CN103023592B - Fairness based cognitive radio-frequency spectrum resource management algorithm - Google Patents

Abstract

The invention discloses a fairness based cognitive radio-frequency spectrum resource management algorithm. The fairness based cognitive radio-frequency spectrum resource management algorithm includes the steps of (1) subjecting underutilized or undistributed spectrum resource to channel division according to different network type; (2) describing parameters of the underutilized or undistributed spectrum resource; (3) introducing three variable parameters a, b and c, increasing the a to increase influence of channel width wt which can be obtained currently, obtaining larger total system revenue and reducing distribution fairness at the same time, and increasing the b and the c to improve the distribution fairness and reduce the total system revenue at the same time, wherein total bandwidth w, the channel width wt of the network type t and the channel quantity Mt of the network type t meet the Formula Mt=floor(w/wt), and the Formula is floor operation.

Description

A kind of cognitive radio frequency spectrum Resource Management Algorithm based on fairness

Technical field

The invention belongs to wireless communication field, specifically a kind of cognitive radio intermediate frequency spectrum Resource Management Algorithm based on fairness.In obtaining optimizer system income, ensure the reliable communication of frequency spectrum resource distributional equity and node.

Background technology

Radio spectrum resources is fixed division by radio spectrum management department always, gives certain business, certain wireless traffic operator (or group, department etc.) by certain frequency range fixed allocation.Frequency range after division can only be used by the user under this operator, and other unauthorized users can not use this frequency range, even the in the situation that of the authorized user free time.But, more and more studies have shown that this method of salary distribution is that utilance is low-down, and can not meet the growing demand to radio spectrum resources of user.

Based on this, JosephMitola in 1999 have proposed the concept of cognitive radio (CR, Cognitive Radio).Although CR is assigned with some, untapped frequency spectrum and some are used but the frequency spectrum that is not fully utilized is looked for out, and again distribute and use, thereby reach higher spectrum utilization efficiency, alleviated to a certain extent the contradiction of current frequency spectrum anxiety and business development.Be accompanied by the rise of cognitive radio technology, Dynamic Spectrum Management problem is rich in challenging problem as one in cognitive radio and is also more and more paid close attention to by more people.

Dynamic Spectrum Management is in all feasible channel allocation, to select optimum channel allocation set U*, reaches the optimization of target function with this.In more than ten years in the past, a lot of researchers have proposed the effective Dynamic Spectrum Management algorithm based on different network environments, different target function.What for example, in the article " Collaboration and fairness in opportunistic spectrum access " of Haitao Zheng and Chunyi Peng, propose has designed three kinds of different algorithms based on three kinds of different target functions.This locality by graph theory for example proposing in the article " List-coloring based channel allocation for open-spectrum wireless networks " of Wei Wangand Xin Liu again problem (LCP, local-coloring problem) of tinting is applied to the algorithm of Dynamic Spectrum Management.But much the research based on graph theory, by introducing the concept of " neighbours ", connects spectrum allocation may problem with neighbours' spectrum allocation may, has simplified between node and has disturbed, and ignored the cumulative interference of non-neighbor user.As at L.Yang, L.Cao, in the article " Physical interference driven dynamic spectrum management " of and H. Zheng, point out, when multiple link, operation can cause enough large cumulative interference, thereby cause bust this, even can not impact transmission when these links move alone.Thereby a kind of system framework PLAN(physical conflict graph generator proposed in this section of article), this framework improves the allocation performance of algorithm by considering the suffered cumulative interference of node.Subsequently, at SooyeolIm, Yunseok Kang, Wonsop Kim, Seunghee Kim, in the article " Dynamic spectrum allocation with efficient SINR-based interference management " of Jinup Kim and HyuckjaeLee, propose the heterogeneous network based on center type, proposed a kind of effective spectrum management algorithm, but this algorithm has been ignored distributional equity problem.

Based on the heterogeneous network of center type, this paper combines equitable proportion algorithm (PF, proportional fairness) and physical disturbance model, has ensured the reliable communication of distributional equity and node in obtaining optimizer system income.

Summary of the invention

The present invention proposes a kind of cognitive radio frequency spectrum Resource Management Algorithm based on fairness that has ensured the reliable communication of distributional equity and node in obtaining optimizer system income.A cognitive radio frequency spectrum Resource Management Algorithm based on fairness, is characterized in that: it comprises the following steps:

The first step: the frequency spectrum resource of underusing or not being assigned with is carried out to channel distribution according to different network type;

Second step: describe the parameter of the frequency spectrum resource of underusing or not being assigned with, the channel quantity of the total bandwidth W of the frequency spectrum resource of wherein underusing or not being assigned with, the channel width Wt of network type t and network type t wherein it is rounding operation;

The 3rd step: introduce three variable element a, b, c, increase a and increase current obtainable channel width w _timpact, and obtain larger system total revenue, reduce fairness in distribution simultaneously, increase b and c and improve fairness in distribution, reduce system total revenue simultaneously.

On the basis of technique scheme, described the 3rd step also comprises, each base station is selected an available network model voluntarily, then the channel of map network type is carried out to competitive bidding, and base station 1 is to channel m _t(m _t=1 ..., M _t) carry out competitive bidding, the target function design in its i time circulation is as follows:

$\underset{\underset{\underset{(l,j)\∈J_i-1}{(n,j)\∈J_i-1}}{(l,m_t)\∉J_i-1}}{\max }\frac{b_{l,m_t}.w_t^a}{{({\mathrm{\Σ}}_{n\∈B_{i-1}}\underset{j=1}{\overset{M_t}{\mathrm{\Σ}}}{w}_t.a_{n,j})}^b.{(\underset{j=1}{\overset{M_t}{\mathrm{\Σ}}}{w}_t.a_{l,j})}^c}$

Wherein J _i-1until the BTS channel pair that the i-1 time circulation distributed after finishing b _i-1j _i-1in all base stations, that base station l is to channel m _tbidding price, w _tthe channel width of network type t, a _n,j(j=1 ..., M _t), (n, j) ∈ J _i-1i-1 the circulation allocation matrix of network type t afterwards, and a _{n, j}be defined as:

On the basis of technique scheme, further consider the cumulative interference between base station, as follows based on physical disturbance model optimization:

Constraints design is as follows:

Be constrained in $\frac{\frac{p_k}{d_k^{\mathrm{\α}}}}{{\mathrm{\Σ}}_{n\∈B_{i^{\′}},n\≠k}\underset{m_n=1}{\overset{M}{\mathrm{\Σ}}}{c}_{m_n,m_k}{a}_{n,m_n}\frac{p_n}{d_{n,k}^{\mathrm{\α}}}+P_N}\≥\mathrm{\β},$

For all (k, m _k) ∈ J _i'

Wherein J _i'=J _i-1∪ (l, m _t), B _i'to belong to J _i'all base stations, i.e. B _i'=B _i-1∪ l, P _kthe through-put power of base station k, P _nnoise power, d _kthe distance between base station k and its terminal use, d _n,kit is the distance between base station n and any user of base station k; Wherein represent that the current base station-channel that will distribute is to (l, m _t) be subject to likely disturb summation, simultaneously be defined as:

When be 1 and show channel m _nwith channel m _kidentical or overlapping, be 0 o'clock channel m _nwith channel m _kbetween noiseless mutually.

Can provide service more flexibly with respect to prior art the present invention, we can obtain the allocation result of high fairness, low system benefit, also can obtain the high yield allocation result of relatively high fairness simultaneously, by utilizing physical disturbance model, effectively suppress the cumulative interference between base station, thereby ensured terminal use's reliable communication.

Brief description of the drawings

Fig. 1 the present invention carries out channel distribution according to different network type to CAB frequency spectrum resource;

Fig. 2 network insertion type of the present invention and bidding price;

Fig. 3 system total revenue contrast of the present invention;

Fig. 4 fairness in distribution contrast of the present invention;

Fig. 5 SINR cumulative distribution function contrast of the present invention (SINR thresholding β=5dB);

Fig. 6 SINR cumulative distribution function contrast of the present invention (SINR thresholding β=10dB).

Embodiment

Please refer to Fig. 1 and Fig. 2, a kind of cognitive radio frequency spectrum Resource Management Algorithm based on fairness of the present invention comprises the following steps,

The first step: the frequency spectrum resource of underusing or not being assigned with is carried out to channel distribution according to different network type;

Second step: describe the parameter of the frequency spectrum resource of underusing or not being assigned with, the channel quantity of the total bandwidth W of the frequency spectrum resource of wherein underusing or not being assigned with, the channel width Wt of network type t and network type t wherein it is rounding operation;

The 3rd step: introduce three variable element a, b, c, increase a and increase current obtainable channel width w _timpact, and obtain larger system total revenue, reduce fairness in distribution simultaneously, increase b and c and improve fairness in distribution, reduce system total revenue simultaneously.

On the basis of technique scheme, described the 3rd step also comprises, each base station is selected an available network model voluntarily, then the channel of map network type is carried out to competitive bidding, and base station 1 is to channel m _t(m _t=1 ..., M _t) carry out competitive bidding, the target function design in its i time circulation is as follows:

$\underset{\underset{\underset{(l,j)\∈J_i-1}{(n,j)\∈J_i-1}}{(l,m_t)\∉J_i-1}}{\max }\frac{b_{l,m_t}.w_t^a}{{({\mathrm{\Σ}}_{n\∈B_{i-1}}\underset{j=1}{\overset{M_t}{\mathrm{\Σ}}}{w}_t.a_{n,j})}^b.{(\underset{j=1}{\overset{M_t}{\mathrm{\Σ}}}{w}_t.a_{l,j})}^c}$

Wherein J _i-1until the BTS channel pair that the i-1 time circulation distributed after finishing b _i-1j _i-1in all base stations, that base station l is to channel m _tbidding price, w _tthe channel width of network type t, a _n,j(j=1 ..., M _t), (n, j) ∈ J _i-1i-1 the circulation allocation matrix of network type t afterwards, and a _{n, j}be defined as:

On the basis of technique scheme, further consider the cumulative interference between base station, as follows based on physical disturbance model optimization:

Constraints design is as follows:

Be constrained in $\frac{\frac{p_k}{d_k^{\mathrm{\α}}}}{{\mathrm{\Σ}}_{n\∈B_{i^{\′}},n\≠k}\underset{m_n=1}{\overset{M}{\mathrm{\Σ}}}{c}_{m_n,m_k}{a}_{n,m_n}\frac{p_n}{d_{n,k}^{\mathrm{\α}}}+P_N}\≥\mathrm{\β},$

For all (k, m _k) ∈ J _i,

Wherein J _i'=J _i-1∪ (l, m _t), B _i'to belong to J _i'all base stations, i.e. B _i'=B _i-1∪ l, P _kthe through-put power of base station k, P _nnoise power, d _kthe distance between base station k and its terminal use, d _n,kit is the distance between base station n and any user of base station k; Wherein represent that the current base station-channel that will distribute is to (l, m _t) be subject to likely disturb summation, simultaneously be defined as:

When be 1 and show channel m _nwith channel m _kidentical or overlapping, be 0 o'clock channel m _nwith channel m _kbetween noiseless mutually.

Next explanation is introduced about the parameter declaration relating in simulation process of the present invention:

Base station number: N

Square region size: 1000 × 1000units:

The through-put power of each base station: P=100mW

The communication distance of each base station: r=50units,

A path loss factor: mistake! Do not find Reference source.

Noise power: P _n=-100dBm

SINR thresholding: 5dB and 10dB.

Can use CAB resource spectral bandwidth: 100MHZ as base station

Then carry out channel distribution according to the network type in Fig. 2.Each base station is being chosen after the network type that oneself will access, and carries out competitive bidding according to the competitive bidding scope of Fig. 2.

Next introduce the correlated performance index of simulation result:

System total revenue can be expressed as

$R=\underset{l=1}{\overset{L}{\mathrm{\Σ}}}\underset{m_t=1}{\overset{M}{\mathrm{\Σ}}}{a}_{l,m_t}{b}_{l,m_t}$

The cumulative distribution function of the SINR of end user location in management area (CDF, cumulative distribution function) is selected ten terminal uses that are positioned at the communication zone boundaries of base station at random, measures its SINR, and draws cumulative distribution function figure.

Fairness in distribution

Definition fairness criteria is:

$F_A={\left(\underset{t=1}{\overset{T}{\mathrm{\Σ}}}{Q}_t\right)}^2/T.\underset{t=1}{\overset{T}{\mathrm{\Σ}}}{Q}_t^2$

F _a=1 mean distribute finish after, all base stations have obtained the frequency spectrum resource of same amount of bandwidth.

Wherein Q _tbe the average bandwidth that base station obtains, and be defined as

$Q_t=w_t\frac{R_t}{N_t}$

R _tafter distribution finishes, the channel multiplexing total degree of network type t, N _tit is the number of base stations of selecting network type t.

Can provide service more flexibly with respect to prior art the present invention, we can obtain the allocation result of high fairness, low system benefit, also can obtain the high yield allocation result of relatively high fairness simultaneously, by utilizing physical disturbance model, effectively suppress the cumulative interference between base station, thereby ensured terminal use's reliable communication.

Claims (2)

1. a method for the cognitive radio frequency spectrum resource management based on fairness, is characterized in that: it comprises the following steps:

The first step: the frequency spectrum resource of underusing or not being assigned with is carried out to channel distribution according to different network type;

Second step: describe the parameter of the frequency spectrum resource of underusing or not being assigned with, the channel quantity of the total bandwidth W of the frequency spectrum resource of wherein underusing or not being assigned with, the channel width Wt of network type t and network type t wherein it is rounding operation;

The 3rd step: introduce three variable element a, b, c, increase the impact that a increases current obtainable channel width wt, and the larger system total revenue of acquisition, reduce fairness in distribution simultaneously, increase b and c and improve fairness in distribution, reduce system total revenue simultaneously, described the 3rd step also comprises, each base station is selected an available network model voluntarily, then the channel of map network type is carried out to competitive bidding, and base station 1 is to channel m _t(m _t=1 ..., M _t) carry out competitive bidding, the target function design in its i time circulation is as follows:

$\underset{(l,j)\∈J_i-1}{\underset{(n,j)\∈J_i-1}{\underset{(l,m_t)\∉J_i-1}{\max }}}\frac{b_{l,m_t}\·w_t^a}{{({\mathrm{\Σ}}_{n\∈B_{i-1}}\underset{j=1}{\overset{M_t}{\mathrm{\Σ}}}{w}_t\·a_{n,j})}^b\·{(\underset{j=1}{\overset{M_t}{\mathrm{\Σ}}}{w}_t\·a_{l,j})}^c}$

Wherein J _i-1until the BTS channel pair that the i-1 time circulation distributed after finishing b _i-1j _i-1in all base stations, that base station l is to channel m _tbidding price, w _tthe channel width of network type t, a _n,j(j=1 ..., M _t), (n, j) ∈ J _i-1i-1 the circulation allocation matrix of network type t afterwards, and a _{n, j}be defined as:

2. the method for a kind of cognitive radio frequency spectrum resource management based on fairness as claimed in claim 1, is characterized in that: further consider the cumulative interference between base station, as follows based on physical disturbance model optimization:

Constraints design is as follows:

Be constrained in $\frac{\frac{p_k}{d_k^{\mathrm{\α}}}}{{\mathrm{\Σ}}_{n\∈B_{i^{\′}}n\≠k}\underset{m_n=1}{\overset{M}{\mathrm{\Σ}}}{c}_{m_n,m_k}{a}_{n,m_n}\frac{p_n}{d_{n,k}^{\mathrm{\α}}}+P_N}\≥\mathrm{\β},$

For all (k, m _k) ∈ J _i'

Wherein J _i'=J _i-1∪ (l, m _t), B _i'to belong to J _i'all base stations, i.e. B _i'=B _i-1∪ l, P _kthe through-put power of base station k, P _nnoise power, d _kthe distance between base station k and its terminal use, d _n,kit is the distance between base station n and any user of base station k; Wherein represent that the current base station-channel that will distribute is to (l, m _t) be subject to likely disturb summation, simultaneously be defined as:

When be 1 and show channel m _nwith channel m _kidentical or overlapping, be 0 o'clock channel m _nwith channel m _kbetween noiseless mutually.