the average waiting time required for two queues to send data packets. and Represented as:

[0096] (33)

[0097] (34)

[0098] Numerical and simulation results:

[0099] This embodiment uses MATLAB simulation to verify the correctness of the results obtained under the above caching strategy. Consider a 150 m × 150 m two-dimensional square cell with corners (0,0), (150,0), (0,150), and (150,150). The base station is deployed at (0,0), and the RIS is deployed near the base station at (5,0). K CUs are evenly distributed in the square cell. S pairs of D2D users are distributed in an area more than 20 m away from the BS. The distance between the D2D transmitter and receiver is 10–50 m. The user's transmit power is 20. Path loss fading coefficient The Gaussian white noise power is -104. Content library , and Cache space is , The user's position is fixed during the simulation, and user mobility is not considered.

[0100] Figure 3 Given Three parameters The graph shows the captain's position as the request arrival rate changes, and the theoretical value matches the simulation results, proving the correctness of the theoretical derivation. (RIS-assisted scheme) The rapid saturation of the RIS as the number of reflection units increases demonstrates that deploying RIS in wireless networks mitigates channel interference and enables faster file retrieval via D2D links, especially with increased request arrival rates.

[0101] Figure 4 The path loss factor is given. Under the change, the effect curve of average latency on request arrival rate is as follows, and and compared to, The average length reaches saturation faster, the path loss factor is smaller, the average throughput is lower, the average delay is longer, and the packet loss rate is higher. The results show that a higher path loss factor reduces the impact of interfering links on the D2D receiver and improves the probability of successful file transmission. The figure shows that the theoretical values ​​and simulation results are in good agreement.

[0102] from Figure 5 As can be seen from The increase in the probability of self-unloading This increase is because it enables the establishment of more D2D links to provide files to requesting users. The requested files are stored in the user cache, increasing the probability of self-unloading. It is shown that the larger the file is, the more reusability it has. The larger the file is, the more likely it is to be requested by users, and the higher the probability of self-unloading is. The numerical calculation and computer simulation are basically consistent.

[0103] In this embodiment, the reconfigurable intelligent surface is applied to the D2D cache communication system, the direct link interference and the reflection link interference in the RIS-assisted D2D communication system are analyzed, and the expression of the SINR is derived. For the D2D-assisted mobile cache network, the D2D cache technology combines the double characteristics of user terminal cache and D2D communication, that is, the content offloading includes two offloading modes of self-unloading and D2D offloading, and the content transmission needs to meet the dual constraints of the receiving end signal-to-interference-and-noise ratio and the D2D distance. A user cache Geo / G / 1 queue model is established to analyze the transmission process of the requested file. The queue state of the file is constructed as a Markov chain, the quasi-birth-and-death process is used to analyze the file queue state under different transmission modes, the expressions of the successful offloading probability, the average queue length and the average waiting time of the cache network and other key performance parameters are obtained, the simulation results and the numerical calculation are basically consistent, and the correctness of the model is illustrated.

[0104] Although the present application has been disclosed with the preferred embodiments as above, the embodiments and the drawings are not intended to limit the present application, and any person skilled in the art can make various changes or modifications without departing from the spirit and scope of the present application, and the changes or modifications are also within the protection scope of the present application. Therefore, the protection scope of the present application should be defined by the protection scope of the claims of the present application.

Claims

1. A RIS-assisted D2D cache network task offloading modeling method, characterized in that: The method comprises the following steps: Step 1: for the cache network scenario of multiple D2D and multiple cellular users, the reconfigurable intelligent surface is applied to the D2D cache network, the expression of SINR is derived by analyzing the direct link interference and reflection link interference in the RIS assisted D2D network, and the base station transmission success probability is obtained; Step 2: for the communication system with all mobile users having cache capability, a file acquisition protocol is given by combining the double constraints of content hit and SINR condition required for successful transmission of content, and the content offloading includes self-offloading and D2D offloading; Step 3: an M / G / 1 queuing model is established at the request user end, the cache hit rate of the file and the actual interference are considered, and the change of the queue state is analyzed; Step 4: based on the generality of service time, queues for storing D2D transmission files and base station transmission files are established, and a two-dimensional discrete time Markov chain of the two states is constructed; Step 5: the steady-state probability distribution of the Markov chain is solved by using the quasi-birth-and-death process method, and the expression of the key performance parameter is obtained; The step 1 specifically comprises: D2D user transmit power is denoted as , the receiving end DR The received signal from the DT is demodulated, and the DR D represents D2D user, R represents receiver, The s-th user is denoted as DT D represents D2D user, T represents transmitter, The s-th user is denoted as DT The received SINR value is denoted as: (1) wherein, and respectively represent the DT to DR and BS to DR direct link channels, and respectively represent the signal transmit power of BS and DT , BS represents Base Station; and respectively represent the reflection channels from DT to IRS and from BS to IRS, IRS represents Intelligent Reflecting Surface, represents the reflection channel from IRS to DR ; assuming that the CSI of all channels is known, CSI represents Channel State Information, the phase shift matrix of RIS is represented as , represents a diagonal matrix, represents a complex matrix of , is a complex field, is the mth diagonal element of the diagonal matrix, represents the mth unit control coefficient of signal amplitude, is a complex exponential form, corresponding to the mth unit control of signal phase, is a phase offset, is an imaginary unit; Based on the continuous phase shift reflection coefficient model, then , ; denotes the CU link reuse factor with the th D2D user pair, CU denotes Cellular User, which means cellular user, when the D2D user pair reuses the resource of CU , then =1, otherwise =0 ; denotes the power of the Additive White Gaussian Noise (AWGN) received at the receiving end DR . For a cellular user, it is able to receive a communication signal from a base station, while also receiving an interference signal, CU The SINR of demodulating the signal from the base station can be represented as: (2) wherein, and denote the direct link channel from the BS to the CU and the DT to the CU , and denote the reflection channel from the BS to the IRS and from the DT to the IRS, denote the reflection channel from the IRS to the CU . 2.The RIS-aided D2D caching network task offloading modeling method of claim 1, wherein, The step 2 specifically comprises: Heterogeneous cellular networks based on D2D communication, base station location distribution obeys PPP , PPP represents Poisson distribution, density is , cellular users and D2D users respectively obey PPP and , density is and , independent and identically distributed; assume that all users have cache function, if the distance between the requesting user and the user who has the target file is less than , D2D communication link is established between users, and the file is transmitted to the requesting user.

3. The RIS assisted D2D cache network task offloading modeling method according to claim 2, characterized in that: A set of multimedia files is given where is the total number of files, is the set of files with dimension , all multimedia files are of the same size , the probability of a user requesting any one file is subject to the file's popularity distribution function , the function is assumed to follow a Zipf distribution, expressed as: （3） wherein reflects the bias of the popularity function, high indicates large file reusability; denotes file number, the physical meaning is the probability, the first few files with small values have large popularity, corresponding to the majority of requests, is a normalization factor to ensure that the sum of the probabilities of all elements is 1, satisfying the basic properties of probability distribution, denotes the normalization parameter; definition The user has a set of probabilities for each file, where This indicates that the user has a file. The probability is that Since the D2D user buffer has limited capacity, Must meet: （4）。 4. The RIS assisted D2D cache network task offloading modeling method according to claim 3, characterized in that: In the D2D cache network, the end user can be a content cache or a content requester; when the request user has a content request demand, the required content can be obtained through three ways of local cache, cache D2D effective range user and core network through backhaul link from the cellular base station, and the successful transmission of the content needs to meet the double constraints of content hit and SINR condition; The file acquisition protocol comprises: Probability of cache hit from request: probability that a requested file can be found in one's own cache since each user caches a random portion of all files For: (5) D2D cache hit probability: the probability that the requested file is not cached in its own memory, but in the memory of a device within a certain D2D communication distance R; if at least one user has the requested content, the user will establish D2D transmission with the nearest user and the cached content; file Probability of being hit by a D2D user is represented as: (6) When the received signal to interference ratio of the requesting user is not less than the decoding threshold, the content transmission is considered successful, and represents the probability of successful transmission of the file through D2D communication; considering the cache hit and the success of the content transmission, in the D2D offloading method, when the content meets the requirements, the location proximity, and the received signal to interference ratio is greater than the decoding threshold, it is considered that the D2D offloading is successful, and its probability can be represented as: (7) wherein SINR of the requesting user equipment not less than a minimum threshold of successfully decoded information the probability of the requesting user equipment (8) wherein represents the probability of occurrence of an event ; Base station transmission success probability: assuming that the request user can always obtain the content from the core network through the backhaul link from the cellular base station, the successful transmission probability of the content obtained from the base station is whether the link SINR is greater than the decoding threshold, which is represented as: (9)。 5.The RIS-aided D2D caching network task offloading modeling method of claim 1, characterized in that, The steps 3 and 4 specifically comprise: Assume that the queue storing the file transmitted by D2D is called , and the queue storing the file transmitted by the base station is called ; when a user requests a file, if the user does not have the file randomly stored, the user needs to acquire the file from the surrounding users in a D2D manner or send the file by the base station, and the user terminal establishes an M / G / 1 queuing model, wherein M represents a Poisson arrival process, G represents a service sending process subject to a general distribution, and 1 indicates the number of service counters; the changes of the queue length of the file transmitted by the D2D link and the queue length of the file transmitted by the base station are discussed respectively, and the state set of the two-dimensional Markov chain is defined as ; rearrange the queue length of the states, where and denote the queue length of the file queue for D2D link transmission and the queue length of the file transmitted by the base station and consider it as a one-dimensional process ; define where denoted by horizontal, is the phase; Therefore, the corresponding probability transition matrix is a cell matrix, which is represented as: (10) Cell matrix Each element in the matrix is also a matrix; the transition between the horizontal is represented by a non-zero matrix Description, diagonal The sub-matrix on the diagonal represents the events in one time slot The horizontal number is constant, while The set of events that add 1, subtract 1 or remain constant in phase, i.e. The sub-matrix below the diagonal represents the events after one service period, Decreasing a file, while The set of events that add 1, subtract 1 or remain constant in phase, i.e. ; represents The sub-matrix contains All events that increase a file.

6. The RIS assisted D2D cache network task offloading modeling method according to claim 5, characterized in that: All sub-matrices The expression of the middle element is: first, for The boundary condition of For the case of and remains constant for one time slot, while The change contains the following three cases: Case 1: remains unchanged, i.e. , indicating that the user did not initiate a request with probability, or generated a request, the file stored by itself can meet, in addition, the event that the file is not successfully transmitted through the D2D link or the backhaul link from the core network also leads to ; Case 2: Add a file, i.e. , indicating the event set that the request is generated and cannot be satisfied by the self cache, and the request user still cannot find the matching content cache user in the effective D2D communication range, and finally the base station transmission is successful and saved. Case 3: decrease a file, i.e. , indicates an event that is deleted after the request is satisfied; ​ (11) Wherein (12) (13) (14) For , indicates when , the change of state; unlike , because in , when a request is generated, in addition to checking whether the user's pre-stored file meets the requirement, it also checks whether there is a required file in the D2D queue and the BS queue. (15) （16) (17) (18) For the boundary conditions of the sub-matrices , (19) Wherein: (20) (21) comprises reducing all events of a file, such a request only occurs when the request itself pre-stored cannot be satisfied, the existing file can satisfy; (22) Wherein: (23) indicates ; sub-matrix contains Add an event to a file that occurs when a request cannot be satisfied by a pre-stored file, neither of the two queues contain the file, the user requests it from the surrounding users and the transfer is successful: (24) Wherein: (25) After substituting formulas (11)-(23) into formula (10), the transition probability matrix of the queuing model is obtained. 7.The RIS-aided D2D caching network task offloading modeling method of claim 1, characterized in that, In the step 5, the key performance parameters include the successful offloading probability, the average queue length and the average waiting time of the cache network, and the specific derivation process is: Assume is the transition probability matrix is the column matrix of the steady-state probability distribution, in order to exploit the regular structure of the block matrix , we divide according to the level of the quasi-birth-and-death process (26) where ; given the normalization and the probabilities, the steady-state distribution is obtained by solving the following system of matrix equations: ; given the normalization and the probabilities, the steady-state distribution is obtained by solving the following system of matrix equations: (27) The following definitions of several important performance parameters are given, firstly considering the uncertainty of the request user requesting content, the successful offloading probability is represented as: (28) and The average length of the queue is represented as: (29) (30) Given a steady-state probability distribution, and Average throughput is defined as and The derivation of their respective expressions is as follows: (31) (32) Little's law is used to evaluate the average waiting time required before two queues send packets and is expressed as: (33) (34)。

Images