CN102664665A - Alternative-optimization and rate-maximization multi-point cooperation wave beam forming method - Google Patents

Abstract

The invention discloses a single base station power constraint and rate maximization multipoint cooperation wave beam forming and power distribution method. The method comprises the following steps: firstly, converting sum rate maximization problem into an emission power minimization problem satisfying a minimum rate requirement; secondly, using a second-order cone programming optimization method to solve the emission power minimization problem so as to obtain an emission wave beam vector and emission power; then, taking the obtained wave beam vector as a constant, converting the sum rate maximization optimization problem into the sum rate maximization optimization problem of a single-input single-output communication network, and using a convex approximation method and a geometric programming optimization method to solve the sum rate maximization optimization problem so as to obtain the emission power of the sum rate maximization when the emission wave beam vector is given. Compared to the current sum rate maximization multipoint cooperation wave beam forming method, by using the method of the invention, calculating complexity is low and the sum rate obtained through the method of the invention approximates the optimal sum rate searched by an exhaustive method.

Description

A kind of alternately optimization and speed maximization multipoint cooperative beam-forming method

Technical field

The invention belongs to wireless communication technology field, be specifically related to a kind of single base station power constraint down maximize multipoint cooperative beam forming and power distribution method with speed.

Background technology

In order to improve the utilization ratio of frequency spectrum; It is that 1 spectrum reuse mode makes up communication network that cell mobile communication systems generally adopts the spectrum reuse factor; So not only realize the increase of frequency spectrum service efficiency, and alleviated the arrangement pressure of plot planning and base-station node.But it is more serious that the cellular mobile network that makes up has by this method produced the co-channel interference influence that serious presence of intercell interference, particularly Cell Edge User receive, and this has caused the throughput performance of cell mobile communication systems network to be had a strong impact on.Recently, in order to improve the systematic function of community marginal user performance and whole communication network, the co-channel interference of utilizing minizone beam forming and Poewr control method to suppress the minizone becomes a big research focus of wireless communication field.At present; The design of multipoint cooperative beam forming and power distribution method mainly concentrates between the cooperative base station under the total power constraint condition; The problem of design beam vector and power division how between the cooperative base station; Under the every base station power constraints of fewer literature research, the problem of design beam vector and power division how between the cooperative base station.For this reason, the multipoint cooperative beam forming under a kind of single base station power constraints that the present invention is based on the duality theory principle design and the optimization method of power division.

Summary of the invention

Technical problem: the invention provides a kind of alternately optimization and speed maximization multipoint cooperative beam-forming method, be a kind of computation complexity lower, can realize that single base station power constraint and the speed that separates on speed Pareto border maximize multipoint cooperative beam forming and power distribution method.

Technical scheme: of the present invention alternately optimization and speed maximization multipoint cooperative beam-forming method may further comprise the steps:

1) Initialize the transmit beam vector get the initial value of the transmit beam vector Initialize transmit power

get initial transmit power value

use beam vector

Transmit Power

and Formulas initialization feasible SINR

obtain an initial feasible SINR

using the initial feasible SINR ratios and formulas initialize

and

get the initial auxiliary variables

and

is

b, the base station transmit beam vector;

is

the transmit power of the base station b;

is that

is the Signal to Interference plus Noise Ratio of user u, and computing formula does

${\mathrm{\γ}}_u^{\left(n\right)}=\frac{p_{b(b=u)}{\left|\right|h_{u,b(b=u)}^H{w}_{b(b=u)}\left|\right|}^2}{\underset{b\≠u}{\overset{K}{\underset{b=1}{\mathrm{\Σ}}}}{p}_b{\left|\right|h_{u,b}^H{w}_b\left|\right|}^2+{\mathrm{\σ}}_u^2};$

For

Be the approximate auxiliary variable of speed of user u, computing formula is: ${\mathrm{\α}}_u=\frac{{\left({\mathrm{\γ}}_u^{\left(n\right)}\right)}^2}{1+{\mathrm{\γ}}_u^{\left(n\right)}},$

For

Be the approximate auxiliary variable of speed of user u, computing formula is: ${\mathrm{\β}}_u=\frac{{\mathrm{\γ}}_u^{\left(n\right)}}{1+{\mathrm{\γ}}_u^{\left(n\right)}}+\mathrm{Ln}(1+{\mathrm{\γ}}_u^{\left(n\right)}),$

K is the quantity of cooperative base station;

N is the algorithm iteration number of times, and initial value is 0;

U is a Customs Assigned Number;

B is the base station numbering;

B=u representes that the serving BS of user u is base station b;

B ≠ u representes that the serving BS of user u is not base station b, is interference base station;

representes b=1;, K;

representes u=1;, K;

2). utilize

and second order cone planing method optimization solving-optimizing problem:

$\underset{{\left\{{\tilde{w}}_b\right\}}_{b=1}^K}{\mathrm{Min}}\underset{b=1}{\overset{K}{\mathrm{\&Sigma;}}}{\left|\right|{\tilde{w}}_b\left|\right|}^{2}s.t.{\mathrm{\&gamma;}}_u^{\left(n\right)}\&le;\frac{{\left|\right|h_{u(b=u),b}^H{\tilde{w}}_b\left|\right|}^2}{\underset{b=1,b\&NotEqual;u}{\overset{K}{\mathrm{\&Sigma;}}}{\left|\right|h_{u,b}^H{\tilde{w}}_b\left|\right|}^2+{\mathrm{\&sigma;}}_{\mathrm{<figure-callout\; id="2"\; label="u"\; filenames="120323100303.png"\; state="\{\{state\}\}">u</figure-callout>}}^2},$

${\left|\right|{\tilde{w}}_b\left|\right|}^2\&le;P_b,$

Obtain separating of optimization problem The compute beam vector With interim transmitting power

h _{U, b}Be the channel coefficients of base station b to user u;

is the noise variance of user u;

h _{U, b} ^HExpression h _{U, b}Grip the transposition computing altogether;

is

is the base station transmit beam temporary vector b;

The optimal solution beam vector of

expression base station b;

P _bMaximum transmission power constraint for base station b;

p _bTransmitting power for base station b;

is the interim transmitting power of base station b;

3). utilize the launching beam vector

Auxiliary variable

With geometric programming optimization method solving-optimizing problem: $\underset{{\left\{p_b\right\}}_{b=1}^K}{\mathrm{Min}}\underset{u=1}{\overset{K}{\mathrm{\&Sigma;}}}\left[{\mathrm{\&alpha;}}_u^{\left(n\right)}{\left|\right|h_{u,b(b=u)}^H{w}_{b(b=u)}^{(n+1)}\left|\right|}^{-2}{p}_{b(b=u)}^{-1}({\mathrm{\&sigma;}}_u^2+\underset{b\&NotEqual;u}{\overset{K}{\underset{b=1}{\mathrm{\&Sigma;}}}}{\left|\right|h_{u,b}^H{w}_b^{(n+1)}\left|\right|}^{-2}{p}_b)\right],$ S.t.0≤p _b≤P _b,

Obtain its optimal solution

4) Use the transmit beam vector and transmit power Update feasible SINR

and update the auxiliary variables

and

5) if.

sets up, and then exports launching beam vector and transmitting power

otherwise gets back to step 3);

ξ is predefined required precision;

The object that the inventive method is used is the multi-base station cooperative communication system, comprises K cooperative base station, and all there is the M transmit antennas each base station, and a single antenna user is only served in each base station, and the maximum transmission power of base station b is P _b

Beneficial effect: the inventive method is compared with the maximized exhaustive search algorithm of speed, and computation complexity is low, and arithmetic speed is fast, and inventive method can realize that the Pareto border of user rate separates.

Description of drawings

Fig. 1 is the system model of the inventive method;

Fig. 2 is single base station power constraint multipoint cooperative beam forming and power distribution method flow chart;

Fig. 3 separates performance curve for the Pareto Boundary border of institute's extracting method;

Fig. 4 is the optimum and rate capability curve of institute's extracting method;

Fig. 5 is institute's extracting method rate of convergence performance curve;

Have among the figure: base station 1, user terminal 2;

Embodiment

Concrete theoretical foundation explanation:

The present invention is directed to the single output of the many inputs multipoint cooperative downlink system of single base station power constraint, turn to optimization aim to optimize solving system and speed maximum, that is:

$\underset{{\{w_b,p_b\}}_{b=1}^K}{\max }\underset{u=1}{\overset{K}{\mathrm{\&Sigma;}}}{R}_u---\left(1\right)$

$s.t.\left|\right|w_b\left|\right|=1,p_b\&le;P_b,\&ForAll;b$

Wherein:

The speed of

expression user u, its unit is Nat/second/hertz;

${\mathrm{\&gamma;}}_u=\frac{p_{b(b=u)}{\left|\right|h_{u,b(b=u)}^H{w}_{b(b=u)}\left|\right|}^2}{\underset{b\&NotEqual;u}{\overset{K}{\underset{b=1}{\mathrm{\&Sigma;}}}}{p}_b{\left|\right|h_{u,b}^H{w}_b\left|\right|}^2+{\mathrm{\&sigma;}}_u^2},$

The Signal to Interference plus Noise Ratio of expression user u;

Ln (x) representes natural logrithm;

Be { w ₁..., w _K, w _bLaunching beam vector for base station b;

Be { p ₁..., p _K, p _bThe transmitting power of base station b;

P _bMaximum transmission power constraint for base station b;

K is the quantity of cooperative base station;

U is a Customs Assigned Number;

B is the base station numbering;

B=u representes that the serving BS of user u is base station b;

B ≠ u representes that the serving BS of user u is not base station b, is interference base station;

representes b=1;, K;

representes u=1;, K;

Because optimization problem (1) is the NP optimization problem, usually, people are difficult to find its optimal solution.Therefore, to the multipoint cooperative communication system of the single output of list input, optimization problem (1) obtains many research, but because the introducing of launching beam vector is more increased the degree of coupling of optimization problem, like this, finding the solution of optimization problem is difficult more.The inventive method utilization replaces optimization method and the convex function approximation method is found the solution above-mentioned optimization problem.At first, given attainable targeted customer's speed or attainable target Signal to Interference plus Noise Ratio are converted into the transmitting power that realizes targeted customer's speed with finding the solution of optimization problem (1) and minimize optimization problem, that is:

$\underset{{\left\{{\tilde{w}}_b\right\}}_{b=1}^K}{\min }\underset{b=1}{\overset{K}{\mathrm{\&Sigma;}}}{\left|\right|{\tilde{w}}_b\left|\right|}^2$

$s.t.{\mathrm{\&gamma;}}_u^T\&le;\frac{{\left|\right|h_{u,b(b=u)}^H{\tilde{w}}_{b(b=u)}\left|\right|}^2}{\underset{b\&NotEqual;u}{\overset{K}{\underset{b=1}{\mathrm{\&Sigma;}}}}{\left|\right|h_{u,b}^H{\tilde{w}}_b\left|\right|}^2+{\mathrm{\&sigma;}}_{\mathrm{<figure-callout\; id="2"\; label="u"\; filenames="120323100303.png"\; state="\{\{state\}\}">u</figure-callout>}}^2},\&ForAll;u---\left(2\right)$

${\left|\right|{\tilde{w}}_b\left|\right|}^2\&le;P_b,\&ForAll;b$

Wherein,

expression user u can realize the target Signal to Interference plus Noise Ratio,

is

is the base station transmit beam temporary vector b;

Utilize existing ripe optimization method second order cone planing method at an easy rate solving-optimizing problem (2) be met the launching beam of attainable targeted customer's speed, make transmitting power user rate minimum and that the end user can realize be targeted customer's speed simultaneously.Optimal beam vector at the launching beam that has obtained to satisfy attainable targeted customer's speed; We can obtain to realize targeted customer's speed accordingly and minimize the beam vector of transmitting power

also with after it substitution and the speed maximization optimization problem; Can be translated into single output of single input and speed maximization optimization problem; But, remain an optimization problem that is difficult for finding the solution like this because transmitting power still intercouples.For this reason, utilize convex function to be similar to inequality

Wherein α and β be two only with an approximate some x ₀Relevant auxiliary variable, their computational methods are:

$\mathrm{\&alpha;}=\frac{x_0^2}{1+x_0}$ (3)

$\mathrm{\&beta;}=\frac{x_0}{1+x_0}+\ln (1+x_0)$

Can realize target Signal to Interference plus Noise Ratio place, single output of single input and speed maximized the optimization problem that optimization problem is converted into the contrary Signal to Interference plus Noise Ratio sum of minimizing Weighted, that is:

$\underset{{\left\{p_b\right\}}_{b=1}^K}{\min }\underset{u=1}{\overset{K}{\mathrm{\&Sigma;}}}\left[{\mathrm{\&alpha;}}_u{\left|\right|h_{u,b(b=u)}^H{w}_{b(b=u)}^T\left|\right|}^{-2}{p}_{b(b=u)}^{-1}({\mathrm{\&sigma;}}_u^2+\underset{b\&NotEqual;u}{\overset{K}{\underset{b=1}{\mathrm{\&Sigma;}}}}{\left|\right|h_{u,b}^H{w}_b^T\left|\right|}^{-2}{p}_b)\right]---\left(4\right)$

$s.t.0\&le;p_b\&le;P_b,\&ForAll;b$

Wherein,

${\mathrm{\&alpha;}}_u=\frac{{\left({\mathrm{\&gamma;}}_u^T\right)}^2}{1+{\mathrm{\&gamma;}}_u^T},\&ForAll;U$ (5)

${\mathrm{\&beta;}}_u=\frac{{\mathrm{\&gamma;}}_n^T}{1+{\mathrm{\&gamma;}}_u^T}+\ln (1+{\mathrm{\&gamma;}}_u^T),\&ForAll;u$

Like this, single output of single input and speed maximization optimization problem can utilize the geometric programming method to find the solution.After finding the solution optimal solution, simultaneously two auxiliary variable α and β are upgraded, can obtain in new transmitting power like this and can realize the largest user speed under targeted customer's rate conditions.

The multipoint cooperative beam forming and the power distribution method of a kind of single base station power constraint of the present invention may further comprise the steps:

1) Initialize the transmit beam vector

get the initial value of the transmit beam vector

Initialize transmit power

get initial transmit power value

use beam vector Transmit Power and Formulas initialization feasible SINR

obtain an initial feasible SINR

using the initial feasible SINR ratios and formulas initialize

and get the initial auxiliary variables

and

is

b, the base station transmit beam vector;

is

the transmit power of the base station b;

For

Be the Signal to Interference plus Noise Ratio of user u, computing formula does ${\mathrm{\&gamma;}}_u^{\left(n\right)}=\frac{p_{b(b=u)}{\left|\right|h_{u,b(b=u)}^H{w}_{b(b=u)}\left|\right|}^2}{\underset{b\&NotEqual;u}{\overset{K}{\underset{b=1}{\mathrm{\&Sigma;}}}}{p}_b{\left|\right|h_{u,b}^H{w}_b\left|\right|}^2+{\mathrm{\&sigma;}}_{\mathrm{<figure-callout\; id="2"\; label="u"\; filenames="120323100303.png"\; state="\{\{state\}\}">u</figure-callout>}}^2};$

For

Be the approximate auxiliary variable of speed of user u, computing formula is: ${\mathrm{\&alpha;}}_u=\frac{{\left({\mathrm{\&gamma;}}_u^{\left(n\right)}\right)}^2}{1+{\mathrm{\&gamma;}}_u^{\left(n\right)}},$

For

Be the approximate auxiliary variable of speed of user u, computing formula is: ${\mathrm{\&beta;}}_u=\frac{{\mathrm{\&gamma;}}_u^{\left(n\right)}}{1+{\mathrm{\&gamma;}}_u^{\left(n\right)}}+\mathrm{Ln}(1+{\mathrm{\&gamma;}}_u^{\left(n\right)}),$

K is the quantity of cooperative base station;

N is the algorithm iteration number of times, and initial value is 0;

U is a Customs Assigned Number;

B is the base station numbering;

B=u representes that the serving BS of user u is base station b;

B ≠ u representes that the serving BS of user u is not base station b, is interference base station;

representes b=1;, K;

representes u=1;, K;

2). utilize and second order cone planing method optimization solving-optimizing problem:

$\underset{{\left\{{\tilde{w}}_b\right\}}_{b=1}^K}{\mathrm{Min}}\underset{b=1}{\overset{K}{\mathrm{\&Sigma;}}}{\left|\right|{\tilde{w}}_b\left|\right|}^{2}s.t.{\mathrm{\&gamma;}}_u^{\left(n\right)}\&le;\frac{{\left|\right|h_{u(b=u),b}^H{\tilde{w}}_b\left|\right|}^2}{\underset{b=1,b\&NotEqual;u}{\overset{K}{\mathrm{\&Sigma;}}}{\left|\right|h_{u,b}^H{\tilde{w}}_b\left|\right|}^2+{\mathrm{\&sigma;}}_{\mathrm{<figure-callout\; id="2"\; label="u"\; filenames="120323100303.png"\; state="\{\{state\}\}">u</figure-callout>}}^2},$

${\left|\right|{\tilde{w}}_b\left|\right|}^2\&le;P_b,$

Obtain separating of optimization problem

The compute beam vector With interim transmitting power

h _{U, b}Be the channel coefficients of base station b to user u;

is the noise variance of user u;

h _{U, b} ^HExpression h _{U, b}Grip the transposition computing altogether;

is

is the base station transmit beam temporary vector b;

The optimal solution beam vector of

expression base station b;

P _bMaximum transmission power constraint for base station b;

p _bTransmitting power for base station b;

is the interim transmitting power of base station b;

3). utilize the launching beam vector

Auxiliary variable

With geometric programming optimization method solving-optimizing problem: $\underset{{\left\{p_b\right\}}_{b=1}^K}{\mathrm{Min}}\underset{u=1}{\overset{K}{\mathrm{\&Sigma;}}}\left[{\mathrm{\&alpha;}}_u^{\left(n\right)}{\left|\right|h_{u,b(b=u)}^H{w}_{b(b=u)}^{(n+1)}\left|\right|}^{-2}{p}_{b(b=u)}^{-1}({\mathrm{\&sigma;}}_u^2+\underset{b\&NotEqual;u}{\overset{K}{\underset{b=1}{\mathrm{\&Sigma;}}}}{\left|\right|h_{u,b}^H{w}_b^{(n+1)}\left|\right|}^{-2}{p}_b)\right],$ S.t.0≤p _b≤Pb,

Obtain its optimal solution

4) Use the transmit beam vector

and transmit power

Update feasible SINR

and update the auxiliary variables

and

5) if.

sets up, and then exports launching beam vector

and transmitting power

otherwise gets back to step 3);

ξ is predefined required precision;

Make an explanation in the face of the performance comparison of the inventive method and additive method down:

In analogous diagram, " user rate border " refers under every base station power constraints, utilizes the optimization of user rate location mode to find the solution the user rate that is obtained under the different rates distribution of weights; " optimum and speed " refers under every base station power constraints, optimum and the speed of utilizing the exhaustive search of user rate location mode to arrive; " method one (two, three, four) " refer under every base station power constraints; At close-to zero beam shaping (high specific sends beam shaping, letter leaks and makes an uproar than beam shaping, accidental beam shaping) initial method with under with the acc power initialization condition, optimum that institute's extracting method is obtained and speed (user rate); " method five " refers to the multipoint cooperative beam forming method of the ratio of signal power and leakage signal power plus noise power sum under every base station power constraints; " method six " refers to the high specific launching beam forming method under every base station power constraints; " method seven " refers to the zero launching beam forming method of compeling under every base station power constraints.

Fig. 3 has provided the whole bag of tricks optimum of two base station collaborations and the user rate Pareto boundary curve of speed and the realization of base station user rate location mode.Emulation shows, institute's extracting method can realize all that in different initial method conditions user rate Pareto border separates; The SGINR method can realize that also user rate Pareto border separates; But additive method can not guarantee to realize that user rate Pareto border separates; When Fig. 4 has provided different initial methods and two or three base station collaborations, the optimum of institute's extracting method and rate capability curve.Simulation result shows that the optimal solution of institute's extracting method is relevant with initial method, and promptly institute's extracting method resulting optimum under different initial method conditions is different with speed, but the optimum and the speed ratio that arrive with the exhaustive search of rate distribution method are more approaching; Fig. 5 has provided rate of convergence and relation iterations between of institute's extracting method under different initial method conditions, and simulation result shows institute's extracting method, and only needs 3 or 4 iteration just can converge to point of safes; Rate distribution method exhaustive search maximum and speed then need hundreds of subrate distribution optimizations to find the solution at least could obtain optimum and speed point.

Claims (1)

1. alternately optimize and speed maximization multipoint cooperative beam-forming method for one kind, it is characterized in that this method may further comprise the steps:

1) Initialize the transmit beam vector?

to obtain the initial value of the transmit beam vector?

Initialize transmit power?

get initial transmit power value?

use beam vector?

transmit power?

and Formulas initialization feasible SINR? obtain an initial feasible SINR?

using the initial feasible SINR and Formulas initialization?

and?

get the initial auxiliary variables?

and?

is?

b, the base station transmit beam vector;

is?

base station transmit power of b;

is that

is the Signal to Interference plus Noise Ratio of user u, and computing formula does

is the speed approximate auxiliary variable of

for user u, and computing formula is:

is the speed approximate auxiliary variable of

for user u, and computing formula is:

K is the quantity of cooperative base station;

N is the algorithm iteration number of times, and initial value is 0;

U is a Customs Assigned Number;

B is the base station numbering;

B=u representes that the serving BS of user u is base station b;

B ≠ u representes that the serving BS of user u is not base station b, is interference base station;

representes b=1;, K;

representes u=1;, K;

2) Use?

and SOCP method optimization for solving optimization problems:?

get optimization problem?

calculate beam vector?

and temporary transmitter power?

h _{U, b}Be the channel coefficients of base station b to user u;

is the noise variance of user u;

h _{U, b} ^HExpression h _{U, b}Grip the transposition computing altogether;

is?

as a temporary base station transmit beam vector b;

The optimal solution beam vector of

expression base station b;

P _bMaximum transmission power constraint for base station b;

p _bTransmitting power for base station b;

is the interim transmitting power of base station b;

3). utilize the launching beam vector

Auxiliary variable With geometric programming optimization method solving-optimizing problem:

S.t.0≤p _b≤P _b,

Obtain its optimal solution

4) Use the transmit beam vector?

and transmit power?

Update feasible SINR? and update the auxiliary variables?

and?

5) if.

sets up, and then exports launching beam vector and transmitting power

otherwise gets back to step 3);

ξ is predefined required precision.

Images