阅读说明：本技术 面向毫米波mec的低时延-高速率的卸载传输方法 (Millimeter wave MEC-oriented low-delay-high-rate unloading transmission method ) 是由 石嘉 赵钟灵 胡新旸 周奕帆 司江勃 李赞 于 2021-06-22 设计创作，主要内容包括：本发明公开了一种面向毫米波MEC的低时延-高速率的卸载传输方法,该方法为：在固定的用户配对情况下,获得最优波束宽度分配Ω～(*)；对CPUE和CMUE进行随机配对；逐一检测所有CMUE的配对关系是否为最优配对；对于非最优配对的CPUE和CMUE进行交换配对,获得对应的最优配对；根据所述最优波束宽度分配Ω～(*)对所有的最优配对的波束宽度进行迭代优化。本发明对用户配对和波束宽度这两项待优化变量进行分配,以求得在多种用户并存的情况下实现不同种类用户的服务性能折衷,进而提高mmWave MEC系统的传输与计算性能。(The invention discloses a low-delay-high-rate unloading transmission method facing millimeter wave MEC, which comprises the following steps: obtaining optimal beamwidth allocation omega under fixed user pairing * (ii) a Carrying out random pairing on the CPUE and the CMUE; detecting whether the pairing relation of all CMIEs is optimal pairing one by one; exchanging and pairing the CPUE and the CMUE which are not optimally paired to obtain a corresponding optimal pairing; allocating omega according to the optimal beam width * And performing iterative optimization on the beam widths of all the optimal pairs. The invention distributes two variables to be optimized, namely user pairing and beam width, so as to obtain the service performance compromise of different types of users under the condition that multiple users coexist, and further improve the transmission and calculation performance of the mmWave MEC system.)

1. A low-delay-high-rate unloading transmission method facing millimeter wave MEC is characterized in that the method comprises the following steps:

obtaining optimal beamwidth allocation omega under fixed user pairing^*；

Carrying out random pairing on the CPUE and the CMUE;

detecting whether the pairing relation of all CMIEs is optimal pairing one by one;

exchanging and pairing the CPUE and the CMUE which are not optimally paired to obtain a corresponding optimal pairing;

allocating omega according to the optimal beam width^*And performing iterative optimization on the beam widths of all the optimal pairs.

2. The millimeter wave MEC-oriented low-latency high-rate offload transmission method according to claim 1, wherein the optimal beam width allocation Ω is obtained under fixed user pairing conditions^*The method specifically comprises the following steps: initializing auxiliary variables x and y, enabling an iteration counter n to be 0, setting an iteration termination threshold zeta to be more than or equal to 0, and performing interior point alignment by adopting an interior point methodSolving to obtain the current beam distribution result omega^[n]When the average time delay difference of the CMCUs iterated twice is smaller than zeta, iteration is terminated, and the optimal beam width distribution omega is obtained^*。

3. The millimeter wave MEC-oriented low-latency high-rate offload transmission method according to claim 2, wherein the auxiliary variables x and y both satisfy The auxiliary variables x and y replace the signal-to-noise ratio of the CPUE and the signal-to-noise sum of the CMUE, respectively.

4. The millimeter wave MEC-oriented low-latency high-rate offload transmission method according to claim 1 or 2, wherein the randomly pairing the CPUE and the CMUE specifically comprises: randomly pairing the CPUE and the CMUE, and generating an allocation state matrix A if A_i,jAnd if the result is 1, indicating that the CPUE i and the CMUT j form a pairing relation, otherwise, indicating that the CPUE i and the CMUT j do not form a pairing relation, generating an exchange control matrix C with the same size, and making C equal to 1-A.

5. The millimeter wave MEC-oriented low-latency high-rate offload transmission method according to claim 3, wherein the step of detecting whether the pairing relationship of all CMCUs is optimal pairing one by one is specifically as follows: and if 1 element exists in the switching matrix C corresponding to any CMUE j, determining that the CPUE i and the CMUE j are not the optimal pairing, and otherwise, determining that the CPUE i and the CMUE j are the optimal pairing.

6. The millimeter wave MEC-oriented low-latency high-rate offload transmission method according to claim 5, wherein the CPUE and the CMUE that are not optimally paired are exchanged and paired to obtain a corresponding optimal pairing, specifically: when the CPUE i and the CMUT j are in non-optimal matching, the CPUE i is released from the matching relation with the CMUT j in the original user group and is combined with the CMUT j ', meanwhile, the CPUE i ' is released from the matching relation with the CMUT j ' in the original user group, the distribution matrix A is updated, and meanwhile, the position where the control matrix C is changed is reset to be 0.

7. The millimeter wave MEC-oriented low-latency high-rate offload transmission method according to claim 6, wherein the Ω -shaped allocation is performed according to the optimal beam width^*After iterative optimization of all the optimally paired beam widths, the method also includesThe method comprises the following steps: and continuously detecting all CMUTs one by one, if 1 element still exists in the exchange control matrix C corresponding to any CMUT, determining that the pair is the CPUE and the CMUT which are not the optimal pair, and continuously carrying out exchange pairing on the CPUE and the CMUT which are not the optimal pair until the whole exchange control matrix is 0.

Technical Field

The invention belongs to the technical field of communication, and particularly relates to a low-delay-high-rate unloading transmission method for millimeter wave MECs.

Background

With the development of internet of things (IoT) technology, in the future B5G/6G communication system, there will be many real-time application scenarios, and the system will also face larger-scale user access. With the push of low-latency data processing requirements, Mobile Edge Computing (MEC) will become an important technology to improve user experience and reduce network cost. In order to meet the low-delay computing requirement of future MEC systems, the existing key technology of the physical layer, particularly millimeter wave (mmWave) technology, needs to be considered at the same time. The MmWave is a promising technology, and has a huge available bandwidth resource, the MEC offloading transmission capability can be improved manyfold, so that a wide research interest has been attracted in recent years, the technology has become an important research trend of the later 5G era and the 6G era, the combination of the MEC and the MmWave will bring a new opportunity for future low-delay implementation offloading and transmission, and in addition, an efficient radio resource management strategy will provide technical support for the system.

However, the current research on the radio resource management technology of the MEC system still has a big problem. First, most radio resource management work still stays in the microwave phase, mainly aiming at resource management of MEC system in microwave band. However, many measured data show that mmWave can achieve higher transmission rate and result in lower time delay in MEC system, and the combination of these two key technologies will be trending greatly. Therefore, development of MEC system resource management technology oriented to mmWave is urgently needed. Secondly, in practical MEC systems, computation class users (CPUE) and communication class users (CMUE) often coexist. However, the existing resource management work usually ignores this point, and most of the existing work mainly aims at the scenario where only CPUE exists and performs optimal allocation of radio resources on the basis, which is undoubtedly far away from the actual MEC system. In addition, the existing work optimization target is single, most of work is concentrated on single target optimization such as calculation delay and unloading energy consumption, which is not consistent with the complicated MEC unloading process, and a more appropriate resource allocation mode should be selected from a plurality of targets with a trade-off relation.

Disclosure of Invention

In view of the above, the main objective of the present invention is to provide a millimeter wave MEC-oriented low latency-high rate offloading transmission method

In order to achieve the purpose, the technical scheme of the invention is realized as follows:

the embodiment of the invention provides a low-delay-high-rate unloading transmission method for millimeter wave MECs, which comprises the following steps:

obtaining optimal beamwidth allocation omega under fixed user pairing^*；

Carrying out random pairing on the CPUE and the CMUE;

detecting whether the pairing relation of all CMIEs is optimal pairing one by one;

exchanging and pairing the CPUE and the CMUE which are not optimally paired to obtain a corresponding optimal pairing;

allocating omega according to the optimal beam width^*And performing iterative optimization on the beam widths of all the optimal pairs.

In the above scheme, the optimal beam width allocation Ω is obtained under the fixed user pairing condition^*The method specifically comprises the following steps: initializing auxiliary variables x and y, enabling an iteration counter n to be 0, setting an iteration termination threshold zeta to be more than or equal to 0, and performing interior point alignment by adopting an interior point methodSolving to obtain the current beam distribution result omega^[n]When the average time delay difference of the CMCUs iterated twice is smaller than zeta, iteration is terminated, and the optimal beam width distribution omega is obtained^*。

In the above scheme, the auxiliary variables x and y both satisfy The auxiliary variables x and y replace the signal-to-noise ratio of the CPUE and the signal-to-noise sum of the CMUE, respectively.

In the foregoing scheme, the randomly pairing the CPUE and the CMUE specifically includes: randomly pairing the CPUE and the CMUE, and generating an allocation state matrix A if A_i，jIndicates CPUE when 1ⁱAnd CMNE^jForming a pairing relation, otherwise, indicating the CPUEⁱAnd CMNE^jNo pairing is formed, and a switching control matrix C of the same size is generated, and C is made equal to 1-a.

In the above scheme, detecting whether the pairing relationships of all CMUEs are optimal pairings one by one specifically includes: if any one CMNE^jIf 1 element exists in the corresponding switching matrix C, the CPUE is determinedⁱAnd CMNE^jThe CPUE is determined to be non-optimal pairing, otherwise, the CPUE is determinedⁱAnd CMNE^jIs the optimal pairing.

In the foregoing solution, the performing exchange pairing on the CPUE and the CMUE that are not optimally paired to obtain a corresponding optimal pairing specifically includes: the CPUEⁱAnd CMNE^jFor non-optimal pairing, the CPUE is usedⁱReleasing CMNE from original user group^jAnd make it match with CMUT^j'Grouping and combining the CPUE^i'Releasing CMNE from original user group^j'And updating the distribution matrix A, and resetting the position of the control matrix C with change to 0.

In the above scheme, the allocation Ω according to the optimal beam width is performed^*After iteratively optimizing the beamwidths of all the optimal pairs, the method further comprises: and continuously detecting all CMUTs one by one, if 1 element still exists in the exchange control matrix C corresponding to any CMUT, determining that the pair is the CPUE and the CMUT which are not the optimal pair, and continuously carrying out exchange pairing on the CPUE and the CMUT which are not the optimal pair until the whole exchange control matrix is 0.

Compared with the prior art, the method and the device distribute two variables to be optimized, namely user pairing and beam width, so that the service performance compromise of different types of users is realized under the condition that multiple users coexist, and the transmission and calculation performance of the mmWave MEC system is further improved.

Drawings

The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this specification, illustrate embodiment(s) of the invention and together with the description serve to explain the invention without limiting the invention. In the drawings:

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

fig. 2 is a diagram of a terry front surface simulation of the proposed algorithm and two other typical algorithms.

Detailed Description

In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

It should be noted that, in this document, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, article, or apparatus. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other like elements in a process, article, or apparatus that comprises the element.

The invention firstly optimizes the beam width variable under the condition of fixed user pairing. In the process, aiming at the optimization problem of the beam width, the non-convex optimization problem is converted into a convex problem by adopting a continuous convex approximation mode, and iterative optimization is carried out on the convex problem by adopting an inner point method to finally obtain convergence.

For the CPUE and CMUE that coexist, it is assumed that one CPUE and one CMUE in each timeslot perform uplink transmission in a non-orthogonal multiple access (NOMA) manner, and therefore, it is necessary to match the two, and add the above iterative optimization process of beam width to the two. We propose an exchange matching strategy based on one-to-one matching.

The embodiment of the invention provides a low-delay-high-rate unloading transmission method for millimeter wave MECs, which is realized by the following steps:

step 1: obtaining optimal beamwidth allocation omega under fixed user pairing^*；

Specifically, initializing auxiliary variables x and y, making an iteration counter n equal to 0, setting an iteration termination threshold zeta equal to or greater than 0, and performing interior point pairing by adopting an interior point methodSolving to obtain the current beam distribution result omega^[n]When the average time delay difference of the CMCUs iterated twice is smaller than zeta, iteration is terminated, and the optimal beam width distribution omega is obtained^*。

Both the auxiliary variables x and y satisfy The auxiliary variables x and y replace the signal-to-noise ratio of the CPUE and the signal-to-noise sum of the CMUE, respectively.

Step 2: carrying out random pairing on the CPUE and the CMUE;

specifically, the CPUE and CMUE are randomly paired and an assignment state matrix A is generated if A_i,jIndicates CPUE when 1ⁱAnd CMNE^jForming a pairing relation, otherwise, indicating the CPUEⁱAnd CMNE^jNo pairing is formed, and a switching control matrix C of the same size is generated, and C is made equal to 1-a.

And step 3: detecting whether the pairing relation of all CMIEs is optimal pairing one by one;

in particular, if any one CMNE^jIf 1 element exists in the corresponding switching matrix C, the CPUE is determinedⁱAnd CMNE^jThe CPUE is determined to be non-optimal pairing, otherwise, the CPUE is determinedⁱAnd CMNE^jIs the optimal pairing.

And 4, step 4: exchanging and pairing the CPUE and the CMUE which are not optimally paired to obtain a corresponding optimal pairing;

specifically, the CPUEⁱAnd CMNE^jFor non-optimal pairing, the CPUE is usedⁱReleasing CMNE from original user group^jAnd make it match with CMUT^j'Grouping and combining the CPUE^i'Releasing CMNE from original user group^j'And updating the distribution matrix A, and resetting the position of the control matrix C with change to 0.

And 5: allocating omega according to the optimal beam width^*And performing iterative optimization on the beam widths of all the optimal pairs.

The allocation omega according to the optimal beam width^*After iteratively optimizing the beamwidths of all the optimal pairs, the method further comprises: and continuously detecting all CMUTs one by one, if 1 element still exists in the exchange control matrix C corresponding to any CMUT, determining that the pair is the CPUE and the CMUT which are not the optimal pair, and continuously carrying out exchange pairing on the CPUE and the CMUT which are not the optimal pair until the whole exchange control matrix is 0.

The invention optimizes the millimeter wave beam width as an optimization variable for the first time, further improves the system performance compared with the prior art, comprehensively considers the transmission and unloading requirements of two users, namely CPUE and CMUE, performs joint optimization on the beam width of an uplink and the user pairing condition, optimizes the multi-task target aiming at the complex optimization problem, and finally obtains the compromise of double targets as the optimization result.

The use scene of the invention is the mmWave MEC uplink with the coexistence of the CPUE and the CMUE, the base station is assumed to be positioned at the center of the cell, and the two users are randomly distributed in the cell; assuming that there are I CPUEs, index with I; there are J cmuts, indexed by J. At the same time, considerThe number of time slots is the same as that of two users, and the task amount to be unloaded of each CP user is C_i. Defining the wave beam gain of the millimeter wave antenna at the sending end as G, which is inversely proportional to the wave beam width and aims at a specific CPUEⁱThe calculation task unloading delay can be expressed as:

wherein D is₀In order to align the time for the beams,wait duration for user i, R_i，tFor CPUEⁱThe transmission rate at time slot t. Due to the NOMA transmission scheme, the transmission rates of two users can be respectively expressed as:

according to the system description above, the problem is modeled as a multi-objective optimization problem aiming at minimizing the transmission delay of the CPUE and maximizing the transmission rate of the CMUE, and can be modeled as

s.t.A_i，j＝0，1， (6)

R_j，t≥R₀， (8)

w₀≤ω_i，t≤ω₁， (10)

w₀≤ω_j，t≤ω₁。 (11)

The above problem targets the average delay of all CPUEs and the transmission and rate of CMUEs, and optimizes user pairing and beam width allocation. Constraints (6) and (7) are user pairing constraints, which indicate a one-to-one pairing relationship between the CPUE and the CMUE; constraint (8) is the lower limit of the transmission rate of the CMUT; constraint (9) is the energy consumption constraint of the CPUE, and constraints (10) and (11) are the upper and lower limits of the beam width of the two users respectively.

However, the multi-objective optimization problem cannot be solved by a low-complexity algorithm, so the rate of the CMUE in the algorithm is converted into a constraint condition, and the whole problem is modeled as:

compared with the original optimization problem, the problem P2 is converted into constraint (13), the original multi-objective optimization problem is converted into a single-objective optimization problem in a mode of setting and rate function lower limit, and R is adjusted₁The pareto optimal solution of the original multi-objective optimization problem is obtained through the values.

For the above optimization problem, we optimize the beam width Ω and the user pairing relationship a separately, and when considering the optimization of the beam width Ω separately, the problem P3 can be expressed as

As can be seen from the problem model, both the objective function and the constraint (9) in the problem are non-convex functions, so the problem is transformed by adopting a continuous convex approximation processing mode, and the specific operations are as follows:

adding auxiliary variable z, converting the target into constraint, and writing the problem as

To solve the non-convex constraint (9), auxiliary variables x and y are introduced, both satisfying

y_j≥p_j，tG_j，t|h_j，t|²L_j+N₀ (18)

Replacing the signal-to-noise ratio of the CPUE and the signal-to-noise sum of the CMUE with auxiliary variables x and y respectively, wherein the constraint (9) is a convex constraint, and the non-convex constraint introduced in the constraint (17) is changed into

x_iy_j≤p_i，tG_i，t|h_i，t|²L_i (19)

Wherein x_iy_jExist in the upper bound

By combining (28) with (29), a

All non-convex constraints are converted into convex constraints through the continuous convex approximation process. At this point the problem can be written

At this time, the optimization for the beam width is a convex optimization problem.

As shown in fig. 1, the implementation steps of the present invention in the above scenario are as follows:

step one, beam width distribution

And under the condition of fixed user pairing, optimizing the beam width by adopting an iteration inner point method. The specific flow is as follows

1.1 initialize the system, initialize the auxiliary variables x and y, let the iteration counter n equal to 0, and set the iteration stop threshold z³0。

1.2 solving the problem (22) by adopting an interior point method and obtaining a current beam distribution result W^[n]

1.3 when the average time delay difference of CMUT of two iterations is less than z, the iteration is terminated and the optimal beam width distribution w is obtained^*。

Step two, user pairing

For the coexisting CPUEs and cmuts, it is assumed that one CPUE and one CMUE in each timeslot perform uplink transmission in a non-orthogonal multiple access (NOMA) manner. It is therefore necessary to match the two and add the above beam width assignment process to them. We propose an exchange matching strategy based on one-to-one matching, the specific scheme is as follows:

2.1 initialization, carry on the random pairing with CPUE and CMUE, at this moment the system produces a distribution state matrix A, if A_i,jIndicates CPUE when 1ⁱAnd CMNE^jForming a pairing relationship. Otherwise, it indicates that the two are not paired. Setting up a switch of equal size simultaneouslyAnd controlling the matrix C, and enabling C to be 1-A.

2.2 detecting all CMIEs one by one, if some CMIEs are found^jThere is 1 element in the corresponding switch matrix C, performing 2.3.

2.3CPUEⁱAnd CPUE^i'Are respectively connected with CMUT^jAnd CMNE^j'Grouping, if the position of CPUE in two pairs of NOMA user groups is interchanged to realize better system performance, the CPUE is groupedⁱReleasing CMNE from original user group^jAnd make it match with CMUT^j'Grouping and combining the CPUE^i'Releasing CMNE from original user group^j'And updating the distribution matrix A, and resetting the changed position of the control matrix C to be 0.

2.4 under the current user pairing relationship, executing the iterative beam width optimization process in the step one. And the optimization result is used as the input of the subsequent matching algorithm.

And 2.5, continuously detecting the CMUTs one by one, and if 1 element is found to exist in the exchange control matrix C corresponding to a certain CMUT, executing 2.3. And stopping the algorithm until the whole exchange control matrix is a 0 matrix.

The effects of the invention can be further illustrated by simulation:

1. simulation conditions are as follows: in the considered mmWave MEC unloading transmission scene, 1 base station is included, the base station has a limited transmission bandwidth of 2GHz, 8 CPUEs and 8 CMUEs, and the transmission condition in 8 time slots is considered, the side lobe gain is set to 0.001, the pilot transmission time is 1ms, and the sector-level beam width is p/4 rad/s.

2. Simulation content: in fig. 2, the terrorist front surface of the proposed algorithm and two other typical algorithms are simulated. The vertical axis is the average time delay of the CPUE, and the horizontal axis is the sum rate of the CMUEs, and it can be seen from the graph that as the sum rate of the CMUEs increases, the average time delay shows an increasing trend, indicating that two users have a competitive relationship with the communication resources. At the same time, the algorithm proposed by us can produce better performance than the typical algorithm at different computation task amounts. This shows that the proposed algorithm can effectively achieve mmWave MEC offload resource allocation and produce better system performance.

The above description is only a preferred embodiment of the present invention, and is not intended to limit the scope of the present invention.