CN113747557A - A dual-layer heterogeneous network power control method - Google Patents

Abstract

The invention discloses a power control method for a double-layer heterogeneous network. The power control model of the double-layer heterogeneous network is established; the positions of stars and quantum positions are initialized and sorted; new galaxies are selected according to a tournament selection mechanism; Quantum rotation angle, use the optimization search process of simulating quantum revolving gate evolution galaxy; judge if the maximum number of cycles is not reached, return to the previous step until the maximum number of cycles, otherwise terminate the cycle, move the star to positive and negative rotational chaotic movement, and search for more Excellent galaxy; if it is judged that the maximum number of iterations has not been reached, go back to the previous step; otherwise, terminate the loop, mix the newly obtained galaxy with the initial galaxy, and select a galaxy of the same scale as the initial galaxy; if it is judged that the maximum number of iterations has not been reached, return to the championship The selection mechanism selects new galaxies; otherwise, the iteration is terminated, and the optimal power distribution scheme is obtained. The present invention can simultaneously optimize the conflicting system throughput and system energy consumption.

Description

A dual-layer heterogeneous network power control method

technical field

The invention belongs to the field of resource management, and relates to a power control method for a double-layer heterogeneous network, in particular to a double-layer heterogeneous network multi-target power control method based on a multi-target quantum galaxy search mechanism.

Background technique

With the advancement of communication technology, it can be seen that the communication between people has become more and more convenient. From hundreds of years ago, messages could only be transmitted by flying pigeons and beating drums, to later mailing letters or telegrams. To achieve mutual communication, now only a phone call, a text message can achieve the purpose of message delivery, and even through video and other methods to meet the requirements of wanting to meet, these progress in communication are clearly visible to us, and the requirements of users are also Increasingly, users expect high-quality communications wherever they are. Obviously, only wireless communication systems can achieve this ideal.

The development process of wireless mobile communication from 1G to 5G, which we are now familiar with, brings convenience to the naked eye. The first generation of mobile communications, at the beginning of its development, had many shortcomings, such as low spectrum utilization, poor security, and bulky mobile devices that were not conducive to portability. When the second-generation mobile communication develops, many problems existing in the first-generation mobile communication have been solved, the spectrum utilization rate has been greatly improved, the confidentiality has been strengthened, and the services have been enriched a lot. The third generation of mobile communication has more multimedia services, such as e-mail, short messages and so on. The application of the fourth-generation mobile communication is more extensive, not only in daily life such as video calls and online games, but also in smart homes, car navigation, monitoring and security, etc. to improve people's living standards. The first 5th generation mobile communication standard was released in 2018, proving that we have now entered the 5G era. 5G has wider coverage, higher capacity and faster speed. It can be seen that in the process of historical development, the field of communication has brought a lot of convenience to people's lives, and people have been inseparable from wireless communication. Communication services play an important role in all aspects, making outstanding contributions to the progress and development of society and the improvement of people's living standards.

However, at present, the number of communication users is increasing explosively, and the number of base stations built is very small compared to the rapidly growing number of communication users, which seriously affects the quality of user services. According to statistics, the energy consumption of mobile communication accounts for 15% to 20% of the total energy consumption of information and communication technology, and the energy consumption of the communication industry accounts for 3% to 7% of the total energy consumption of all industries. With the increase in the diversity of communication services, the energy consumption of the communication industry also increases rapidly. The carbon dioxide emissions generated by the communication and information industry exceed 5% of the total global emissions. At this time, it is particularly important to seek ways to reduce energy consumption for green development. In order to meet the increasing user demand, the number of high-power base stations needs to be greatly increased, but too many high-power base stations are obviously not in line with the current concept of green and sustainable development worldwide, and the equipment of high-power base stations is relatively complex. , there will also be inflexibility problems, but it will be more convenient to use low-power base stations, and will not generate excessive energy consumption, and heterogeneous networks will therefore be proposed. Mobile communication is now an indispensable part of people's lives, and power control as its core technology has also been studied many times.

Through the retrieval of prior art literature, it was found that Zhang Xin et al. published a "Power Control Algorithm Based on Resource Allocation in Heterogeneous Networks" published in Computer Engineering and Design (2017, 38(6): 1446-1451). Resource allocation mechanism based on user division and power control according to resource allocation strategy. By continuously adjusting the transmit power of the base station, the interference between heterogeneous networks is reduced and the throughput of the system is improved. However, this mechanism only considers the change of system throughput and does not consider the change of system energy consumption. It is better to take the system throughput and system energy consumption as objective functions at the same time, which can better show the advantages of this power control method. Qian Jin et al. published "Resource Allocation Algorithm Design for Heterogeneous Networks Based on Energy Efficiency Optimization" published in Communication Technology (2013(1):79-82), aiming at the interference between two-layer heterogeneous networks, through the multi-objective genetic mechanism The research of this paper proposes a resource allocation mechanism with system energy efficiency as the optimization goal. However, the mechanism is too complicated, and the optimization effect is not obvious, and the multi-target quantum galaxy search mechanism can be used to obtain a more robust power control method suitable for various situations.

The retrieval results of existing literature show that most of the existing multi-objective power control methods transform the system throughput and energy consumption into a single-objective problem under a certain linear weight, but in practical situations, in order to improve the quality of user communication. , it is necessary to increase the system throughput and increase the transmission power of the device according to various communication requirements, but the increase of the transmission power of the device will increase the energy consumption, which is not expected. The energy consumption is used as the objective function for multi-objective optimization. Therefore, it is urgent to solve the high-dimensional optimization problem of power distribution in two-layer heterogeneous networks, and to obtain a feasible power distribution scheme under any energy consumption and throughput requirements.

SUMMARY OF THE INVENTION

In view of the above prior art, the technical problem to be solved by the present invention is to provide a dual-layer heterogeneous network power control method based on a multi-target quantum galaxy search mechanism, which simultaneously optimizes the conflicting system throughput and system energy consumption.

In order to solve the above-mentioned technical problems, a method for controlling power in a dual-layer heterogeneous network of the present invention includes the following steps:

Step 1, establishing a dual-layer heterogeneous network power control model;

Step 2: Initialize the positions and quantum positions of stars, sort all stars by non-dominated solution level, and then sort all stars in each level according to the crowding degree;

Step 3: Carry out the tournament selection mechanism to select a new galaxy;

Step 4, update the quantum rotation angle according to the chaotic change of the position, and use the optimization search process of simulating the quantum revolving gate evolution galaxy;

作为最优星系，

设定旋转混沌移动的最大循环次数为K₂，循环次数标号为k₂，k₂∈[K₁+1,K₁+K₂]。则第g次迭代中的第k₂次循环中第l个星体的位置为

在旋转混沌移动循环中的初始星系为

Step 5, determine whether the maximum number of cycles K ₁ is reached, if not, return to step 4 until the maximum number of cycles is reached, otherwise terminate the cycle and put the first

As the best galaxy,

The maximum number of cycles of the rotational chaotic movement is set as K ₂ , and the number of cycles is labeled as k ₂ , where k ₂ ∈ [K ₁ +1, K ₁ +K ₂ ]. Then the position of the l-th star in the k _- th cycle in the g-th iteration is

The initial galaxy in the rotating chaotic movement cycle is

Step 6: Move all the stars to positive and negative rotational chaotic movements to find better galaxies;

作为最优结果，

Step 7, judge whether the maximum number of cycles K ₁ +K ₂ is reached, if not, set k ₂ =k ₂ +1, and return to step 6; otherwise, terminate the cycle and use the new galaxy obtained in the gth iteration

As the best result,

Step 8: Mix the new galaxies obtained with the initial galaxies, and select galaxies with the same scale as the initial galaxies;

返回到步骤三；否则终止迭代，将第G次迭代中的Pareto最优星体集合输出，根据具体的通信需求，得到最优的功率分配方案。Step 9: Determine whether the maximum number of iterations G is reached, if not, let g=g+1,

Return to step 3; otherwise, terminate the iteration, output the Pareto optimal star set in the Gth iteration, and obtain the optimal power allocation scheme according to the specific communication requirements.

The present invention also includes:

1. In step 1, the power control model of the two-layer heterogeneous network is established as follows:

Constitute a Macrocell/Femtocell dual-layer heterogeneous network: In a macro cell with a radius of R _m , F Femtocells with a radius of R _f are randomly distributed. N _m macro base station user MUEs are randomly distributed in the Macrocell, and N _f home base station user FUEs are randomly distributed in each Femtocell, and then the spectrum resources are divided into Q sub-channels in a shared frequency band, and all MUEs and FUEs are used together. Q sub-channels; the channel allocation method is: first randomly select Q MUEs from the N _m MUEs and distribute them on the Q sub-channels on average, and then randomly assign the N _m -Q macro base station users to the Q sub-channels, where 2Q>N _m >Q, randomly assign FN _f home base station users to Q sub-channels, where Q>FN _f , the interference situation is as follows:

其中，

为Macrocell到其用户i的信道增益，

为Macrocell给干扰用户y所分配的功率，

为Macrocell到其用户i的路径损耗；从Macrocell到其用户i的信道增益进行建模为：

表示在Macrocell与MUE之间的距离；Interference at the same layer of the Macrocell layer: When Macrocell wants to send a signal to user i, if user i and user y of Macrocell occupy the channel at the same time, then user y is an interfering user, and the interference of Macrocell to user i satisfies:

in,

is the channel gain from Macrocell to its user i,

is the power allocated by Macrocell to interfering user y,

is the path loss from the Macrocell to its user i; the channel gain from the Macrocell to its user i is modeled as:

Indicates the distance between Macrocell and MUE;

其中，

为Femtocell到正常用户i的信道增益，

为Femtocell为其用户y所分配的功率，

为Femtocell到正常用户i的路径损耗，Femtocell对MUE的信道增益满足：

Z_c为一个损耗因子；Z^F通过

求得，λ为波长，

为FBS和室内MUE之间的距离，第i个MUE受到的干扰和为

Cross-layer interference of Femtocell to MUE: When Macrocell wants to send a signal to its user i, if the Femtocell within the coverage of Macrocell also allocates the same sub-channel to transmit the signal sent by Macrocell to its user i, then This Femtocell is the interfering base station, and the interference to Macrocell's user i is:

in,

is the channel gain from Femtocell to normal user i,

is the power allocated by the Femtocell to its user y,

is the path loss from Femtocell to normal user i, the channel gain of Femtocell to MUE satisfies:

Z _c is a loss factor; Z ^F passes

Obtained, λ is the wavelength,

is the distance between the FBS and the indoor MUE, and the sum of the interference received by the i-th MUE is

其中，

为Macrocell到正常用户i的信道增益，

为Macrocell为干扰用户y所分配的功率，

为Macrocell到正常用户i的路径损耗，Macrocell到FUE的信道增益与Macrocell到MUE的信道情况相同，所以模型为：

表示在Macrocell与FUE之间的距离，上述公式中提到的路径损耗可以定义为：宏基站会安装三个天线，每个天线负责120度的扇形区域，三个天线负责的面积加在一起就会完整覆盖宏基站的圆形作用区域，当干扰用户和被干扰用户在同一个天线负责的扇形区域内时，基站发送给干扰用户的功率到达被干扰用户时只剩下原来的二分之一，即路径损耗为二分之一；当干扰用户和被干扰用户不在同一个天线负责的扇形区域内时，功率减小为原来的的四分之一，即路径损耗为四分之一。Macrocell's cross-layer interference to FUE: When Femtocell wants to send a signal to its user i, Macrocell also allocates the same sub-channel for transmitting Femtocell to its user i signal to its user y. At this time, Macrocell is the interfering base station. The interference to user i of Femtocell is:

in,

is the channel gain from Macrocell to normal user i,

is the power allocated by Macrocell for interfering user y,

is the path loss from Macrocell to normal user i, the channel gain from Macrocell to FUE is the same as the channel from Macrocell to MUE, so the model is:

Represents the distance between the Macrocell and the FUE. The path loss mentioned in the above formula can be defined as: the macro base station will install three antennas, each antenna is responsible for a 120-degree sector area, and the areas responsible for the three antennas are added together. It will completely cover the circular action area of the macro base station. When the interfering user and the interfered user are in the sector area responsible for the same antenna, the power sent by the base station to the interfering user reaches the interfered user and only half of the original power is left. , that is, the path loss is one-half; when the interfering user and the interfered user are not in the sector area that the same antenna is responsible for, the power is reduced to a quarter of the original, that is, the path loss is a quarter.

Using additive white Gaussian noise as the ambient noise;

N_m是系统中宏基站用户总数量；FN_f是系统中家庭基站用户总数量；

和

分别表示第i个宏基站用户和第y个家庭基站用户的吞吐量，根据香农公式吞吐量通过

和

求出，其中，

和

分别表示第i个宏基站用户和第y个家庭基站用户的信噪比，具体结果可以通过

和

求出，H_i和H_y表示在正常通信时第i个宏基站用户和第y个家庭基站用户相对应的基站对其的信道增益，P_i和P_y分别表示第i个宏基站用户和第y个家庭基站用户对应的基站为其分配的功率；G_i和G_y为正常通信时第i个宏基站用户和第y个家庭基站用户相对应基站对其的路径损耗；n₀表示环境噪声，在上述公式中M_i和F_y分别代表第i个宏基站用户和第y个家庭基站用户的跨层干扰和同层干扰的干扰和；The system throughput in a heterogeneous network is modeled as

N _m is the total number of macro base station users in the system; FN _f is the total number of home base station users in the system;

and

Respectively represent the throughput of the i-th macro base station user and the y-th home base station user. According to Shannon's formula, the throughput passes through

and

find out, where,

and

respectively represent the signal-to-noise ratio of the i-th macro base station user and the y-th home base station user. The specific results can be obtained by

and

To find out, H _i and H _y represent the channel gain of the base station corresponding to the i-th macro base station user and the y-th home base station user during normal communication, and P _i and P _y represent the i-th macro base station user and P y respectively. The power allocated by the base station corresponding to the y-th home base station user; G _i and G _y are the path losses of the base station corresponding to the i-th macro base station user and the y-th home base station user during normal communication; n ₀ represents the environment Noise, in the above formula, M _i and F _y represent the interference sum of the cross-layer interference and the same-layer interference of the i-th macro base station user and the y-th home base station user respectively;

Taking the first objective function system throughput maximization and the second objective function system energy consumption minimization as objective functions at the same time, the energy consumption minimization problem is transformed into a maximization problem, and the mathematical model is established as the formula:

P_max是系统所能发射的最大功率，系统所能发射的最大功率是基站的发射功率和电路损耗的总和，求解公式为

为Macrocell最大的总功率，

为每个Femtocell最大的总功率，

和

分别表示Macrocell层和Femtocell层的电路损耗，P是系统的实际发射功率，求解公式为

P_i ^m和

分别为第i个宏基站用户和第y个家庭基站用户的实际功率，P_max不变的情况下，P越小，第2个目标函数值f₂越优，

和

分别表示宏基站用户和家庭基站用户实际功率可以达到的最小值和最大值。

P _max is the maximum power that the system can transmit, and the maximum power that the system can transmit is the sum of the transmit power of the base station and the circuit loss, and the solution formula is:

is the maximum total power of the Macrocell,

is the maximum total power per Femtocell,

and

respectively represent the circuit loss of the Macrocell layer and the Femtocell layer, P is the actual transmit power of the system, and the solution formula is

P _i ^m and

are the actual powers of the i-th macro base station user and the y-th home base station user, respectively. When P _max remains unchanged, the smaller P is, the better the second objective function value f ₂ is.

and

Represents the minimum and maximum values that can be achieved by the actual power of macro base station users and home base station users, respectively.

2. Step 2: Initialize the positions and quantum positions of the stars, sort all the stars by non-dominated solution level, and then sort all the stars in each level according to the degree of crowding. Specifically:

l＝1,2,…,L，N_m为MUE总数量，F为家庭基站总数量，N_f为每个家庭基站中的FUE总数量，第1代将星体位置初始化为

和

分别表示宏基站用户和家庭基站用户实际功率可以达到的最小值和最大值；设定

为支配星体

的星体个数，初始

中存放星体

所支配的所有星体的标号；将星体

的两个适应度值分别与星系中所有星体的适应度值比较，若是

并且

l＝1,2,…,L，q＝1,2,…,L，且当等号不同时成立时，则代表星体

支配星体

将星体

的编号q存在

中，若是

并且

当等号不同时成立时，则代表星体

被星体

支配，则

若是

代表星体

没有被任何星体支配，则它的非支配解等级为1；接着遍历非支配解等级为1的星体支配的所有星体，若是除了非支配解等级为1的星体没有其他星体支配，则这些星体的非支配解等级为2，按照此规律找出所有星体的非支配解等级；Set the maximum number of iterations as G, the number of iterations labeled as g, g∈[1,G], set the number of stars in the galaxy as L, and the position of the lth star in the gth iteration as

l=1,2,...,L, N _m is the total number of MUEs, F is the total number of home base stations, N _f is the total number of FUEs in each home base station, the first generation initializes the star position as

and

Represent the minimum and maximum values that can be achieved by the actual power of macro base station users and home base station users; set

to rule the stars

number of stars, initial

store stars

the labels of all the stars that govern; the stars

The two fitness values of , respectively, are compared with the fitness values of all the stars in the galaxy, if

and

l=1,2,…,L, q=1,2,…,L, and when the equal signs are different, it represents a star

dominating star

the astral

The number q exists

in, if

and

When the equal signs are not established at the same time, it represents a star

by astral

dominate, then

if

Representing stars

If it is not dominated by any star, its non-dominated solution level is 1; then traverse all the stars dominated by the non-dominated solution level 1 star, if there is no other star dominated except the non-dominated solution level 1 star, then these stars have The non-dominated solution level is 2, and the non-dominated solution level of all stars is found according to this law;

a＝1,2，f_max和f_min表示该非支配解等级中所有星体的最大的适应度值和最小的适应度值，排序后第一个和最后一个星体的拥挤距离设定为无限大；因为有几个适应度，就有几个对应的拥挤距离，所以第l个星体的平均拥挤距离

和

分别表示第l个星体两个适应度值分别对应的拥挤距离，计算公式为

和

Then sort the galaxies from the non-dominated solution level to 1, and determine the number of stars in each non-dominated solution level. If there are multiple stars with the same non-dominated solution level, sort these stars according to the degree of crowding from large to small. Sort again; the method for judging the degree of crowding is: the degree of crowding is expressed by the normalization of the crowding distance, and all the stars are sorted according to the fitness value from small to large, and the crowding distance of the a-th fitness function of the l-th star is equal to The normalization is calculated as

a=1, 2, f _max and f _min represent the maximum fitness value and minimum fitness value of all stars in the non-dominated solution level, and the crowding distance of the first and last stars after sorting is set to be infinite ; Because there are several fitness, there are several corresponding crowding distances, so the average crowding distance of the lth star

and

respectively represent the crowding distance corresponding to the two fitness values of the l-th star, and the calculation formula is:

and

其中

l＝1,2,…,L，j＝1,2,…,N_m+FN_f，

和

分别被定义为

和

第g次迭代中第k₁次循环中第l个星体的位置为

l＝1,2,…,L。Set the maximum number of cycles as K ₁ , the number of cycles is labeled as k ₁ , k ₁ ∈ [ ₁ ,K ₁ ], the quantum position of the lth star in the k1th cycle in the gth iteration can be expressed as

in

l=1,2,...,L, j=1,2,...,N _m +FN _f ,

and

are defined as

and

The position of the l-th star in the k _- th iteration in the g-th iteration is

l=1,2,...,L.

3. In step 3, the tournament selection mechanism is performed, and the new galaxy is selected as follows:

个星体组成新的星系，L为初始星系中星体个数，T为每次进行选择的星体个数，将选出的新的星系中每个星体的序号存入

中，然后根据

找出对应的星体的量子位置。Select T different stars among all the stars in the galaxy, the stars can only be compared once, compare the non-dominated solution levels, select the non-dominated solution level with the smallest level, if the non-dominated solution levels are the same, select the largest crowding distance, if the crowding distance the same, select the first one, and through multiple selections, select the

Each star forms a new galaxy, L is the number of stars in the initial galaxy, T is the number of stars selected each time, and the serial number of each star in the selected new galaxy is stored in

, then according to

Find the quantum position of the corresponding star.

4. Step 4: Update the quantum rotation angle according to the positional chaotic change, and use the simulation quantum revolving gate evolution galaxy to optimize the search process as follows:

产生[0,1]均匀随机数

若

则有第g次迭代中的第k₁+1次循环中第l个星体第j维的量子旋转角所对应的量子旋转门对应为

使用模拟量子旋转门对第g次迭代中的第k₁+1次循环中第l个星体第j维的量子位置进行更新，可以表示为

j＝1,2,…,N_m+FN_f；第g次迭代中的第k₁+1次循环中星系中的第l个星体第j维对应量子旋转角

是第g次迭代中的第k₁+1次循环中第l个星体位置变化指数，在初次循环中第l个星体的位置变化指数设定为

abs()代表对括号中的向量的每一维变量取绝对值操作；

是第g次迭代中的第k₁+1次循环中每个星体第j维的对应的位置变化步长，通过

计算，b_max是最大位置变化步长，

是第g次迭代中的第k₁+1次循环第l个星体的混沌指数，可以通过混沌序列计算，公式如下

j＝1,2,…,N_m+FN_f，初始值

是个[0,1]的随机数；将星体的量子位置映射，得到第g迭代中的第k₁+1次循环中第l个星体对应的位置为

和

分别表示宏基站用户和家庭基站用户实际功率可以达到的最小值和最大值；当

并且

时，第g迭代中的第k₁+1次循环中第l个星体位置变化指数为

否则，

令k₁＝k₁+1，返回步骤四。For the lth star in the _k1th cycle,

Generate [0,1] uniform random numbers

like

Then there is a quantum rotation gate corresponding to the quantum rotation angle of the jth dimension of the lth star in the k1+ ₁ cycle in the gth iteration.

The quantum position of the jth dimension of the lth star in the k1+ ₁ cycle in the gth iteration is updated using the simulated quantum revolving gate, which can be expressed as

j=1,2,...,N _m +FN _f ; the j-th dimension of the l-th star in the galaxy in the k ₁ +1-th cycle in the g-th iteration corresponds to the quantum rotation angle

is the position change index of the lth star in the k1+ ₁ cycle in the gth iteration, and the position change index of the lth star in the first cycle is set as

abs() represents the operation of taking the absolute value of each dimension variable of the vector in parentheses;

is the corresponding position change step size of each star in the jth dimension of the k1+ _1th cycle in the gth iteration, through

Calculate, b _max is the maximum position change step size,

is the chaotic index of the l-th star in the k ₁ +1 cycle in the g-th iteration, which can be calculated by the chaotic sequence, the formula is as follows

j=1,2,...,N _m +FN _f , initial value

is a random number of [0,1]; the quantum position of the star is mapped to obtain the position corresponding to the l-th star in the k ₁ +1 cycle in the g-th iteration as

and

respectively represent the minimum and maximum power that can be achieved by macro base station users and home base station users; when

and

When , the position change index of the l-th star in the k ₁ +1 cycle in the g-th iteration is

otherwise,

Let k ₁ =k ₁ +1, and return to step 4.

则改变第g迭代中的第k₁+1次循环中第l个星体第j维的量子旋转角为

第l个星体第j维所对应的量子旋转门可以表示为

使用模拟量子旋转门，对第k₁+1次循环中第l个星体第j维的量子位置进行更新为

j＝1,2,…,N_m+FN_f；将星体的量子位置映射，得到第g迭代中的第k₁+1次循环中第l个星体第j维对应的位置为：

当

并且

时，第g迭代中的第k₁+1次循环中第l个星体位置变化指数为

否则，

令k₁＝k₁+1，返回步骤四。like

Then change the quantum rotation angle of the jth dimension of the lth star in the k1+ ₁ cycle in the gth iteration to be

The quantum revolving gate corresponding to the jth dimension of the lth star can be expressed as

Using an analog quantum revolving gate, update the quantum position of the jth dimension of the lth star in the k1+ _1th cycle as

j=1,2,...,N _m +FN _f ; map the quantum position of the star to obtain the position corresponding to the j-th dimension of the l-th star in the k ₁ +1 cycle in the g-th iteration:

when

and

When , the position change index of the l-th star in the k ₁ +1 cycle in the g-th iteration is

otherwise,

Let k ₁ =k ₁ +1, and return to step 4.

5. Step 6: Perform positive and negative rotational chaotic movements on all stars to find a better galaxy. Specifically:

产生均匀随机数

若

则星体经过旋转运动，在第g次迭代中的第k₂+1次循环时，将星系中的第l个星体位置第j维更新为

j＝1,2,…,N_m+F×N_f；

是第g次迭代中的第k₂+1次循环第l个星体的混沌指数，它由混沌序列确定，混沌序列公式可以表示为

初始混沌值

是个[0,1]的均匀随机数；

是第g次迭代中的第k₂+1次循环中每个星体第j维对应的位置变化步长，计算公式为

初次循环中的第l个星体的位置变化步长

也由混沌序列产生；第g次迭代中的第k₂+1次循环中第l个星体的旋转角为

初次循环的第l个星体的旋转角

是初次循环中第l个星体的混沌指数；当

时，

加入判断机制，即

和

分别表示宏基站用户和家庭基站用户实际功率可以达到的最小值和最大值；通过贪婪机制保留更新位置前后的较优解

For the l-th star in the k _- th cycle,

generate uniform random numbers

like

Then the star undergoes rotational motion, and in the k ₂ +1 cycle in the g-th iteration, the j-th dimension of the position of the l-th star in the galaxy is updated as

j=1,2,...,N _m +F×N _f ;

is the chaotic index of the lth star in the k2 ₊ 1 cycle in the gth iteration, which is determined by the chaotic sequence, and the chaotic sequence formula can be expressed as

initial chaos value

is a uniform random number of [0,1];

is the position change step size corresponding to the jth dimension of each star in the k2 ₊ 1 cycle in the gth iteration, and the calculation formula is

The position change step size of the lth star in the first cycle

is also produced by the chaotic sequence; the rotation angle of the lth star in the k2 ₊ 1 cycle in the gth iteration is

The rotation angle of the l-th star in the first cycle

is the chaos index of the l-th star in the first cycle; when

hour,

Join the judgment mechanism, that is

and

Represent the minimum and maximum actual power that can be achieved by macro base station users and home base station users, respectively; the better solution before and after updating the position is reserved by the greedy mechanism

再次将第g次迭代中的第k₂+1次循环中第l个星体进行负向旋转，更新后的第j维位置可以表示为

j＝1,2,…,N_m+FN_f；同样地，加入判断机制，即

和

分别表示宏基站用户和家庭基站用户实际功率可以达到的最小值和最大值；再次选出更新位置前后的较优解

like

Negatively rotate the lth star in the k2 ₊ 1 cycle in the gth iteration again, the updated jth dimension position can be expressed as

j=1,2,...,N _m +FN _f ; similarly, a judgment mechanism is added, namely

and

Represent the minimum and maximum values that can be achieved by the actual power of macro base station users and home base station users; select the optimal solution before and after updating the location again.

6. Step 8: Mix the new galaxies obtained with the initial galaxies, and select galaxies of the same scale as the initial galaxies as follows:

与演化前得到的星系

进行混合构成临时星系，再次根据步骤二方式进行非支配解和拥挤度进行排序；在这个临时星系中，先从非支配解等级小的的进行选取，直到选取出L个星体；若是加上该非支配解等级中的所有星体，总星体数目没有超过L，则将该非支配解等级中的所有星体全部选择，若是超出了L，则将该非支配解等级的星体按照拥挤度排序，优先选择拥挤距离较大的，直到达到L个星体，得到新的星系

new galaxy

with pre-evolutionary galaxies

Mix and form temporary galaxies, and then sort the non-dominated solutions and crowding degree according to the second method; in this temporary galaxies, first select from the non-dominated solutions with the lowest level, until L stars are selected; if adding this For all the stars in the non-dominated solution level, if the total number of stars does not exceed L, all the stars in the non-dominated solution level are selected. If it exceeds L, the stars in the non-dominated solution level are sorted according to the degree of crowding, with priority. Choose the one with a larger crowding distance until you reach L stars and get new galaxies

Beneficial effects of the present invention: The present invention aims at the same-layer and cross-layer interference problems that occur in the Macrocell/Femtocell double-layer heterogeneous network. The present invention designs an intelligent power control method, which combines the conflicting system throughput and system energy consumption. As the optimization goal, the multi-objective quantum galaxy search mechanism is used to solve the high-dimensional problem of continuous vector optimization of power distribution, so that the system throughput and energy consumption are optimized at the same time, and better optimization results are obtained than other multi-objective mechanisms, reflecting the superiority of this mechanism.

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

(1) The present invention proposes a Macrocell/Femtocell double-layer heterogeneous network in view of the current rapidly growing communication user group scale and users' high requirements for communication quality and speed, and arranges low-power base station nodes within the coverage of traditional macro base stations, The low-power base station equipment is simple and small in size, convenient and flexible, and does not generate excessive energy consumption, which is in line with the green development concept of energy saving and emission reduction.

(2) The present invention is based on the consideration of the actual situation. In practical engineering applications, in order to improve the communication quality of users, it is necessary to increase the system throughput and increase the transmission power of the equipment, but the improvement of the transmission power of the equipment will increase the energy consumption, which It is not expected to be seen, so the two conflicting problems of system throughput and system energy consumption are taken as the optimization goals at the same time, and the multi-objective optimization problem is solved.

(3) To solve the multi-objective problem, the present invention proposes a multi-objective quantum galaxy search mechanism, which realizes the simultaneous optimization of system throughput and system energy consumption. The superiority of the designed multi-target quantum galaxy search mechanism.

Description of drawings

1 is a schematic diagram of the research on the power control method of the double-layer heterogeneous network based on the multi-target quantum galaxy search mechanism designed by the present invention;

Fig. 2 is a plane structure diagram of a two-layer heterogeneous network;

Fig. 3 is the MOQGBSA, NSGA-II and MOPSO method system throughput and relative energy consumption relation diagram of the present invention;

Fig. 4 is the relation diagram of system throughput and relative energy consumption of MOQGBSA, NSGA-II and MOPSO method of the present invention under the condition that the number of channels is 35;

Fig. 5 is the relation diagram of system throughput and relative energy consumption of MOQGBSA, NSGA-II and MOPSO method of the present invention under the condition that the number of channels is 45;

Figure 6 shows the optimization results obtained when the upper limit of the actual power of the FUE is 50mW, 100mW and 150mW, and the upper limit of the actual power of the MUE is 100mW, 200mW and 300mW respectively.

Detailed ways

The present invention will be further described below with reference to the accompanying drawings and specific embodiments.

1 and 2, the specific embodiment of the present invention includes the following steps:

Step 1, establish a dual-layer heterogeneous network power control model.

In a macro cell with a radius of R _m , F Femtocells with a radius of R _f are randomly distributed, and the coverage of the Femtocells are all within the macro cell and do not intersect each other, thus forming a Macrocell/Femtocell double-layer heterogeneous network . Among them, N _m macro base station users (MUE) are randomly distributed in the entire Macrocell, and N _f home base station users (FUE) are randomly distributed in each Femtocell. Then, the spectrum resource is divided into Q sub-channels in a way of sharing frequency bands, and all the MUEs and FUEs share the Q sub-channels. The channel allocation method is as follows: first randomly select Q MUEs from N _m MUEs and distribute them on Q sub-channels on average, and then randomly assign N _m -Q (2Q>N _m >Q) macro base station users to sub-channels, and FN _f (Q>FN _f ) home base station users are randomly allocated to sub-channels. When different users occupy the same channel, interference will occur, and because the present invention sets parameters, the total number of home base station users will not exceed the number of channels, so there will be no same-layer interference at the Femtocell layer. The following interferences for different situations can be expressed as:

其中，

为Macrocell到其用户i的信道增益，

为Macrocell给干扰用户y所分配的功率，

为Macrocell到其用户i的路径损耗。从Macrocell到其用户i的信道增益，可对其进行建模为：

表示在Macrocell与MUE之间的距离。Interference on the same layer of the Macrocell layer: When Macrocell wants to send a signal to user i, if user i and user y of Macrocell occupy the channel at the same time, then user y is an interfering user, and the interference of macrocell to user i can be calculated according to the formula:

in,

is the channel gain from Macrocell to its user i,

is the power allocated by Macrocell to interfering user y,

is the path loss from the Macrocell to its user i. The channel gain from the Macrocell to its user i can be modeled as:

Indicates the distance between the Macrocell and the MUE.

其中，

为Femtocell到正常用户i的信道增益，

为Femtocell为其用户y所分配的功率，

为Femtocell到正常用户i的路径损耗。Femtocell对MUE的信道增益通过下式计算：

Z_c为一个损耗因子；Z^F通过

求得，λ为波长，

为FBS和室内MUE之间的距离。所以第i个MUE受到的干扰和为

Cross-layer interference of Femtocell to MUE: When Macrocell wants to send a signal to its user i, if the Femtocell within the coverage of Macrocell also allocates the same sub-channel to transmit the signal sent by Macrocell to its user i, then This Femtocell is the interfering base station, and the interference to Macrocell's user i can be calculated as:

in,

is the channel gain from Femtocell to normal user i,

is the power allocated by the Femtocell to its user y,

is the path loss from Femtocell to normal user i. The channel gain of Femtocell to MUE is calculated by the following formula:

Z _c is a loss factor; Z ^F passes

Obtained, λ is the wavelength,

is the distance between the FBS and the indoor MUE. So the interference sum of the i-th MUE is

其中，

为Macrocell到正常用户i的信道增益，

为Macrocell为干扰用户y所分配的功率，

为Macrocell到正常用户i的路径损耗。Macrocell到FUE的信道增益与Macrocell到MUE的信道情况相同，所以模型为：

表示在Macrocell与FUE之间的距离。上述公式中提到的路径损耗可以定义为：宏基站会安装三个天线，每个天线负责120度的扇形区域，三个天线负责的面积加在一起就会完整覆盖宏基站的圆形作用区域。当干扰用户和被干扰用户在同一个天线负责的扇形区域内时，基站发送给干扰用户的功率到达被干扰用户时只剩下原来的二分之一，即路径损耗为二分之一；当干扰用户和被干扰用户不在同一个天线负责的扇形区域内时，功率减小为原来的的四分之一，即路径损耗为四分之一。Macrocell's cross-layer interference to FUE: When Femtocell wants to send a signal to its user i, Macrocell also allocates the same sub-channel for transmitting Femtocell to its user i signal to its user y. At this time, Macrocell is the interfering base station. The interference to user i of Femtocell can be calculated by the following formula:

in,

is the channel gain from Macrocell to normal user i,

is the power allocated by Macrocell for interfering user y,

is the path loss from Macrocell to normal user i. The channel gain from Macrocell to FUE is the same as the channel from Macrocell to MUE, so the model is:

Indicates the distance between the Macrocell and the FUE. The path loss mentioned in the above formula can be defined as: the macro base station will install three antennas, each antenna is responsible for a 120-degree sector area, and the areas responsible for the three antennas are added together to completely cover the circular active area of the macro base station. . When the interfering user and the interfered user are in the sector area that the same antenna is responsible for, the power sent by the base station to the interfering user reaches the interfered user with only one half of the original power, that is, the path loss is one half; when When the interfering user and the interfered user are not in the sector area that the same antenna is responsible for, the power is reduced to a quarter of the original, that is, the path loss is a quarter.

There are plenty of other noises in the actual environment. White Gaussian noise is often used as noise used in communication system analysis, and additive white Gaussian noise is also used as environmental noise in this patent.

N_m是系统中宏基站用户总数量；FN_f是系统中家庭基站用户总数量；

和

分别表示第i个宏基站用户和第y个家庭基站用户的吞吐量，根据香农公式可以知道吞吐量可以通过

和

求出，其中，

和

分别表示第i个宏基站用户和第y个家庭基站用户的信噪比，具体结果可以通过

和

求出，H_i和H_y表示在正常通信时第i个宏基站用户和第y个家庭基站用户相对应的基站对其的信道增益，P_i和P_y分别表示第i个宏基站用户和第y个家庭基站用户对应的基站为其分配的功率；G_i和G_y为正常通信时第i个宏基站用户和第y个家庭基站用户相对应基站对其的路径损耗；n₀表示环境噪声。在上述公式中M_i和F_y分别代表第i个宏基站用户和第y个家庭基站用户的跨层干扰和同层干扰的干扰和。In a heterogeneous network, the system throughput can be modeled as

N _m is the total number of macro base station users in the system; FN _f is the total number of home base station users in the system;

and

respectively represent the throughput of the i-th macro base station user and the y-th home base station user. According to Shannon's formula, it can be known that the throughput can be obtained by

and

find out, where,

and

respectively represent the signal-to-noise ratio of the i-th macro base station user and the y-th home base station user. The specific results can be obtained by

and

To find out, H _i and H _y represent the channel gain of the base station corresponding to the i-th macro base station user and the y-th home base station user during normal communication, and P _i and P _y represent the i-th macro base station user and P y respectively. The power allocated by the base station corresponding to the y-th home base station user; G _i and G _y are the path losses of the base station corresponding to the i-th macro base station user and the y-th home base station user during normal communication; n ₀ represents the environment noise. In the above formula, M _i and F _y represent the interference sum of the cross-layer interference and the same-layer interference of the ith macro base station user and the y th home base station user, respectively.

P_max是系统所能发射的最大功率，系统所能发射的最大功率是基站的发射功率和电路损耗的总和，求解公式为

为Macrocell最大的总功率，

为每个Femtocell最大的总功率，

和

分别表示Macrocell层和Femtocell层的电路损耗。P是系统的实际发射功率，求解公式为

P_i ^m和

分别为第i个宏基站用户和第y个家庭基站用户的实际功率。P_max不变的情况下，P越小，第2个目标函数值f₂越优，

和

分别表示宏基站用户和家庭基站用户实际功率可以达到的最小值和最大值。The first objective function system throughput maximization and the second objective function system energy consumption minimization are both used as objective functions. In order to facilitate analysis, the energy consumption minimization problem is transformed into a maximization problem, so the mathematical model is established as the formula:

P _max is the maximum power that the system can transmit, and the maximum power that the system can transmit is the sum of the transmit power of the base station and the circuit loss, and the solution formula is:

is the maximum total power of the Macrocell,

is the maximum total power per Femtocell,

and

represent the circuit losses of the Macrocell layer and Femtocell layer, respectively. P is the actual transmit power of the system, and the solution formula is

P _i ^m and

are the actual powers of the i-th macro base station user and the y-th home base station user, respectively. When P _max is unchanged, the smaller P is, the better the second objective function value f ₂ is.

and

Represents the minimum and maximum values that can be achieved by the actual power of macro base station users and home base station users, respectively.

Step 2: Initialize the positions and quantum positions of the stars, sort all the stars by non-dominated solution level, and then sort all the stars in each level according to the crowding degree. The specific method is as follows:

l＝1,2,…,L，N_m为MUE总数量，F为家庭基站总数量，N_f为每个家庭基站中的FUE总数量，第1代将星体位置初始化为

和

分别表示宏基站用户和家庭基站用户实际功率可以达到的最小值和最大值。设定

为支配星体

的星体个数，初始

中存放星体

所支配的所有星体的标号。将星体

的两个适应度值分别与星系中所有星体的适应度值比较，若是

并且

l＝1,2,…,L，q＝1,2,…,L，且当等号不同时成立时，则代表星体

支配星体

将星体

的编号q存在

中，若是

并且

当等号不同时成立时，则代表星体

被星体

支配，则

若是

代表星体

没有被任何星体支配，则它的非支配解等级为1。接着遍历非支配解等级为1的星体支配的所有星体，若是除了非支配解等级为1的星体没有其他星体支配，则这些星体的非支配解等级为2。按照此规律找出所有星体的非支配解等级。First, set the maximum number of iterations as G, and the number of iterations as g, g∈[1,G]. Let the number of stars in the galaxy be L, and the position of the l-th star in the g-th iteration is

l=1,2,...,L, N _m is the total number of MUEs, F is the total number of home base stations, N _f is the total number of FUEs in each home base station, the first generation initializes the star position as

and

Represents the minimum and maximum values that can be achieved by the actual power of macro base station users and home base station users, respectively. set up

to rule the stars

number of stars, initial

store stars

The labels of all the stars it rules. the astral

The two fitness values of , respectively, are compared with the fitness values of all the stars in the galaxy, if

and

l=1,2,…,L, q=1,2,…,L, and when the equal signs are different, it represents a star

dominating star

the astral

The number q exists

in, if

and

When the equal signs are not established at the same time, it represents a star

by astral

dominate, then

if

Representing stars

If it is not dominated by any star, its non-dominated solution level is 1. Then traverse all the stars dominated by the star whose non-dominated solution level is 1. If no other star dominates except the star whose non-dominated solution level is 1, the non-dominated solution level of these stars is 2. According to this law, find out the non-dominated solution levels of all stars.

a＝1,2，f_max和f_min表示该非支配解等级中所有星体的最大的适应度值和最小的适应度值，排序后第一个和最后一个星体的拥挤距离设定为无限大。因为有几个适应度，就有几个对应的拥挤距离，所以第l个星体的平均拥挤距离

和

分别表示第l个星体两个适应度值分别对应的拥挤距离，计算公式为

和

Then sort the galaxies from the non-dominated solution level to 1, and determine the number of stars in each non-dominated solution level. If there are multiple stars with the same non-dominated solution level, sort these stars according to the degree of crowding from large to small. Sort again. The method of judging the degree of crowding is: the degree of crowding is expressed by the normalization of the crowding distance, and all the stars are sorted according to the fitness value from small to large, and the normalization of the crowding distance of the a-th fitness function of the l-th star is calculated. Calculated as

a=1, 2, f _max and f _min represent the maximum fitness value and minimum fitness value of all stars in the non-dominated solution level, and the crowding distance of the first and last stars after sorting is set to be infinite . Because there are several fitness, there are several corresponding crowding distances, so the average crowding distance of the lth star

and

respectively represent the crowding distance corresponding to the two fitness values of the l-th star, and the calculation formula is:

and

其中

l＝1,2,…,L，j＝1,2,…,N_m+FN_f，

和

分别被定义为

和

第g次迭代中第k₁次循环中第l个星体的位置为

The maximum number of cycles is set as K ₁ , the number of cycles is labeled as k ₁ , and k ₁ ∈ [1, K ₁ ]. The quantum position of the l-th star in the k _- th cycle in the g-th iteration can be expressed as

in

l=1,2,...,L, j=1,2,...,N _m +FN _f ,

and

are defined as

and

The position of the l-th star in the k _- th iteration in the g-th iteration is

Step 3: Perform the tournament selection mechanism to select a new galaxy. The specific steps are:

个星体组成新的星系，L为初始星系中星体个数，T为每次进行选择的星体个数，将选出的新的星系中每个星体的序号存入

中，然后根据

找出对应的星体的量子位置。The selection is made through the tournament selection mechanism. The specific method is: select T different stars among all the stars in the galaxy. The stars can only be compared once, and the non-dominated solution levels are compared, and the non-dominated solution level is selected. The smallest, if the non-dominated solution levels are the same , select the one with the largest crowding distance, if the crowding distance is the same, select the first one, and select the

Each star forms a new galaxy, L is the number of stars in the initial galaxy, T is the number of stars selected each time, and the serial number of each star in the selected new galaxy is stored in

, then according to

Find the quantum position of the corresponding star.

Step 4: Update the quantum rotation angle according to the positional chaotic change, and use the optimization search process to simulate the evolution of the quantum revolving gate. The specific steps are:

产生[0,1]均匀随机数

若

则有第g次迭代中的第k₁+1次循环中第l个星体第j维的量子旋转角所对应的量子旋转门对应为

使用模拟量子旋转门对第g次迭代中的第k₁+1次循环中第l个星体第j维的量子位置进行更新，可以表示为

j＝1,2,…,N_m+FN_f。第g次迭代中的第k₁+1次循环中星系中的第l个星体第j维对应量子旋转角

是第g次迭代中的第k₁+1次循环中第l个星体位置变化指数，在初次循环中第l个星体的位置变化指数设定为

abs()代表对括号中的向量的每一维变量取绝对值操作。

是第g次迭代中的第k₁+1次循环中每个星体第j维的对应的位置变化步长，可以通过

计算，b_max是最大位置变化步长，

是第g次迭代中的第k₁+1次循环第l个星体的混沌指数，可以通过混沌序列计算，公式如下

j＝1,2,…,N_m+FN_f，初始值

是个[0,1]的随机数。将星体的量子位置映射，得到第g迭代中的第k₁+1次循环中第l个星体对应的位置为

和

分别表示宏基站用户和家庭基站用户实际功率可以达到的最小值和最大值。当

并且

时，第g迭代中的第k₁+1次循环中第l个星体位置变化指数为

否则，

令k₁＝k₁+1，返回步骤四。For the lth star in the _k1th cycle,

Generate [0,1] uniform random numbers

like

Then there is a quantum rotation gate corresponding to the quantum rotation angle of the jth dimension of the lth star in the k1+ ₁ cycle in the gth iteration.

The quantum position of the jth dimension of the lth star in the k1+ ₁ cycle in the gth iteration is updated using the simulated quantum revolving gate, which can be expressed as

j=1,2,...,N _m +FN _f . Quantum rotation angle corresponding to the jth dimension of the lth star in the galaxy in the k1+ ₁ cycle in the gth iteration

is the position change index of the lth star in the k1+ ₁ cycle in the gth iteration, and the position change index of the lth star in the first cycle is set as

abs() represents the operation of taking the absolute value of each dimension of the vector in parentheses.

is the corresponding position change step size of each star in the jth dimension of the k ₁ +1 cycle in the gth iteration, which can be obtained by

Calculate, b _max is the maximum position change step size,

is the chaotic index of the l-th star in the k ₁ +1 cycle in the g-th iteration, which can be calculated by the chaotic sequence, the formula is as follows

j=1,2,...,N _m +FN _f , initial value

is a random number in [0,1]. Map the quantum position of the star to obtain the position corresponding to the lth star in the k ₁ +1 cycle in the gth iteration as

and

Represents the minimum and maximum values that can be achieved by the actual power of macro base station users and home base station users, respectively. when

and

When , the position change index of the l-th star in the k ₁ +1 cycle in the g-th iteration is

otherwise,

Let k ₁ =k ₁ +1, and return to step 4.

则改变第g迭代中的第k₁+1次循环中第l个星体第j维的量子旋转角为

第l个星体第j维所对应的量子旋转门可以表示为

使用模拟量子旋转门，对第k₁+1次循环中第l个星体第j维的量子位置进行更新为

j＝1,2,…,N_m+FN_f。将星体的量子位置映射，得到第g迭代中的第k₁+1次循环中第l个星体第j维对应的位置为

当

并且

时，第g迭代中的第k₁+1次循环中第l个星体位置变化指数为

否则，

令k₁＝k₁+1，返回步骤四。like

Then change the quantum rotation angle of the jth dimension of the lth star in the k1+ ₁ cycle in the gth iteration to be

The quantum revolving gate corresponding to the jth dimension of the lth star can be expressed as

Using an analog quantum revolving gate, update the quantum position of the jth dimension of the lth star in the k1+ _1th cycle as

j=1,2,...,N _m +FN _f . Map the quantum position of the star to obtain the position corresponding to the jth dimension of the lth star in the k1+ ₁ cycle in the gth iteration as

when

and

When , the position change index of the l-th star in the k ₁ +1 cycle in the g-th iteration is

otherwise,

Let k ₁ =k ₁ +1, and return to step 4.

作为最优星系，

设定旋转混沌移动的最大循环次数为K₂，循环次数标号为k₂，k₂∈[K₁+1,K₁+K₂]。则第g次迭代中的第k₂次循环中第l个星体的位置为

在旋转混沌移动循环中的初始星系为

Step 5, determine whether the maximum number of cycles K ₁ is reached, if not, return to step 4 until the maximum number of cycles is reached, otherwise terminate the cycle and put the first

As the best galaxy,

The maximum number of cycles of the rotational chaotic movement is set as K ₂ , and the number of cycles is labeled as k ₂ , where k ₂ ∈ [K ₁ +1, K ₁ +K ₂ ]. Then the position of the l-th star in the k _- th cycle in the g-th iteration is

The initial galaxy in the rotating chaotic movement cycle is

Step 6: Move all the stars to positive and negative rotational chaotic movements to find better galaxies. The specific steps are:

产生均匀随机数

若

则星体经过旋转运动，在第g次迭代中的第k₂+1次循环时，将星系中的第l个星体位置第j维更新为

j＝1,2,…,N_m+F×N_f。

是第g次迭代中的第k₂+1次循环第l个星体的混沌指数，它由混沌序列确定，混沌序列公式可以表示为

初始混沌值

是个[0,1]的均匀随机数。

是第g次迭代中的第k₂+1次循环中每个星体第j维对应的位置变化步长，计算公式为

初次循环中的第l个星体的位置变化步长

也由混沌序列产生。第g次迭代中的第k₂+1次循环中第l个星体的旋转角为

初次循环的第l个星体的旋转角

是初次循环中第l个星体的混沌指数。当

时，

并且因为用户实际功率有限制，所以需要加入判断机制，即

和

分别表示宏基站用户和家庭基站用户实际功率可以达到的最小值和最大值。通过贪婪机制保留更新位置前后的较优解

For the l-th star in the k _- th cycle,

generate uniform random numbers

like

Then the star undergoes rotational motion, and in the k ₂ +1 cycle in the g-th iteration, the j-th dimension of the position of the l-th star in the galaxy is updated as

j=1,2,...,N _m +F×N _f .

is the chaotic index of the lth star in the k2 ₊ 1 cycle in the gth iteration, which is determined by the chaotic sequence, and the chaotic sequence formula can be expressed as

initial chaos value

is a uniform random number of [0,1].

is the position change step size corresponding to the jth dimension of each star in the k2 ₊ 1 cycle in the gth iteration, and the calculation formula is

The position change step size of the lth star in the first cycle

Also generated by chaotic sequences. The rotation angle of the lth star in the k2 ₊ 1 cycle in the gth iteration is

The rotation angle of the l-th star in the first cycle

is the chaos index of the lth star in the first cycle. when

hour,

And because the actual power of the user is limited, a judgment mechanism needs to be added, that is,

and

Represents the minimum and maximum values that can be achieved by the actual power of macro base station users and home base station users, respectively. Preserve the optimal solution before and after updating the position through a greedy mechanism

再次将第g次迭代中的第k₂+1次循环中第l个星体进行负向旋转，更新后的第j维位置可以表示为

j＝1,2,…,N_m+FN_f。同样地，需要加入判断机制，即

和

分别表示宏基站用户和家庭基站用户实际功率可以达到的最小值和最大值。再次选出更新位置前后的较优解

like

Negatively rotate the lth star in the k2 ₊ 1 cycle in the gth iteration again, the updated jth dimension position can be expressed as

j=1,2,...,N _m +FN _f . Similarly, a judgment mechanism needs to be added, that is,

and

Represents the minimum and maximum values that can be achieved by the actual power of macro base station users and home base station users, respectively. Select the better solution before and after updating the position again

作为最优结果，

Step 7, judge whether the maximum number of cycles K ₁ +K ₂ is reached, if not, set k ₂ =k ₂ +1, and return to step 6; otherwise, terminate the cycle and use the new galaxy obtained in the gth iteration

As the best result,

Step 8: Mix the new galaxies obtained with the initial galaxies, and select galaxies with the same scale as the initial galaxies. The specific steps are:

与演化前得到的星系

进行混合构成临时星系，再次根据步骤二方式进行非支配解和拥挤度进行排序。在这个临时星系中，先从非支配解等级小的的进行选取，直到选取出L个星体。若是加上该非支配解等级中的所有星体，总星体数目没有超过L，则将该非支配解等级中的所有星体全部选择，若是超出了L，则将该非支配解等级的星体按照拥挤度排序，优先选择拥挤距离较大的，直到达到L个星体，得到新的星系

new galaxy

with pre-evolutionary galaxies

Mixing is performed to form temporary galaxies, and the non-dominated solution and crowding degree are sorted again according to the second method. In this temporary galaxy, first select from the non-dominated solution with a smaller level, until L stars are selected. If all the stars in the non-dominated solution level are added, and the total number of stars does not exceed L, then all the stars in the non-dominated solution level are selected. Sort by degree, give preference to those with larger crowding distances, until L stars are reached, and new galaxies are obtained

返回到步骤三；否则终止迭代，将第G次迭代中的Pareto最优星体集合输出，根据具体的通信需求，得到最优的功率分配方案。Step 9: Determine whether the maximum number of iterations G is reached, if not, let g=g+1,

Return to step 3; otherwise, terminate the iteration, output the Pareto optimal star set in the Gth iteration, and obtain the optimal power allocation scheme according to the specific communication requirements.

In Fig. 3, Fig. 4 and Fig. 5, the power control method of the double-layer heterogeneous network based on the multi-target quantum galaxy search mechanism proposed by the present invention is denoted as MOQGBSA; the power control method of the double-layer heterogeneous network based on the multi-target genetic mechanism The method is denoted as NSGA-II; the power control method of double-layer heterogeneous network based on multi-objective particle swarm mechanism is denoted as MOPSO. The parameter selection of NSGA-II is based on "Multi-objective power distribution optimization using NSGA-II" published in International Journal for Computational Methods in Engineering Science and Mechanic (2021, 22(3): 235-243) by Jain Kunal, Gupta Shashank and Kumar Divya ". The parameter selection of MOPSO is based on "Multi-Objective Particle Swarm Optimization Based on Multi-population Dynamic Cooperation" published by Yu H, Wang Y.J. and Chen Q in Electronic Science and Technology (2019, 32(10): 28-33). The rest of the parameter selections are consistent with MOQGBSA.

Figures 4 and 5 are optimized results obtained with the number of channels Q being 35 and 45, respectively

Fig. 6 The upper limit of the actual power of the FUE is 100mW, 200mW and 300mW, and the upper limit of the actual power of the MUE is 50mW, 100mW and 150mW.

T＝2，K₁＝100，K₂＝600，G＝1000，

b_max＝2，Z_c＝-30dB，a₁＝1，a₂＝4，a₃＝0.01，a₄＝0.05，a₅＝0.01，a₆＝0.001，a₇＝0.01，a₈＝-1，a₉＝2。The simulation parameters are set as follows: N _m =50, N _f =5, F = 5, R _m =500 m, R _f =70 m, L = 80, Q = 40,

T= ₂ , K1 ₌ 100, K2=600, G=1000,

b _max =2, Z _c =-30dB, a ₁ =1,a ₂ =4,a ₃ =0.01,a ₄ =0.05,a ₅ =0.01,a ₆ =0.001,a ₇ =0.01,a ₈ =- 1, a ₉ =2.

It can be seen from the simulation diagram FIG. 3 that the power control method of the MOQGbSA-based dual-layer heterogeneous network designed by the present invention optimizes the system throughput and system energy consumption at the same time, and obtains better performance than NSGA-II and MOPSO. The optimization results of MOQGbSA are better than NSGA-II and MOPSO in terms of system throughput and system energy consumption. Comparing the simulation results in Fig. 3, Fig. 4 and Fig. 5, when the channel is changed, the optimization results of MOQGbSA are also stronger than that of NSGA-II and MOPSO, which verifies the superiority of the algorithm. It can be seen from the simulation Figure 6 that when the actual power upper limits of FUE and MUE are both low, the actual power allocated by FUE and MUE is generally low, the energy consumption of the system is relatively small, and the relative energy consumption is relatively large, but the throughput The amount will decrease as the actual power allocated to the FUE and MUE is low. On the contrary, when the actual power upper limit of FUE and MUE is high, the power actually allocated by FUE and MUE is generally high, and the throughput will increase accordingly, but the energy consumption of the system will decrease accordingly, and the relative energy consumption will increase relatively. .

Claims (7)

1. A power control method for a double-layer heterogeneous network is characterized by comprising the following steps:

step one, establishing a double-layer heterogeneous network power control model;

initializing the positions and quantum positions of the stars, sorting all the stars in the non-dominated solution levels, and sorting all the stars in each level according to the crowding degree;

step three, a championship selection mechanism is carried out to select a new galaxy;

updating the quantum rotation angle according to the position chaotic change, and using an optimization search process of simulating the quantum rotation gate evolution galaxy;

step five, judging whether the maximum cycle number K is reached₁If not, returning to the step four until the maximum circulation times are reached, otherwise, terminating the circulation, and returning to the step four

As the optimal galaxy, the star system,

setting the maximum cycle number of the rotary chaotic movement to be K₂Number of cycles is denoted by k₂，k₂∈[K₁+1,K₁+K₂]. The kth iteration in the g-th iteration₂The position of the first star in the subcycle is

Initial star system in a rotating chaotic moving cycle

Step six, performing positive and negative rotation chaotic movement on all stars to search a more optimal galaxy;

step seven, judging whether the maximum cycle number K is reached₁+K₂If not, let k₂＝k₂+1, returning to the step six; otherwise, the loop is terminated, and the new galaxy obtained in the g iteration is used

As a result of the optimum result,

step eight, mixing the obtained new galaxies with the initial galaxies, and selecting galaxies with the same scale as the initial galaxies;

step nine, judging whether the maximum iteration times G is reached, if not, making G equal to G +1,

returning to the step three; otherwise, terminating the iteration, outputting the Pareto optimal star set in the G iteration, and obtaining an optimal power distribution scheme according to the specific communication requirement.

2. The method according to claim 1, wherein the method comprises: step one, establishing a power control model of a double-layer heterogeneous network specifically comprises:

forming a Macrocell/Femtocell double-layer heterogeneous network: in one halfDiameter of R_mIn the macro cell, F numbers of R radii are randomly distributed_fThe coverage areas of the Femtocell are all in the Macrocell and are not intersected with each other, wherein N is randomly distributed in the whole Macrocell_mThe MUE of each macro base station user is randomly distributed in each Femtocell_fEach family base station user FUE adopts a mode of sharing frequency bands to divide frequency spectrum resources into Q subchannels, and all MUEs and FUEs share the Q subchannels; the channel allocation mode is as follows: firstly, N is_mRandomly extracting Q MUEs from the MUEs, equally distributing the Q MUEs to Q subchannels, and then distributing N_m-Q macro base station users are randomly allocated to Q sub-channels, where 2Q > N_m> Q, to FN_fRandomly distributing the family base station users to Q sub-channels, wherein Q > FN_fThe interference condition is specifically as follows:

same-layer interference of Macrocell layer: when the Macrocell wants to send a signal to a user i, if the user i and the user y of the Macrocell occupy channels simultaneously, the user y is an interfering user, and the interference of the Macrocell on the user i satisfies the following conditions:

wherein,

for the channel gain of Macrocell to its user i,

the power allocated to interfering user y for Macrocell,

the path loss from Macrocell to its user i; the channel gain from Macrocell to its user i is modeled as:

represents the distance between Macrocell and MUE;

cross-layer interference of Femtocell to MUE: when a Macrocell wants to send a signal to a user i thereof, if a Femtocell in the Macrocell coverage area also allocates a sub-channel for transmitting the signal sent by the Macrocell to the user i thereof to a user y thereof, the Femtocell is an interfering base station at the moment, and the interference to the user i of the Macrocell is as follows:

wherein,

for the Femtocell to normal user i channel gain,

the power allocated by the Femtocell for its user y,

for the path loss from the Femtocell to the normal user i, the channel gain of the Femtocell to the MUE satisfies:

Z_cis a loss factor; z^FBy passing

It was found that λ is the wavelength,

the distance between the FBS and the indoor MUE, the interference suffered by the ith MUE and the distance between the FBS and the indoor MUE are

Cross-layer interference of macrocells with FUEs: when a Femtocell wants to send a signal to its user i, the Macrocell also allocates the same transmission Femto to its user yWhen the cell sends a sub-channel of a signal of a user i, the Macrocell is an interference base station, and the interference on the user i of the Femtocell is as follows:

wherein,

for the channel gain of Macrocell to normal user i,

the power allocated for Macrocell for interfering user y,

for the path loss from Macrocell to normal user i, the channel gain from Macrocell to FUE is the same as the channel condition from Macrocell to MUE, so the model is:

expressing the distance between Macrocell and FUE, the path loss mentioned in the above formula can be defined as: the macro base station is provided with three antennas, each antenna is responsible for a 120-degree sector area, the areas responsible for the three antennas are added together to completely cover a circular action area of the macro base station, when an interfering user and an interfered user are in the sector area responsible for the same antenna, only one half of the original power is left when the power sent to the interfering user by the base station reaches the interfered user, namely the path loss is one half; when the interfering user and the interfered user are not in the sector area for which the same antenna is responsible, the power is reduced to the original quarter, i.e., the path loss is a quarter.

Adopting additive white Gaussian noise as environmental noise;

modeling system throughput in heterogeneous networks as

N_mIs the total number of macro base station users in the system; FN (FN)_fIs the total number of home base station users in the system;

and

respectively representing the throughput of the ith macro base station user and the throughput of the yth home base station user according to the Shannon formula

And

obtaining the results of the steps of, among others,

and

respectively representing the signal-to-noise ratio of the ith macro base station user and the yth femtocell user, and the specific result can be obtained

And

is found out of_iAnd H_yIndicates the channel gain, P, of the base station corresponding to the ith macro base station user and the yth home base station user during normal communication_iAnd P_yRespectively representing the power distributed to the base stations corresponding to the ith macro base station user and the yth femtocell user; g_iAnd G_yPath loss of the corresponding base station of the ith macro base station user and the yth femtocell user to the ith macro base station user during normal communicationConsumption; n is₀Representing ambient noise, M in the above formula_iAnd F_yRespectively representing the interference sum of cross-layer interference and same-layer interference of the ith macro base station user and the yth femtocell user;

taking the 1 st objective function system throughput maximization and the 2 nd objective function system energy consumption minimization as objective functions at the same time, converting the energy consumption minimization problem into a maximization problem, and establishing a mathematical model as a formula:

P_maxthe maximum power which can be transmitted by the system is the sum of the transmission power of the base station and the circuit loss, and the solution formula is

The maximum total power of the Macrocell,

for the maximum total power per Femtocell,

and

respectively representing the circuit loss of the Macrocell layer and the Femtocell layer, P is the actual transmitting power of the system, and the solving formula is

P_i ^mAnd

respectively being the ith macro base station user and the yth home base stationActual power of the user, P_maxThe smaller P, the less the value of the 2 nd objective function f₂The more preferred is that the more preferred is,

and

respectively representing the minimum value and the maximum value which can be reached by the actual power of the macro base station user and the home base station user.

3. The method according to claim 1, wherein the method comprises: initializing the positions and quantum positions of the stars, sorting all the stars in the non-dominated solution levels, and sorting all the stars in each level according to the crowding degree specifically comprises the following steps:

setting the maximum iteration times as G, the iteration number label as G, and G belonging to [1, G ]]Setting the number of stars in the galaxy to be L and the position of the first star in the g iteration to be

N_mFor the total number of MUEs, F is the total number of femtocells, N_fFor the total number of FUEs in each home base station, generation 1 initializes the constellation position to

And

respectively representing the minimum value and the maximum value which can be reached by the actual power of a macro base station user and a home base station user; setting up

To dominate the star

Number of stars, initial

Middle storage star

The designation of all stars dominated; will star body

The two fitness values are respectively compared with the fitness values of all the stars in the star system, if so, the fitness values are respectively compared with the fitness values of all the stars in the star system

And is

And when the equal signs are not simultaneously established, the star represents the star

Dominating stars

Will star body

Number q of (a) exists

In, if it is

And is

When the equal signs are not simultaneously established, the star is represented

Quilt body

At the main, then

If it is

Representing stars

If not dominated by any stars, its non-dominated solution rank is 1; traversing all stars dominated by the stars with the non-dominated solution level of 1, if the stars except the stars with the non-dominated solution level of 1 are not dominated by other stars, the non-dominated solution level of the stars is 2, and finding out the non-dominated solution level of all the stars according to the rule;

then, sorting the galaxies from the non-dominated solution level to 1, judging the number of the stars in each non-dominated solution level, and if a plurality of stars have the same non-dominated solution level, sorting the stars again from large to small according to the congestion degree; the method for judging the crowding degree comprises the following steps: the congestion degree is represented by the normalization of congestion distance, all stars are sorted from small to large according to the fitness value, and the calculation mode of the normalization of the congestion distance of the a-th fitness function of the ith star is

f_maxAnd f_minRepresents the non-dominant solutionThe maximum fitness value and the minimum fitness value of all stars in the level, and the crowding distance between the first star and the last star after sorting is set to be infinite; since there are several fitness degrees and there are several corresponding crowding distances, the average crowding distance of the first star

And

respectively represents the crowding distance corresponding to the two fitness values of the first star body, and the calculation formula is

And

setting the maximum cycle number to K₁Number of cycles is denoted by k₁，k₁∈[1,K₁]Kth iteration of the g₁The quantum position of the first star in the subcycle can be expressed as

Wherein

And

are respectively defined as

And

kth iteration of g₁The position of the first star in the subcycle is

4. The method according to claim 1, wherein the method comprises: step three, the championship selection mechanism is carried out, and the selection of a new galaxy is specifically as follows:

selecting T different stars from all stars in the galaxy, comparing the star with each other only once, comparing the non-dominated solution levels, selecting the non-dominated solution level with the smallest value, selecting the crowding distance with the largest value if the non-dominated solution levels are the same, selecting the first one if the crowding distances are the same, and selecting the first one by multiple selections

The individual stars form a new galaxy, L is the number of stars in the initial galaxy, T is the number of stars selected each time, the serial number of each star in the selected new galaxy is stored

In then according to

And finding out the quantum position of the corresponding star.

5. The method according to claim 1, wherein the method comprises: fourthly, updating the quantum rotation angle according to the position chaotic change, wherein the optimizing search process of using the simulated quantum rotation gate evolution galaxy specifically comprises the following steps:

for the k-th₁The first star in the secondary cycle,

produce [0,1]Uniform random number

If it is

There is the kth iteration in the g-th iteration₁The quantum rotation gate corresponding to the j dimension quantum rotation angle of the l star in the +1 circulation corresponds to

Using analog quantum rotating gate to k in g iteration₁The j-dimension quantum position of the l star in the +1 circulation is updated and can be expressed as

Kth iteration in g₁The j dimension of the first star in the star system in +1 circulation corresponds to the quantum rotation angle

Is the kth iteration in the g₁The index of change in the position of the first star in the +1 cycle is set to

abs () represents an absolute value operation on each dimension variable of a vector in parentheses;

is the kth iteration in the g₁The corresponding position change step length of j dimension of each star in +1 cycle

Calculation of b_maxIs the maximum position change step size and,

is the kth iteration in the g₁The chaos index of the first star of the +1 circulation can be calculated through a chaos sequence, and the formula is as follows

Initial value

Is a [0,1 ]]The random number of (2); mapping the quantum position of the star to obtain the kth iteration₁The position corresponding to the first star in the +1 circulation is

And

respectively representing the minimum value and the maximum value which can be reached by the actual power of a macro base station user and a home base station user; when in use

And is

Then, the k-th iteration in the g-th iteration₁The first star position change index in the +1 cycle is

If not, then,

let k₁＝k₁+1, return to step four.

If it is

Change the kth iteration in the g iteration₁The j-th dimension of the l star in +1 cycle has a quantum rotation angle of

The quantum revolving gate corresponding to the j dimension of the ith star can be expressed as

Using analog quantum revolving gate, for the kth₁The j-dimension quantum position of the l star in the +1 circulation is updated to

Mapping the quantum position of the star to obtain the kth iteration₁The j-th dimension of the l star in the +1 circulation corresponds to a position:

when in use

And is

Then, the k-th iteration in the g-th iteration₁The first star position change index in the +1 cycle is

If not, then,

let k₁＝k₁+1, return to step four.

6. The method according to claim 1, wherein the method comprises: step six, performing positive and negative rotation chaotic movement on all stars, and finding a more optimal star system specifically comprises the following steps:

for the k-th₂The first star in the secondary cycle,

generating uniform random numbers

If it is

The star undergoes a rotational movement, at the kth iteration in the g-th iteration₂During +1 circulation, the jth dimension of the first star position in the star system is updated to

Is the kth iteration in the g₂The chaos index of the first star of the +1 circulation is determined by the chaos sequence, and the chaos sequence formula can be expressed as

Initial chaotic value

Is a [0,1 ]]A uniform random number of;

is the kth iteration in the g₂The j-th dimension of each star in the +1 circulation corresponds to the position change step length, and the calculation formula is

Position change step of the first star in the initial cycle

Also produced by chaotic sequences; kth iteration in g₂The rotation angle of the first star in +1 cycle is

Rotation angle of the first star of the first cycle

Is the chaos index of the first star in the initial cycle; when in use

When the temperature of the water is higher than the set temperature,

adding a judging mechanism, i.e.

And

respectively representing the minimum value and the maximum value which can be reached by the actual power of a macro base station user and a home base station user; preserving better solutions before and after updating a location through a greedy mechanism

If it is

The kth iteration in the g-th iteration is repeated again₂The ith star makes negative rotation in +1 cycle, and the updated jth dimension position can be expressed as

Likewise, a judgment mechanism is added, i.e.

And

respectively representing the minimum value and the maximum value which can be reached by the actual power of a macro base station user and a home base station user; reselecting better solution before and after updating position

7. The method according to claim 1, wherein the method comprises: step eight, mixing the obtained new galaxies with the initial galaxies, and selecting the galaxies with the same scale as the initial galaxies specifically comprises the following steps:

new star systems

And the astrology S obtained before evolution_l ^g(1) Mixing to form a temporary galaxy, and sequencing non-dominated solutions and crowdedness according to the second step; in the temporary galaxy, firstly, selecting from the small non-dominated solution level until L stars are selected; if the total star number does not exceed L after all stars in the non-dominated solution level are added, all the stars in the non-dominated solution level are selected, if the total star number exceeds L, the stars in the non-dominated solution level are sorted according to the crowding degree, the crowding distance is preferentially selected to be larger until L stars are reached, and a new galaxy is obtained

Images