CN102427586B - Power and relay combination optimization method based on Fountain code and cooperation communication method thereof - Google Patents

Abstract

The invention discloses a power and relay combination optimization method based on a Fountain code. The method is characterized by: using the Fountain code to carry out coding transmission to a signal; based on an information accumulation characteristic of the Fountain code and an actual channel environment, constructing a power and relay combination optimization model; and under a condition of guaranteeing reliability of the information transmission, optimizing a validity index so as to realize a good compromise between the validity and the reliability of the information transmission; on the basis, solving an optimal solution of a theory of the combination optimization model as the basis of the cooperation communication. The invention also discloses a cooperation communication method based on the combination optimization method. The cooperation communication method of the power and relay combination optimization method based on the Fountain code of the invention is very simple and is easy to realize. An application prospect is good.

Description

Power based on the Fountain code and relaying combined optimization method and collaboration communication method thereof

Technical field

The invention belongs to wireless communication technology field, particularly power and relaying combined optimization method and the collaboration communication method based on this combined optimization method.

Background technology

Cooperative communication technology is to utilize internodal cooperation, shares antenna each other, forms virtual antenna array and obtains space diversity, thereby effectively resist the wireless channel decline, greatly improves the quality of transmission, attracts wide attention in recent years.Cooperative communication technology makes between a plurality of communication entities and improves ability to work and efficiency by cooperation, jointly completes communication task, to reaching the effect of " integral body is greater than the part sum ".According to the pass-through mode of via node, collaboration communication can be divided into following 5 classes: detection forwarding, amplification forwarding, decoding forward, coding forwards and compress forwarding etc.Above-mentioned all kinds show at aspects such as computation complexity, diversity gain, spatial multiplexing gain that each is variant, need to, according to different environment and the demand of application, select to adopt which kind of collaboration communication type.Certainly, collaboration communication, as a kind of emerging technology, still has many problems to have to be solved, and wherein the core difficult point is: (1) " with whose cooperation ", i.e. How to choose collaboration relay node; (2) " how to cooperate ", i.e. how distributing radio resource, particularly power resource.Therefore, emphasis and the core of following cooperative communication technology research the combined optimization research of power and relaying have been become.

At present, correlative study all proposes corresponding prioritization scheme for 3 coordination models (as shown in Figure 1) of classics mostly.According to the difference of optimizing index, the combined optimization scheme of power and relaying can be divided into prioritization scheme, the prioritization scheme based on capacity and the prioritization scheme based on bit error rate based on outage probability.Above scheme is all by the optimization to certain index, and expectation reaches the target of power optimized distribution and relaying optimal selection.Yet still there is following shortcoming in all kinds of prioritization schemes: (1) is difficult to realize the good compromise of validity and reliability, often considers wherein to ignore more in addition; (2) scheme had is easy to by the simplification to Optimized model and approximate obtaining the optimisation strategy realized, but has larger difference with theoretical model, optimization that also just can't realize target; (3) scheme had, by the calculating of Optimized model optimal solution, obtains corresponding optimisation strategy, but often too complicated and be difficult to realize.

Summary of the invention

Technical problem: the object of the invention is to provides a kind of new power based on the Fountain code and relaying combined optimization method for the deficiencies in the prior art, and a kind of collaboration communication method is provided on the basis of this combined optimization method.Optimization method of the present invention adopts the Fountain code to carry out coding transmission to signal, characteristic based on the Fountain code, under the reliability prerequisite of guarantee information transmission, its Validity Index of optimization (propagation delay time), thereby realize the good compromise of the validity and reliability of communication, can realize the optimization of optimization aim.Simultaneously, this scheme very simply is easy to again realize having good application prospect.

Technical scheme:

For achieving the above object, the present invention specifically adopts following technical scheme:

A kind of power and relaying combined optimization method based on the Fountain code, it is characterized in that adopting the Fountain code to carry out coding transmission to signal, information accumulation characteristics based on the Fountain code and actual channel environment construction power and relaying combined optimization model, thereby under the reliability prerequisite of guarantee information transmission, its Validity Index of optimization, good compromise with the validity and reliability of realizing communication, on this basis, solve the foundation of the theoretical optimal solution of combined optimization model as collaboration communication, described power and relaying combined optimization model are as follows:

$\{C^{*},P_{1S}^{*},P_{2S}^{*},P_{2C}^{*},{\mathrm{\Δ}}_1^{*},{\mathrm{\Δ}}_2^{*}\}=\underset{\{C/C\∈N\},\{P_{1S},P_{2S},P_{2C}\},\{{\mathrm{\Δ}}_i/1\≤i\≤2\}}{\arg \min }\underset{i=1}{\overset{2}{\mathrm{\Σ}}}{\mathrm{\Δ}}_i$

subject to 0≤P _1S≤P _max

0≤P _2S≤P _max (1)

P _1C＝0

0≤P _2C≤P _max

Δ _i≥0，1≤i≤2

$f_{\mathrm{SC}}^1\left(P_{1S}\right){\mathrm{\Δ}}_1\≥I$

$\underset{j=S,C}{\mathrm{\Σ}}\underset{i=1}{\overset{k}{\mathrm{\Σ}}}{f}_{\mathrm{jD}}^i\left(P_{\mathrm{ij}}\right){\mathrm{\Δ}}_i\≥I,1\≤k\≤2$

Wherein, Δ ₁Expression sends data to from source node S the time interval that via node C solves data, i.e. the 1st time slot; Δ ₂Expression solves data from via node C and solves the time interval of data to destination node D, i.e. the 2nd time slot, P _1S, P _2S, P _1C, P _2CMean that source node S and via node C are respectively in the transmitting power of 2 time slots, N means candidate relay set of node, P _maxThe transmitting power maximum that means arbitrary node,

Mean respectively variable C, P _1S, P _2S, P _2C, Δ ₁, Δ ₂Optimal solution, last 2 constraints separate provision of Optimized model node C and node D solve the required satisfied condition of data, wherein,

Be illustrated in the node S of time slot 1 and the channel transmission rate between node C,

Be illustrated in the node j of time slot i and the channel transmission rate between node D,

$f_{\mathrm{SC}}^1\left(P_{1S}\right)=W{\log }_2(1+\frac{h_{\mathrm{SC}}{P}_{1S}}{N_{C}W})$ 、 $f_{\mathrm{SD}}^1\left(P_{1S}\right)=W{\log }_2(1+\frac{h_{\mathrm{SD}}{P}_{1S}}{N_{D}W})$ 、 $f_{\mathrm{SD}}^2\left(P_{2S}\right)=W{\log }_2(1+\frac{h_{\mathrm{SD}}{P}_{2S}}{h_{\mathrm{CD}}{P}_{2C}{N}_{D}W})$ 、 $f_{\mathrm{CD}}^2\left(P_{2C}\right)=W{\log }_2(1+\frac{h_{\mathrm{CD}}{P}_{2C}}{h_{\mathrm{SD}}{P}_{2S}{N}_{D}W}),$ Here, W means signal bandwidth, N _CAnd N _DMean respectively to take the additive Gaussian noise power spectral density of the channel that node C and node D are receiving terminal.H _SC, h _SDAnd h _CDMean respectively channel l _SC, l _SDAnd l _CDChannel gain.

Collaboration communication method based on above-mentioned combined optimization method provided by the invention comprises following steps:

The first step: set up network topology, the initialization network environment, comprise source node S, destination node D, candidate relay set of node

N={C ₁, C ₂..., C _Γ, signal bandwidth W, data bit count I, transmitting power maximum P _max

Second step: source node S, by environment perception technology, obtains various channel condition informations, comprises channel gain h _SC, h _SDAnd h _CDChannel noise power spectrum density N _CAnd N _D, here,

The 3rd step: according to formula

$P_{1S}^{*}=P_{\max }$

${\mathrm{\Δ}}_1^{*}=\frac{I}{f_{\mathrm{SC}}^1\left(P_{1S}^{*}\right)}=\frac{I}{W{\log }_2(1+\frac{h_{\mathrm{SC}}{P}_{1S}^{*}}{N_{C}W})},---\left(4\right)$

With

$\max f_D^2(P_{2S},P_{2C})=$

$\left\{\begin{array}{cc}{f}_D^2(0,P_{\max }),& \mathrm{when}{h}_{\mathrm{CD}}{P}_{\max }\&GreaterEqual;\sqrt{N_{D}W(h_{\mathrm{SD}}{P}_{\max }+N_{D}W)}\&h_{\mathrm{CD}}>h_{\mathrm{SD}}\\ f_D^2(P_{\max },0)& \mathrm{when}{h}_{\mathrm{SD}}{P}_{\max }\&GreaterEqual;\sqrt{N_{D}W(h_{\mathrm{CD}}{P}_{\max }+N_{D}W)}\&h_{\mathrm{CD}}<h_{\mathrm{SD}}\\ f_D^2(0,P_{\max })=f_D^2(P_{\max },0),& \mathrm{when}{h}_{\mathrm{CD}}{P}_{\max }\&GreaterEqual;\sqrt{N_{D}W(h_{\mathrm{SD}}{P}_{\max }+N_{D}W)}\&h_{\mathrm{SD}}{P}_{\max }\&GreaterEqual;\sqrt{N_{D}W(h_{\mathrm{CD}}{P}_{\max }+N_{D}W)}\&h_{\mathrm{CD}}=h_{\mathrm{SD}}\\ f_D^2(P_{\max },P_{\max }),& \mathrm{when}{h}_{\mathrm{CD}}{P}_{\max }<\sqrt{N_{D}W(h_{\mathrm{SD}}{P}_{\max }+N_{D}W)}\&h_{\mathrm{SD}}{P}_{\max }<\sqrt{N_{D}W\left(h_{\mathrm{CD}}{P}_{\max }{+N}_{D}W\right)}\end{array}\right.---\left(23\right)$

Wherein, $f_D^2(P_{2S},P_{2C})=W{\log }_2(1+\frac{h_{\mathrm{SD}}{P}_{2S}}{h_{\mathrm{CD}}{P}_{2C}+N_{D}W})+W{\log }_2(1+\frac{h_{\mathrm{CD}}{P}_{2C}}{h_{\mathrm{SD}}{P}_{2S}+N_{D}W})$ Be illustrated in the total transmission rate of the 2nd time slot to node D, at first source node S calculates the power division that obtains optimum while selecting any via node C, at this, when

The time, the dump energy size of comparison source node S and via node C: if the source node S dump energy is larger, order Otherwise $\max f_D^2(P_{2S},P_{2C})=f_D^2(0,P_{\max });$

The 4th step: according to formula

$P_{1S}^{*}=P_{\max }$

${\mathrm{\&Delta;}}_1^{*}=\frac{I}{f_{\mathrm{SC}}^1\left(P_{1S}^{*}\right)}=\frac{I}{W{\log }_2(1+\frac{h_{\mathrm{SC}}{P}_{1S}^{*}}{N_{C}W})},---\left(4\right)$

With

${\mathrm{\&Delta;}}_2^{*}=\frac{{\mathrm{<figure-callout\; id="2"\; label="I\; D"\; filenames="HDA0000116512500000012.png"\; state="\{\{state\}\}">I</figure-callout>}}_D^{}}{\max f_D^2(P_{2S},P_{2C})}---\left(24\right)$

And

${\mathrm{\&Delta;}}_C^{*}={\mathrm{\&Delta;}}_1^{*}+{\mathrm{\&Delta;}}_2^{*}=\frac{I}{f_{\mathrm{SC}}^1\left(P_{1S}^{*}\right)}+\frac{{\mathrm{<figure-callout\; id="2"\; label="I\; D"\; filenames="HDA0000116512500000012.png"\; state="\{\{state\}\}">I</figure-callout>}}_D^{}}{\max f_D^2\left(P_{2S,}{P}_{2C}\right)}---\left(25\right)$

Wherein,

Be illustrated in second time slot node D decoding information needed amount,

Source node S is calculated the propagation delay time of optimum while selecting any via node C;

The 5th step: when C ∈ N, the more corresponding optimal transmission time delay of source node S, according to formula

$C^{*}=\underset{C\&Element;N}{\arg \min }{\mathrm{\&Delta;}}_C^{*}---\left(26\right)$

Obtain optimal relay node C ^*, wherein

At optimal relay node C ^*In situation, the optimal power allocation of trying to achieve and optimum propagation delay time are the theoretical optimal solution of combined optimization model.

The 6th step: source node S sends message to via node C ^*, notify this node to complete the collaboration communication process to destination node D according to optimum power distribution result.

Beneficial effect:

1) the present invention is based on the information accumulation characteristics of Fountain code and power and the relaying combined optimization model of actual channel environment construction, under the reliability prerequisite of guarantee information transmission, its Validity Index of optimization, can realize the good compromise of the validity and reliability of communication.

2) the present invention is resolved into 2 subproblems by the combined optimization model: relay selection problem and power division problem, at first fixed relay is selected problem, solves the power division problem, and then selects best relay, thereby obtain the theoretical optimal solution of Optimized model, simplified solution procedure.

3) the present invention is based on the power of Fountain code and the collaboration communication method of relaying combined optimization method very simply is easy to realize having good application prospect.

The accompanying drawing explanation

Fig. 1 is 3 collaboration communication model schematic diagrames.

Fig. 2 is function

The border schematic diagram.

Fig. 3 is power based on the Fountain code and the collaboration communication method flow chart of relaying combined optimization.

Embodiment

As shown in Figure 1, in 3 collaboration communication models, have three category nodes: source node S, destination node D and collaboration relay node C, source node S and via node C work in coordination with as destination node D transmission information.These 3 coordination models can be divided into 2 channels: from source node S to via node C with the broadcast channel of destination node D; Access channel from source node S and via node C to destination node D.Suppose that via node C adopts TDD mode, node C can send information or reception information, but can not send and receive information simultaneously.Based on this kind of supposition, the collaboration communication process of this model can be divided into 2 time slots: the 1st time slot S sends a message to C and D; Second time slot S and C cooperation send a message to D.

In this coordination model, how information is sent to destination node effectively reliably, be to need the key problem solved.For this reason, this patent will utilize the good characteristic of Fountain code, and by the optimum allocation of Internet resources, to solve an above-mentioned difficult problem.The Fountain code is a class rate-compatible code, has very little encoding and decoding complexity, comprises LT code and Raptor code etc.Utilize the Fountain code, transmitting terminal can be encoded into initial data the code stream of indefinite length, continuously coded message is sent to receiving terminal.If a plurality of transmitting terminals are arranged, and each transmitting terminal all adopts identical Fountain code, and receiving terminal will obtain the multiple copy of same code-word symbol, is similar to rake, i.e. energy accumulation.If each transmitting terminal adopts the Fountain code independently produced based on same initial data, receiving terminal has been accumulated the information of a plurality of transmitting terminals but not energy can be realized the accumulation of its mutual information.Now, the necessary and sufficient condition that success is decoded is: all (transmitting terminal) information sums that receiving terminal is received surpass the bit number I of initial data.Here, the size of I and the characteristic of channel are irrelevant.

In collaboration communication model of the present invention, establishing source node is S, and destination node is D, and optimal relay node is C ^*.Wherein, C ^*In Γ candidate relay node, select to obtain.Suppose that all nodes all adopt the Fountain code to carry out encoding and decoding, and source node and via node adopt identical frequency band to send information, there is the phase mutual interference in interchannel.In addition, each channel all is made as the desirable additive Gaussian channel of frequency-flat, and its transmission rate (capacity) can be obtained by classical shannon formula.Difficult, as long as the transmission rate between transmitting terminal and receiving terminal was not 0 (letter is dry is not tending towards 0 than SINR), pass through the regular hour, receiving terminal can be received abundant information bit, thereby realizes the decoding of initial data.Therefore, under the guaranteed prerequisite of reliability, how effectively (rapidly) transmission information becomes optimization aim of the present invention.Concrete Optimized model is expressed as follows:

$\{C^{*},P_{1S}^{*},P_{2S}^{*},P_{2C}^{*},{\mathrm{\&Delta;}}_1^{*},{\mathrm{\&Delta;}}_2^{*}\}=\underset{\{C/C\&Element;N\},\{P_{1S},P_{2S},P_{2C}\},\{{\mathrm{\&Delta;}}_i/1\&le;i\&le;2\}}{\arg \min }\underset{i=1}{\overset{2}{\mathrm{\&Sigma;}}}{\mathrm{\&Delta;}}_i$

subject to 0≤P _1S≤P _max

0≤P _2S≤P _max

P _1C＝0

0≤P _2C≤P _max

Δ _i≥0，1≤i≤2

$f_{\mathrm{SC}}^1\left(P_{1S}\right){\mathrm{\&Delta;}}_1\&GreaterEqual;I$

$\underset{j=S,C}{\mathrm{\&Sigma;}}\underset{i=1}{\overset{k}{\mathrm{\&Sigma;}}}{f}_{\mathrm{jD}}^i\left(P_{\mathrm{ij}}\right){\mathrm{\&Delta;}}_i\&GreaterEqual;I,1\&le;k\&le;2---\left(1\right)$

Wherein, Δ ₁Expression sends data to from source node S the time interval (the 1st time slot) that via node C solves data; Δ ₂Expression solves data from via node C and solves the time interval (the 2nd time slot) of data to destination node D.P _1S, P _2S, P _1C, P _2CMean that source node S and via node C are respectively in the transmitting power of 2 time slots.N means candidate relay set of node N={C ₁, C ₂..., C _Γ, P _maxThe transmitting power maximum that means arbitrary node,

Mean respectively variable C, P _1S, P _2S, P _2C, Δ ₁, Δ ₂Optimal solution.Last 2 constraints separate provision of Optimized model node C and node D solve the required satisfied condition of data.

Be illustrated in the node S of time slot 1 and the transmission capacity (speed) between node C,

Be illustrated in the node j of time slot i and the transmission capacity (speed) between node D.Be expressed as follows respectively:

$f_{\mathrm{SC}}^1\left(P_{1S}\right)=W{\log }_2(1+\frac{h_{\mathrm{SC}}{P}_{1S}}{N_{C}W})$

$f_{\mathrm{SD}}^1\left(P_{1S}\right)=W{\log }_2(1+\frac{h_{\mathrm{SD}}{P}_{1S}}{N_{D}W})$

$f_{\mathrm{CD}}^1\left(P_{1C}\right)=W{\log }_2(1+\frac{h_{\mathrm{CD}}{P}_{1C}}{N_{D}W})=0$

$f_{\mathrm{SD}}^2\left(P_{2S}\right)=W{\log }_2(1+\frac{h_{\mathrm{SD}}{P}_{2S}}{h_{\mathrm{CD}}{P}_{2C}{N}_{D}W})$

$f_{\mathrm{CD}}^2\left(P_{2C}\right)=W{\log }_2(1+\frac{h_{\mathrm{CD}}{P}_{2C}}{h_{\mathrm{SD}}{P}_{2S}{N}_{D}W})---\left(2\right)$

Wherein, W means the bandwidth of each node transmitted signal.Because the noise characteristic of channel depends primarily on receiving terminal, therefore make N _CAnd N _DMean respectively to take the additive Gaussian noise power spectral density of the channel that node C and node D are receiving terminal.H _SC, h _SDAnd h _CDMean respectively channel l _SC, l _SDAnd l _CDChannel gain.

Can find out, this Optimized model can resolve into 2 subproblems: relay selection problem and power division problem.For obtaining the optimal solution of Optimized model, at first fixed relay is selected problem (C is constant for the supposition via node), solves the power division problem.Optimized model can be reduced to:

$\{P_{1S}^{*},P_{2S}^{*},P_{2C}^{*},{\mathrm{\&Delta;}}_1^{*},{\mathrm{\&Delta;}}_2^{*}\}=\underset{\{P_{1S},P_{2S},P_{2C}\},\{{\mathrm{\&Delta;}}_i/1\&le;i\&le;2\}}{\arg \min }\underset{i=1}{\overset{2}{\mathrm{\&Sigma;}}}{\mathrm{\&Delta;}}_i---\left(3\right)$

Wherein constraint is same as formula (1).Above-mentioned optimizing process is actual to be divided into 2 time slots and to carry out, and the optimum results of 2 time slots has correlation, and the optimum results of the 2nd time slot is subject to the impact of the optimum results of the 1st time slot.Therefore,

But, still can obtain by the method for analyzing the optimal solution of the 1st time slot: in the 1st time slot, for making Δ ₁Minimize, need to maximize transmission rate

Thereby can obtain P _1S=P _maxFor making Δ ₂Minimize node D decoding information needed amount in the time of need to minimizing the 2nd time slot and start

Still can obtain P _1S=P _max.Therefore, the optimal solution of the 1st time slot is:

$P_{1S}^{*}=P_{\max }$

${\mathrm{\&Delta;}}_1^{*}=\frac{I}{f_{\mathrm{SC}}^1\left(P_{1S}^{*}\right)}=\frac{I}{W{\log }_2(1+\frac{h_{\mathrm{SC}}{P}_{1S}^{*}}{N_{C}W})}---\left(4\right)$

On the basis that obtains the 1st time slot optimal solution, by the Optimized model abbreviation and be equivalent to and maximize the overall transmission rate function, and then analyze it and be worth most a little and extreme point, can not lead a little and the relation between boundary point, finally by function, the value at boundary point merges by comparison and parameter set, thereby obtains the optimal solution of the 2nd time slot.

When the 2nd time slot starts, the node D required minimal information amount of decoding is:

${\mathrm{<figure-callout\; id="2"\; label="I\; D"\; filenames="HDA0000116512500000012.png"\; state="\{\{state\}\}">I</figure-callout>}}_D^{}=I-f_{\mathrm{SD}}^1\left(P_{1S}^{*}\right){\mathrm{\&Delta;}}_1^{*}=I(1-\frac{W{\log }_2(1+\frac{h_{\mathrm{SD}}{P}_{1S}^{*}}{N_{D}W})}{W{\log }_2(1+\frac{h_{\mathrm{SC}}{P}_{1S}^{*}}{N_{C}W})})---\left(5\right)$

At the 2nd time slot, total transmission rate that node S and node C send information to node D is:

$f_D^2(P_{2S},P_{2C})=f_{\mathrm{SD}}^2\left(P_{2S}\right)+f_{\mathrm{CD}}^2\left(P_{2C}\right)=W{\log }_2(1+\frac{h_{\mathrm{SD}}{P}_{2S}}{h_{\mathrm{CD}}{P}_{2C}+N_{D}W})+W{\log }_2(1+\frac{h_{\mathrm{CD}}{P}_{2C}}{h_{\mathrm{SD}}{P}_{2S}+N_{D}W})---\left(6\right)$

Therefore, the Optimized model of the 2nd time slot can be expressed as:

$\{P_{2S}^{*},P_{2C}^{*},{\mathrm{\&Delta;}}_2^{*}\}=\underset{\{P_{2S},P_{2C}\},\left\{{\mathrm{\&Delta;}}_2\right\}}{\arg \min }{\mathrm{\&Delta;}}_2$

$=\underset{\{P_{2S},P_{2C}\},\left\{{\mathrm{\&Delta;}}_2\right\}}{\arg \min }\frac{I_D^2}{f_D^2(P_{2S},P_{2C})}---\left(7\right)$

Wherein,

In I, W, N _DAnd h _SDBe known parameters, N _CAnd h _SCAlso suppose constant (via node C supposition is constant), and

It is also known parameters.Therefore,

Can regard fixed amount as.Therefore, the 2nd time slot Optimized model (formula

(7)) can equivalence be converted into:

$\{P_{2S}^{*},P_{2C}^{*},{\mathrm{\&Delta;}}_2^{*}\}=\underset{\{P_{2S},P_{2C}\},\left\{{\mathrm{\&Delta;}}_2\right\}}{\arg \min }{f}_D^2(P_{2S},P_{2C})$

$=\underset{\{P_{2S},P_{2C}\},\left\{{\mathrm{\&Delta;}}_2\right\}}{\arg \max }(W{\log }_2(1+\frac{h_{\mathrm{SD}}{P}_{2S}}{h_{\mathrm{CD}}{P}_{2C}+N_{D}W})+W{\log }_2(1+\frac{h_{\mathrm{CD}}{P}_{2C}}{h_{\mathrm{SD}}{P}_{2S}+N_{D}W}))$

subject to 0≤P _2S≤P _max (8)

0≤P _2C≤P _max

${\mathrm{\&Delta;}}_2=\frac{I_D^2}{f_D^2(P_{2S},P_{2C})}\&GreaterEqual;0$

Can find out, how solved function Maximum become the very corn of a subject.Typically, continuously the value point of bounded function is present in extreme point, can not leads a little and among boundary point.At first, analyze

Value whether be present in its extreme point, utilize method of derivation to obtain its all extreme points:

$\frac{\&PartialD;f_D^2(P_{2S},P_{2C})}{\&PartialD;P_{2S}}=\frac{W}{\ln 2}\frac{h_{\mathrm{SD}}\&CenterDot;(h_{\mathrm{SD}}{P}_{2S}+N_{D}W-h_{\mathrm{CD}}{P}_{2C})}{(h_{\mathrm{SD}}{P}_{2S}+h_{\mathrm{CD}}{P}_{2C}+N_{D}W)\&CenterDot;(h_{\mathrm{SD}}{P}_{2S}+N_{D}W)}=0$ (9)

$\frac{\&PartialD;f_D^2(P_{2S},P_{2C})}{\&PartialD;P_{2C}}=\frac{W}{\ln 2}\frac{h_{\mathrm{CD}}\&CenterDot;(h_{\mathrm{CD}}{P}_{2C}+N_{D}W-h_{\mathrm{SD}}{P}_{2S})}{(h_{\mathrm{SD}}{P}_{2S}+h_{\mathrm{CD}}{P}_{2C}+N_{D}W)\&CenterDot;(h_{\mathrm{CD}}{P}_{2C}+N_{D}W)}=0$

Can be obtained by formula (9), work as N _DW=0 and h _SDP _2S=h _CDP _2CThe time, function

There is extreme point, and be infinite a plurality of minimum point (proof slightly).Certainly, according to N _DThe physical significance of W (being the noise power of node D), known N _DW>0.Therefore, under actual environment,

There do not is extreme point.In addition, due to N _DW>0, known in conjunction with formula (9),

The single order partial derivative at variable (P _2S, P _2C) whole span in all exist, thereby do not exist and can not lead a little.

Therefore, known by above-mentioned analysis, function

Value should be present among its boundary point.As shown in Figure 2, this function has 4 borders (Boundary), and mathematical notation is as follows:

Boundary 1：P _2S＝0and 0≤P _2C≤P _max

Boundary 2：P _2S＝P _max and 0≤P _2C≤P _max (10)

Boundary 3：P _2C＝0and 0≤P _2S≤P _max

Boundary 4：P _2C＝P _maxand 0≤P _2S≤P _max

Because the value of function is present in above-mentioned 4 borders, below at first ask the maximum of function on every border, then relatively try to achieve Function Extreme value.For this reason, first ask function

Maximum on border 1 (Boundary1).Now, corresponding optimization problem is expressed as follows:

$\left\{P_{2C}^{*}\right\}=\arg {\max f}_D^2(0,P_{2C})$

$=\arg \max {W\log }_2(1+\frac{h_{\mathrm{CD}}{P}_{2C}}{N_{D}W})---\left(11\right)$

subject to 0≤P _2C≤P _max

Obviously, function

Variable P _2CMonotonically increasing function, thereby can obtain function

Maximum on Boundary 1:

$\max f_D^2(0,P_{2C})=f_D^2(0,P_{\max })={W\log }_2(1+\frac{h_{\mathrm{CD}}{P}_{\max }}{N_{D}W})---\left(12\right)$

Ask again function

Maximum on Boundary 2.Now, optimization problem is expressed as:

$\left\{P_{2C}^{*}\right\}=\arg {\max f}_D^2{(P}_{\max },P_{2C})$

$=\arg \max (W{\log }_2(1+\frac{h_{\mathrm{SD}}{P}_{\max }}{h_{\mathrm{CD}}{P}_{2C}+N_{D}W})+W{\log }_2(1+\frac{h_{\mathrm{CD}}{P}_{2C}}{h_{\mathrm{SD}}{P}_{\max }+N_{D}W}))---\left(13\right)$

subject to 0≤P _2C≤P _max

Right Differentiate, can obtain $\frac{\&PartialD;f_D^2(P_{\max },P_{2C})}{\&PartialD;P_{2C}}=\frac{W}{\ln 2}\frac{h_{\mathrm{CD}}\&CenterDot;(h_{\mathrm{CD}}{P}_{2C}+N_{D}W-h_{\mathrm{SD}}{P}_{\max })}{(h_{\mathrm{SD}}{P}_{\max }+h_{\mathrm{CD}}{P}_{2C}+N_{D}W)\&CenterDot;(h_{\mathrm{CD}}{P}_{2C}+N_{D}W)},$ And then try to achieve its stationary point

Due to 0≤P _2C≤ P _max, therefore only work as N _DW≤h _SDP _max≤ h _CDP _max+ N _DDuring W, just there is above-mentioned stationary point.And,

To function

The second order differentiate, can obtain $\frac{d^2{f}_D^2(P_{\max },P_{2C})}{d{\left(P_{2C}\right)}^2}>0.$ Hence one can see that, $[0,\frac{h_{\mathrm{SD}}{P}_{\max }-N_{D}W}{h_{\mathrm{CD}}}]$ With $(\frac{h_{\mathrm{SD}}{P}_{\max }-N_{D}W}{h_{\mathrm{CD}}},P_{\max }]$ Be respectively successively decreasing of function interval and increase progressively interval, the maximum in 2 intervals is respectively:

$f_D^2(P_{\max },P_{\max })=W{\log }_2(1+\frac{h_{\mathrm{SD}}{P}_{\max }}{h_{\mathrm{CD}}{P}_{\max }+N_{D}W})+W{\log }_2(1+\frac{h_{\mathrm{CD}}{P}_{\max }}{h_{\mathrm{SD}}{P}_{\max }+N_{D}W}).$ The maximum in 2 intervals relatively, can obtain the maximum in whole interval.By can be calculated,

$\max f_D^2(P_{\max },P_{2C})=\left\{\begin{array}{cc}{f}_D^2(P_{\max },0),& \mathrm{when}\sqrt{N_{D}W(h_{\mathrm{CD}}{P}_{\max }+N_{D}W)}<h_{\mathrm{SD}}{P}_{\max }\&le;h_{\mathrm{CD}}{P}_{\max }+N_{D}W\\ f_D^2(P_{\max },0)=f_D^2(P_{\max },P_{\max },)& \mathrm{when}{h}_{\mathrm{SD}}{P}_{\max }=\sqrt{N_{D}W(h_{\mathrm{CD}}{P}_{\max }+N_{D}W)}\\ f_D^2(P_{\max },P_{\max }),& \mathrm{when}{N}_{D}W\&le;h_{\mathrm{SD}}{P}_{\max }<\sqrt{N_{D}W(h_{\mathrm{CD}}{P}_{\max }+N_{D}W)}\end{array}\right.---\left(14\right)$

Work as h _SDP _max<N _DW or h _SDP _max>h _CDP _max+ N _DDuring W, function

At [0, P _max] all there do not is stationary point, now Be respectively monotonically increasing function and monotonic decreasing function (symbol by the test function first derivative can be demonstrate,proved, and omits) herein.Can obtain thus,

$\max f_D^2(P_{\max },P_{2C})=\left\{\begin{array}{cc}{f}_D^2(P_{\max },0),& \mathrm{when}{h}_{\mathrm{SD}}{P}_{\max }>h_{\mathrm{CD}}{P}_{\max }+N_{D}W\\ f_D^2(P_{\max },P_{\max }),& \mathrm{when}{h}_{\mathrm{SD}}{P}_{\max }<N_{D}W\end{array}\right.---\left(15\right)$

Result to formula (14) and formula (15) is merged, and in the situation that the function maximum is identical that the parameter interval is merged, thereby obtains

Maximum on Boundary 2:

$\max f_D^2(P_{\max },P_{2C})=\left\{\begin{array}{cc}{f}_D^2(P_{\max },0),& \mathrm{when}{h}_{\mathrm{SD}}{P}_{\max }>\sqrt{N_{D}W(h_{\mathrm{CD}}{P}_{\max }+N_{D}W)}\\ f_D^2(P_{\max },0)=f_D^2(P_{\max },P_{\max },)& \mathrm{when}{h}_{\mathrm{SD}}{P}_{\max }=\sqrt{N_{D}W(h_{\mathrm{CD}}{P}_{\max }+N_{D}W)}\\ f_D^2(P_{\max },P_{\max }),& \mathrm{when}{h}_{\mathrm{SD}}{P}_{\max }<\sqrt{N_{D}W(h_{\mathrm{CD}}{P}_{\max }+N_{D}W)}\end{array}\right.---\left(16\right)$

From the angle of saving power, consider, when maximum is

The time, make maximum be only

Therefore, formula (16) can be reduced to:

$\max f_D^2(P_{\max },P_{2C})=\left\{\begin{array}{cc}{f}_D^2(P_{\max },0),& \mathrm{when}{h}_{\mathrm{SD}}{P}_{\max }\&GreaterEqual;\sqrt{N_{D}W{(h}_{\mathrm{CD}}{P}_{\max }+N_{D}W)}\\ f_D^2(P_{\max },P_{\max }),& \mathrm{when}{h}_{\mathrm{SD}}{P}_{\max }<\sqrt{N_{D}W(h_{\mathrm{CD}}{P}_{\max }+N_{D}W)}\end{array}\right.---\left(17\right)$

To the derivation be similar at Boundary 1, can obtain function

Maximum on Boundary 3:

$\max f_D^2(P_{2S},0)=f_D^2(P_{\max },0)W{\log }_2(1+\frac{h_{\mathrm{SD}}{P}_{\max }}{N_{D}W})---\left(\text{18}\right)$

Be similar to the derivation at Boundary 2, can obtain function

Maximum on Boundary 4:

$\max f_D^2(P_{2S},P_{\max })=\left\{\begin{array}{cc}{f}_D^2(0,P_{\max }),& \mathrm{when}{h}_{\mathrm{CD}}{P}_{\max }\&GreaterEqual;\sqrt{N_{D}W{(h}_{\mathrm{SD}}{P}_{\max }+N_{D}W)}\\ f_D^2(P_{\max },P_{\max }),& \mathrm{when}{h}_{\mathrm{CD}}{P}_{\max }<\sqrt{N_{D}W(h_{\mathrm{SD}}{P}_{\max }+N_{D}W)}\end{array}\right.---\left(19\right)$

According to formula (12) and (19), can obtain the maximum of function on Boundary 1 and Boundary 4:

$\max \{f_D^2(0,P_{2C}),f_D^2(P_{2S},P_{\max })\}=\left\{\begin{array}{cc}{f}_D^2(0,P_{\max }),& \mathrm{when}{h}_{\mathrm{CD}}{P}_{\max }\&GreaterEqual;\sqrt{N_{D}W{(h}_{\mathrm{SD}}{P}_{\max }+N_{D}W)}\\ f_D^2(P_{\max },P_{\max }),& \mathrm{when}{h}_{\mathrm{CD}}{P}_{\max }<\sqrt{N_{D}W(h_{\mathrm{SD}}{P}_{\max }+N_{D}W)}\end{array}\right.---\left(20\right)$

According to formula (17) and (18), can obtain the maximum of function on Boundary 2 and Boundary 3:

$\max \{f_D^2(P_{\max },P_{2C}),=f_D^2(P_{2D},0)\}=\left\{\begin{array}{cc}{f}_D^2(P_{\max },0),& \mathrm{when}{h}_{\mathrm{SD}}{P}_{\max }\&GreaterEqual;\sqrt{N_{D}W{(h}_{\mathrm{CD}}{P}_{\max }+N_{D}W)}\\ f_D^2(P_{\max },P_{\max }),& \mathrm{when}{h}_{\mathrm{SD}}{P}_{\max }<\sqrt{N_{D}W(h_{\mathrm{CD}}{P}_{\max }+N_{D}W)}\end{array}\right.---\left(21\right)$

In conjunction with formula (20) and formula (21), can obtain function

Maximum:

$\max f_D^2(P_{2S},P_{2C})=$

$\left\{\begin{array}{cc}{f}_D^2(0,P_{\max }),& \mathrm{when}{h}_{\mathrm{CD}}{P}_{\max }\&GreaterEqual;\sqrt{N_{D}W(h_{\mathrm{SD}}{P}_{\max }+N_{D}W)}\&h_{\mathrm{SD}}{P}_{\max }\&GreaterEqual;\sqrt{N_{D}W(h_{\mathrm{CD}}{P}_{\max }+N_{D}W)}\&h_{\mathrm{CD}}>h_{\mathrm{SD}}\\ f_D^2(P_{\max },0),& \mathrm{when}{h}_{\mathrm{CD}}{P}_{\mathrm{amx}}\&GreaterEqual;\sqrt{N_{D}W(h_{\mathrm{SD}}{P}_{\max }+N_{D}W)}\&h_{\mathrm{SD}}{P}_{\max }\&GreaterEqual;\sqrt{N_{D}W(h_{\mathrm{CD}}{P}_{\max }+N_{D}W)}\&h_{\mathrm{CD}}<h_{\mathrm{SD}}\\ f_D^2\left({0,P}_{\max }\right)=f_D^2(P_{\max },0),& \mathrm{when}{h}_{\mathrm{CD}}{P}_{\max }\&GreaterEqual;\sqrt{N_{D}W(h_{\mathrm{SD}}{P}_{\max }+N_{D}W)}\&h_{\mathrm{SD}}{P}_{\max }\&GreaterEqual;\sqrt{N_{D}W(h_{\mathrm{CD}}{P}_{\max }+N_{D}W)}\&h_{\mathrm{CD}}=h_{\mathrm{SD}}\\ f_D^2\left({0,P}_{\max }\right),& \mathrm{when}{h}_{\mathrm{CD}}{P}_{\max }\&GreaterEqual;\sqrt{N_{D}W(h_{\mathrm{SD}}{P}_{\max }+N_{D}W)}\&h_{\mathrm{SD}}{P}_{\max }<\sqrt{N_{D}W(h_{\mathrm{CD}}{P}_{\max }+N_{D}W)}\\ f_D^2(P_{\max },0),& \mathrm{when}{h}_{\mathrm{CD}}{P}_{\max }<\sqrt{N_{D}W(h_{\mathrm{SD}}{P}_{\max }+N_{D}W)}\&h_{\mathrm{SD}}{P}_{\max }\&GreaterEqual;\sqrt{N_{D}W(h_{\mathrm{CD}}{P}_{\max }+N_{D}W)}\\ f_D^2(P_{\max },P_{\max }),& \mathrm{when}{h}_{\mathrm{CD}}{P}_{\max }<\sqrt{N_{D}W(h_{\mathrm{SD}}{P}_{\max }+N_{D}W)}\&h_{\mathrm{SD}}{P}_{\max }<\sqrt{N_{D}W(h_{\mathrm{CD}}{P}_{\mathrm{amx}}+N_{D}W)}\end{array}\right.$

$\left(22\right)$

Fu Hao &amp wherein; Mean " getting union ".To carrying out abbreviation (merging the parameter interval in the identical situation of maximum) in formula (22), can obtain:

$\max f_D^2(P_{2S},P_{2C})=$

$\left\{\begin{array}{cc}{f}_D^2(0,P_{\max }),& \mathrm{when}{h}_{\mathrm{CD}}{P}_{\max }\&GreaterEqual;\sqrt{N_{D}W(h_{\mathrm{SD}}{P}_{\max }+N_{D}W)}\&h_{\mathrm{CD}}>h_{\mathrm{SD}}\\ f_D^2(P_{\max },0)& \mathrm{when}{h}_{\mathrm{SD}}{P}_{\max }\&GreaterEqual;\sqrt{N_{D}W(h_{\mathrm{CD}}{P}_{\max }+N_{D}W)}\&h_{\mathrm{CD}}<h_{\mathrm{SD}}\\ f_D^2(0,P_{\max })=f_D^2(P_{\max },0),& \mathrm{when}{h}_{\mathrm{CD}}{P}_{\max }\&GreaterEqual;\sqrt{N_{D}W(h_{\mathrm{SD}}{P}_{\max }+N_{D}W)}\&h_{\mathrm{SD}}{P}_{\max }\&GreaterEqual;\sqrt{N_{D}W(h_{\mathrm{CD}}{P}_{\max }+N_{D}W)}\&h_{\mathrm{CD}}=h_{\mathrm{SD}}\\ f_D^2(P_{\max },P_{\max }),& \mathrm{when}{h}_{\mathrm{CD}}{P}_{\max }<\sqrt{N_{D}W(h_{\mathrm{SD}}{P}_{\max }+N_{D}W)}\&h_{\mathrm{SD}}{P}_{\max }<\sqrt{N_{D}W\left(h_{\mathrm{CD}}{P}_{\max }{+N}_{D}W\right)}\end{array}\right.$

(23)

Can obtain optimum Δ by formula (5) and formula (23) ₂:

${\mathrm{\&Delta;}}_2^{*}=\frac{{\mathrm{<figure-callout\; id="2"\; label="I\; D"\; filenames="HDA0000116512500000012.png"\; state="\{\{state\}\}">I</figure-callout>}}_D^{}}{\max f_D^2(P_{2S},P_{2C})}---\left(24\right)$

From formula (4) and formula (24), in the situation that via node C is fixing, optimum propagation delay time can be obtained by following formula:

${\mathrm{\&Delta;}}_C^{*}={\mathrm{\&Delta;}}_1^{*}+{\mathrm{\&Delta;}}_2^{*}=\frac{I}{f_{\mathrm{SC}}^1\left(P_{1S}^{*}\right)}+\frac{{\mathrm{<figure-callout\; id="2"\; label="I\; D"\; filenames="HDA0000116512500000012.png"\; state="\{\{state\}\}">I</figure-callout>}}_D^{}}{\max f_D^2\left(P_{2S,}{P}_{2C}\right)}---\left(25\right)$

Therefore, relatively Γ the corresponding optimal transmission time delay of via node in candidate relay set of node N, can obtain optimal relay node C ^*:

$C^{*}=\underset{C\&Element;N}{\arg \min }{\mathrm{\&Delta;}}_C^{*}---\left(26\right)$

Corresponding to optimal relay node C ^*Propagation delay time and the power division optimal solution that is Optimized model (formula (1)).

The present invention is based on the collaboration communication method flow chart of the power of Fountain code and relaying combined optimization as shown in Figure 3.

The first step: set up network topology, the initialization network environment is (as source node S, destination node D, candidate relay set of node N={C ₁, C ₂..., C _Γ, signal bandwidth W, data bit count I, transmitting power maximum P _max).

Second step: source node S, by environment perception technology, obtains various channel condition informations (CSI): channel gain h _SC, h _SDAnd h _CDChannel noise power spectrum density N _CAnd N _D.Here,

The 3rd step: according to formula (4) and formula (23), at first source node S calculates the power division that obtains optimum while selecting any via node C.Need to illustrate, when The time, the dump energy size of comparison source node S and via node C: if the source node S dump energy is larger, order

Otherwise $\max f_D^2(P_{2S},P_{2C})=f_D^2(0,P_{\max }).$

The 4th step: source node S, according to formula (4), formula (24) and formula (25), is calculated the propagation delay time of optimum while selecting any via node C.

The 5th step: when C ∈ N, the more corresponding optimal transmission time delay of source node S, obtain optimal relay node C according to formula (26) ^*.At optimal relay node C ^*In situation, the optimal power allocation of trying to achieve and optimum propagation delay time are the theoretical optimal solution of combined optimization model.

The 6th step: source node S sends message to via node C ^*, notify this node to complete the collaboration communication process to destination node D according to optimum power distribution result.

Above a kind of collaboration communication scheme that the embodiment of the present invention is provided is described in detail, for one of ordinary skill in the art, thought according to the embodiment of the present invention, all will change in specific embodiments and applications, in sum, this description should not be construed as limitation of the present invention.

Claims (2)

1. power and the relaying combined optimization method based on the Fountain code, it is characterized in that adopting the Fountain code to carry out coding transmission to signal, information accumulation characteristics based on the Fountain code and actual channel environment construction power and relaying combined optimization model, thereby under the reliability prerequisite of guarantee information transmission, its Validity Index of optimization, good compromise with the validity and reliability of realizing communication, on this basis, solve the foundation of the theoretical optimal solution of combined optimization model as collaboration communication, described power and relaying combined optimization model are as follows:

$\{C^{*},P_{1S}^{*},P_{2S}^{*},P_{2C}^{*},{\mathrm{\&Delta;}}_1^{*},{\mathrm{\&Delta;}}_2^{*}\}=\underset{\{C/C\&Element;N\},\{P_{1S},P_{2S},P_{2C}\},\{{\mathrm{\&Delta;}}_i/1\&le;i\&le;2\}}{\arg \min }\underset{i=1}{\overset{2}{\mathrm{\&Sigma;}}}{\mathrm{\&Delta;}}_i$

subject to0≤P _1S≤P _max

0≤P _2S≤P _max （1）

P _1C＝0

0≤P _2C≤P _max

Δ _i≥0,1≤i≤2

$f_{\mathrm{SC}}^1\left(P_{1S}\right){\mathrm{\&Delta;}}_1\&GreaterEqual;I$

$\underset{j=S,C}{\mathrm{\&Sigma;}}\underset{i=1}{\overset{k}{\mathrm{\&Sigma;}}}{f}_{\mathrm{jD}}^i\left(P_{\mathrm{ij}}\right){\mathrm{\&Delta;}}_i\&GreaterEqual;I,1\&le;k\&le;2$

Wherein, Δ ₁Expression sends data to from source node S the time interval that via node C solves data, i.e. the 1st time slot; Δ ₂Expression solves data from via node C and solves the time interval of data to destination node D, i.e. the 2nd time slot, P _1S, P _2S, P _1C, P _2CMean that source node S and via node C are respectively in the transmitting power of 2 time slots, Ν means candidate relay set of node, P _maxThe transmitting power maximum that means arbitrary node, C ^*,

Mean respectively variable C, P _1S, P _2S, P _2C, Δ ₁, Δ ₂Optimal solution, last 2 constraints separate provision of Optimized model node C and node D solve the required satisfied condition of data, wherein, Be illustrated in the node S of time slot 1 and the channel transmission rate between node C, Ι is the data bit number,

J=S, C, be illustrated in the node j of time slot i and the channel transmission rate between node D,

$f_{\mathrm{SC}}^1\left(P_{1S}\right)=W{\log }_2(1+\frac{h_{\mathrm{SC}}{P}_{1S}}{N_{C}W}),f_{\mathrm{SD}}^1\left(P_{1S}\right)=W{\log }_2(1+\frac{h_{\mathrm{SD}}{P}_{1S}}{N_{D}W}),f_{\mathrm{SD}}^2\left(P_{2S}\right)=W{\log }_2(1+\frac{h_{\mathrm{SD}}{P}_{2S}}{h_{\mathrm{CD}}{P}_{2C}+N_{D}W}),$ $f_{\mathrm{CD}}^2\left(P_{2C}\right)={W\log }_2(1+\frac{h_{\mathrm{CD}}{P}_{2C}}{h_{\mathrm{SD}}{P}_{2S}+N_{D}W}),$ Here, W means signal bandwidth, N _CAnd N _DMean respectively to take the additive Gaussian noise power spectral density of the channel that node C and node D are receiving terminal, h _SC, h _SDAnd h _CDMean respectively channel l _SC, l _SDAnd l _CDChannel gain;

The method that solves the theoretical optimal solution of combined optimization model is, the combined optimization model is resolved into to 2 subproblems: relay selection problem and power division problem, at first fixed relay is selected problem, the associating Optimized model is carried out to abbreviation and solve its power division problem, and then solve the relay selection problem, thereby obtain the theoretical optimal solution of combined optimization model, the method that solves the power division problem is as follows:

Fixed relay is selected problem, by the Optimized model abbreviation, is:

$\{P_{1S}^{*},P_{2S}^{*},P_{2C}^{*},{\mathrm{\&Delta;}}_1^{*},{\mathrm{\&Delta;}}_2^{*}\}=\underset{\{P_{1S},P_{2S},P_{2C}\},\{{\mathrm{\&Delta;}}_i/1\&le;i\&le;2\}}{\arg \min }\underset{i=1}{\overset{2}{\mathrm{\&Sigma;}}}{\mathrm{\&Delta;}}_i,---\left(3\right)$

Wherein constraints is same as formula (1),

Optimizing process is divided into to 2 time slots, at first by theoretical analysis method, obtains the optimal solution of the 1st time slot;

$P_{1S}^{*}=P_{\max }$

${\mathrm{\&Delta;}}_1^{*}=\frac{I}{f_{\mathrm{SC}}^1\left(P_{1S}^{*}\right)}=\frac{I}{W{\log }_2(1+\frac{h_{\mathrm{SC}}{P}_{1S}^{*}}{N_{C}W})},---\left(4\right)$

On this basis, obtain the Optimized model of the 2nd time slot, its equivalence be converted into:

$\{P_{2S}^{*},P_{2C}^{*},{\mathrm{\&Delta;}}_2^{*}\}=\underset{\{P_{2S},P_{2C}\},\left\{{\mathrm{\&Delta;}}_2\right\}}{\arg \max }{f}_D^2(P_{2S},P_{2C})$

$=\underset{\{P_{2S},P_{2C}\},\left\{{\mathrm{\&Delta;}}_2\right\}}{\arg \max }({W\log }_2(1+\frac{h_{\mathrm{SD}}{P}_{2S}}{h_{\mathrm{CD}}{P}_{2C}+N_{D}W})+{W\log }_2(1+\frac{h_{\mathrm{CD}}{P}_{2C}}{h_{\mathrm{SD}}{P}_{2S}+N_{D}W}))$

subject to0≤P _2S≤P _max

0≤P _2C≤P _max

${\mathrm{\&Delta;}}_2=\frac{I_D^2}{f_D^2(P_{2S},P_{2C})}\&GreaterEqual;0---\left(8\right)$

Wherein,

Be illustrated in second time slot node D decoding information needed amount,

$f_D^2(P_{2S},P_{2C})={W\log }_2(1+\frac{h_{\mathrm{SD}}{P}_{2S}}{h_{\mathrm{CD}}{P}_{2C}+N_{D}W})+{W\log }_2(1+\frac{h_{\mathrm{CD}}{P}_{2C}}{h_{\mathrm{SD}}{P}_{2S}+N_{D}W})$ Be illustrated in the total transmission rate of the 2nd time slot to node D,

Thereby solve the optimal solution that obtains the 2nd time slot: ${\mathrm{\&Delta;}}_2^{*}=\frac{I_D^2}{\max f_D^2(P_{2S},P_{2C})},---\left(24\right)$

Wherein

Solve on power division problem basis, solve the relay selection problem, thereby obtain the theoretical optimal solution of combined optimization model, concrete grammar is: compare the corresponding optimal transmission time delay of each via node in candidate relay set of node Ν, obtain optimal relay node C ^*: $C^{*}=\underset{C\&Element;N}{\arg \min }{\mathrm{\&Delta;}}_C^{*},$ ${\mathrm{\&Delta;}}_C^{*}={\mathrm{\&Delta;}}_1^{*}+{\mathrm{\&Delta;}}_2^{*}$ Mean the optimal transmission time delay, ${\mathrm{\&Delta;}}_2^{*}=\frac{I_D^2}{\max f_D^2(P_{2S},P_{2C})}$ The optimal solution that means the 2nd time slot, therefore, optimal relay node C ^*Be the theoretical optimal solution of combined optimization model with corresponding optimal power allocation.

2. the collaboration communication method based on the described combined optimization method of claim 1 is characterized in that comprising following steps:

The first step: set up network topology, the initialization network environment, comprise source node S, destination node D, candidate relay set of node Ν={ C ₁, C ₂..., C _Γ, signal bandwidth W, data bit count Ι, transmitting power maximum P _max

Second step: source node S, by environment perception technology, obtains various channel condition informations, comprises channel gain h _SC, h _SDAnd h _CDChannel noise power spectrum density N _CAnd N _D, here, C ∈ Ν;

The 3rd step: according to formula

$P_{1S}^{*}=P_{\max }$

${\mathrm{\&Delta;}}_1^{*}=\frac{I}{f_{\mathrm{SC}}^1\left(P_{1S}^{*}\right)}=\frac{I}{{W\log }_2(1+\frac{h_{\mathrm{SC}}{P}_{1S}^{*}}{N_{C}W})},---\left(4\right)$

With

Wherein, $f_D^2(P_{2S},P_{2C})={W\log }_2(1+\frac{h_{\mathrm{SD}}{P}_{2S}}{h_{\mathrm{CD}}{P}_{2C}+N_{D}W})+{W\log }_2(1+\frac{h_{\mathrm{CD}}{P}_{2C}}{h_{\mathrm{SD}}{P}_{2S}+N_{D}W})$ Be illustrated in the total transmission rate of the 2nd time slot to node D, at first source node S calculates the power division that obtains optimum while selecting any via node C, at this, when

The time, the dump energy size of comparison source node S and via node C: if the source node S dump energy is larger, order

Otherwise $\max f_D^2(P_{2S},P_{2C})=f_D^2(0,P_{\max });$

The 4th step: according to formula

$P_{1S}^{*}=P_{\max }$

${\mathrm{\&Delta;}}_1^{*}=\frac{I}{f_{\mathrm{SC}}^1\left(P_{1S}^{*}\right)}=\frac{I}{{W\log }_2(1+\frac{h_{\mathrm{SC}}{P}_{1S}^{*}}{N_{C}W})},---\left(4\right)$

With

${\mathrm{\&Delta;}}_2^{*}=\frac{I_D^2}{\max f_D^2(P_{2S},P_{2C})}---\left(24\right)$

And

${\mathrm{\&Delta;}}_C^{*}={\mathrm{\&Delta;}}_1^{*}+{\mathrm{\&Delta;}}_2^{*}=\frac{I}{f_{\mathrm{SC}}^1\left(P_{1S}^{*}\right)}+\frac{I_D^2}{\max f_D^2(P_{2S},P_{2C})}---\left(25\right)$

Wherein,

Be illustrated in second time slot node D decoding information needed amount,

Source node S is calculated the propagation delay time of optimum while selecting any via node C;

The 5th step: when C ∈ Ν, the more corresponding optimal transmission time delay of source node S, according to formula

$C^{*}=\underset{C\&Element;N}{\arg \min }{\mathrm{\&Delta;}}_C^{*}---\left(26\right)$

Obtain optimal relay node C ^*, wherein

At optimal relay node C ^*In situation, the optimal power allocation of trying to achieve and optimum propagation delay time are the theoretical optimal solution of combined optimization model;

The 6th step: source node S sends message to via node C ^*, notify this node to complete the collaboration communication process to destination node D according to optimum power distribution result.

Images