CN108810885A - Uplink dual-connection data distribution method based on joint confidentiality degree and power consumption optimization - Google Patents

Abstract

An uplink dual-connection data splitting method based on joint confidentiality and power consumption optimization, wherein the optimization problem of minimizing the total power consumption of MU under the condition of satisfying data confidentiality requirements and energy effectiveness is described as a multivariable non-convex optimization problem; problem P1 is converted into a bottom subproblem P1-Sub and a top problem P1-Top for optimization and solution; the bottom subproblem P1-Sub is equivalently converted into the P2-E problem for multiple times; the problem optimization solution is obtained when the control variable range in the P2-E problem is given; the control variable ranges of the P2-E problem under different conditions are listed and substituted into the P2-E problem optimization solution obtained when the control variable range is given; the optimal solution of the bottom subproblem P1-Sub is obtained; after obtaining the optimal solution of the bottom subproblem P1-Sub, the top problem P1-Top is solved by using a linear search method to obtain the optimal solution of the top problem, and finally the optimal solution of the entire optimization problem is obtained. The present invention has high efficiency and high flexibility.

Description

An uplink dual connection data distribution based on joint security degree and power consumption optimization stream method

technical field

The invention relates to a wireless network, in particular to an uplink dual connection data distribution method based on joint security degree and power consumption optimization.

Background technique

In the past decade, the explosive growth of smart mobile terminals, the increasing popularity of mobile network services, and the huge traffic generated in cellular networks. In the multi-layer structure of the radio access network, a large number of heterogeneous small base stations are densely covered in the unit of the macro base station, and the macro base station offloads mobile traffic to the small base station. This method is data offloading. As an effective and cost-effective method, data offloading has played a significant role in alleviating traffic congestion in macro base station cellular networks. However, the efficiency and flexibility of this single method are relatively lacking. In order to better distribute data and manage resources more flexibly, the Third Generation Partnership Project proposes "dual connectivity" technology, which enables users (MobileUsers, MUs) to use two different radio interfaces and macro base stations (macro Base Station) , BS) to communicate, and at the same time transmit the offloaded data to a small base station (small-cell Access Point, AP).

Contents of the invention

In order to overcome the disadvantages of low efficiency and low flexibility in the prior art, the present invention provides an uplink dual connection data offloading method based on joint security degree and power consumption optimization in a wireless network with high efficiency and high flexibility.

The technical solution adopted by the present invention to solve its technical problems is:

A method for offloading uplink dual connection data based on joint confidentiality and power consumption optimization, said method comprising the following steps:

(1) There is a mobile user MU under the coverage of the base station BS, and a small cell auxiliary network access point AP is deployed to provide data offloading services for the MU through "dual connectivity";

In a wireless network, the optimization problem of minimizing the total power consumption of a MU while satisfying data confidentiality requirements and energy efficiency is described as a non-convex optimization problem P1 shown below, which is expressed as follows:

min p _iA +p _iB

limitation factor:

x _iA ≥ 0

x _iB ≥ 0

Control variables: (x _iA ,p _iA ) and (x _iB ,p _iB )

In the P1 problem, x _iB represents the maximum data flow required by the MU on the BS side, p _iB represents the energy consumed by the MU on the BS side; x _iA represents the maximum data flow required by the MU on the AP side, and p _iA represents the energy consumed by the MU on the AP side Energy consumed by MU; P _out is a function of p _iA and x _iA , expressed as P _out (p _iA , x _iA ), formula (1-5) is obtained through Shannon's theorem;

The meaning of each variable in the question is explained as follows:

p _iA : energy consumed by the MU on the AP side/W;

p _iB : energy consumed by the MU on the BS side/W;

x _iB : the maximum data flow required by the MU on the BS side;

x _iA : the maximum data flow required by the MU on the AP side;

W _B : channel bandwidth/HZ from MU to BS;

W _A : channel bandwidth/HZ from MU to AP;

g _iA : channel gain from MU to AP;

g _iB : channel gain from MU to BS;

g _iE : channel gain from MU to eavesdropper;

n _A : background noise power from MU to AP/W;

n _B : background noise power from MU to BS/W;

n _E : background noise power from MU to eavesdropper/W;

The maximum confidential data throughput that can be obtained from MU to AP;

P _out : probability of confidentiality overflow when AP provides data offload service to MU The maximum energy consumption from MU to AP/W;

The maximum energy consumption from MU to BS/W;

The upper bound of the confidentiality overflow probability of MU;

∈ _i : Confidentiality overflow probability of MU;

α _i : the average value of channel gain from MU to eavesdropper;

(2) Through the analysis of the P1 problem, the P1 problem is decomposed into a bottom sub-problem P1-Sub and a top-level problem P1-Top for optimal solution. The bottom sub-problem P1-Sub is as follows:

V(∈ _i )=minp _iA +p _iB

Constraints: P _out (p _iA , x _iA )＝∈ _i (2-1)

x _iA ≥ 0

x _iB ≥ 0

Control variables: (x _iA ,p _iA ) and (x _iB ,p _iB )

The top-level problem P1-Top looks like this:

min V(ε _i )

limitation factor:

Control variable: ε _i

In the process of optimizing the solution of the P1 problem, the underlying sub-problem P1-Sub is first optimized and solved step by step;

(3) The expression of the probability function P _out (p _iA , x _iA ) of confidentiality overflow is as follows:

in the above formula Indicates the maximum confidential data throughput that can be obtained from MU to AP, and its expression is as follows:

Substitute formula (3-2) into (3-1) to get the expression of P _out (p _iA , x _iA ) as follows:

define an auxiliary Indicates the effective channel power gain from MU to AP, and its expression is as follows:

Combining formula (3-4) to get the expression of P _out (p _iA , x _iA ) as follows:

(4) Through the simultaneous analysis of (1-1) and (3-5), the restricted expression of (1-1) is obtained as follows:

Define a new variable θ _iA to quantify the impact of confidentiality requirements, the expression of θ _iA is as follows:

By further transforming (4-1), the equivalent expression of (1-1) is obtained as follows:

In the optimization scheme of the P1 problem, the above formula is a strict constraint of the problem, and in the problem analysis, the split data flow rate of the MU satisfies the following expression:

Through the analysis of (2-2) and (4-4), the following expressions are obtained:

Therefore, by combining (2-5) and (4-5), the expression can be obtained as follows:

(5) The equivalent transformation of the P1-Sub problem, substituting (4-5), (4-6) and the above are related to the P1-Sub problem, and the P2 problem is expressed as follows:

limitation factor:

Control variable: p _iA

Perform equivalent transformation on (5-1), and get the following expression:

Similarly, (5-2) is also converted equivalently, and the expression is as follows:

Through (5-3) and (5-4), the P2 problem is equivalently transformed into a P2-E problem, and "E" means equivalence, as follows:

Restrictions: Conditions (1-3)

Conditions (5-3)

Conditions (5-4)

Control variable: p _iA

The constraints (5-3) and (5-4) in the P2-E problem are both linearly related to p _iA , so in the parameter setting, the three constraints (5-3), (5-4), ( 1-3) Generate a feasible interval about p _iA , namely

(6) The P2-E problem is regarded as a convex optimization problem, and the first-order derivation is performed on the objective function in P2-E, and the expression of the first-order derivative is obtained as follows:

know by analysis is an increasing function about p _iA ;

(7) In the given and In the case of , the algorithm SolP2E to solve the problem according to the monotonicity of the first derivative of the objective function in the P2-E problem is as follows:

Step 7.1: Set the tolerance value of the calculation error to γ, flag=1;

Step 7.2: If established, then Go to step 7.6; if established, then Go to step 7.6, otherwise go to step 7.3;

Step 7.3 Setup

Step 7.4: When flag=1, get if established, then Set flag=0 at the same time, execute step 7.6;

Step 7.5: If established when satisfied when, update Return to step 7.4; when satisfied when, update Return to step 7.4;

Step 7.6: end the loop;

Step 7.7: Output the current optimal solution of the P2-E problem as

(8) In the algorithm SolP2E, it is the upper bound of the given p _iA with the nether The case is calculated, so the upper bound of p _iA with the nether To solve it, it is necessary to consider various situations and Define two new parameters K and L whose expressions are as follows:

Through parameters K and L, (5-3) and (5-4) are transformed into the following expressions:

Based on the above two formulas (8-3) and (8-4) to get and It is necessary to consider K and L in different situations. First, through the analysis (8-3), two different situations are obtained, namely CaseⅠ: KL≥1; CaseⅡ: KL<1;

Case I is the case under KL≥1, in this case, satisfy Indicates that the BS can meet all the traffic demands of the MU, and does not require the AP to perform data offloading; on the contrary, when KL<1 in Case II, the BS cannot meet all the data traffic demands of the MU. Therefore, the P2-E problem may occur in Case II is not feasible;

If KL≥1, it is Case I. Through analysis, two sub-cases are obtained, as follows:

CaseⅠ.1: when when, get

CaseⅠ.2: when when, get

If KL<1, it is Case II. Five sub-cases are obtained through analysis, as follows:

CaseⅡ.1: when The P2-E problem is infeasible;

CaseⅡ.2a: when and when, get

CaseⅡ.2b: when and When , the P2-E problem is infeasible;

CaseⅡ.3a: When and when, get

CaseⅡ.3b: when and When , the P2-E problem is infeasible;

Eventually the following situations are obtained:

Case I.1

case

I.2

Case II.1 The P2-E problem is infeasible;

Case II.2a( and ):

Case II.2b( and ): P2-E problem is infeasible; CaseⅡ.3a( and ):

Case II.3b( and ):

The P2-E problem is infeasible;

Obtained through the above process and Substitute into the algorithm SolP2E to get the optimal solution The optimal solution obtained by Get the other three optimal solutions corresponding to the P2-E problem as follows:

The above is the optimal solution of the P2-E problem, that is, the optimal solution of the energy consumed by the MU on the AP side in the P1-Sub problem Optimal solution for MU's data distribution requirements on the AP side The Optimal Solution of Energy Consumption by MU on the BS Side Optimal Solution of MU's Data Requirements on BS Side

(9) The optimal solution of the top-level problem P1-Top. Through the analysis of the bottom-level sub-problems, the top-level problem P1-Top is expressed as follows:

limitation factor:

Control variable: ∈ _i

The algorithm SolP1Top to solve P1-Top according to the linear search method of ∈ _i in the feasible range is as follows:

Step 9.1: Set the current optimal solution CBS as an empty set, the current optimal energy consumption value CBV = ∞, and set the initial value of ∈ _i to Δ, and the step size to Δ;

Step 9.2: If ∈ _i satisfies Then go to step 9.3; otherwise go to step 9.6;

Step 9.3: Bring ∈ _i into the objective function of the top-level problem P1-Top, and judge whether the obtained V(∈ _i ) is less than the current optimal energy consumption value CBV;

Step 9.4: If V(∈ _i )≥CBV is established, then update ∈ _i =∈ _i +Δ, return to step 9.2;

Step 9.5: If V(∈ _i )<CBV holds, then update the current optimal solution The current optimal energy consumption value is V ^* (∈ _i ), while updating ∈ _i =∈ _i +Δ, return to step 9.2;

Step 9.6: end the loop;

Step 9.7: If the current optimal energy consumption value CBV is ∞, then the P1 problem is not feasible, otherwise output the current optimal solution The current optimal energy consumption value is V ^* (∈ _i );

(10) Through the hierarchical solution to the P1 problem, the optimal solution of the energy consumed by the MU on the AP side is obtained Optimal solution for MU's data distribution requirements on the AP side The Optimal Solution of Energy Consumption by MU on the BS Side Optimal Solution of MU's Data Requirements on BS Side Optimal Solution of MU's Secrecy Degree The optimal energy consumption value of MU is V ^* (∈ _i ).

The invention is an optimized design of wireless network data distribution based on dual connection. Considering that in the wireless network, the AP that provides the data offload service for the MU works in an unlicensed frequency band, this leads to an eavesdropper being able to eavesdrop on the data traffic offloaded to the AP in the unlicensed frequency band. Therefore, what the present invention researches is the optimization of the joint security degree and power consumption under the dual connection, and the scheme design of optimizing the data offloading service of the AP. Aiming at the idea proposed above, the present invention studies the design of an uplink dual connection data distribution scheme based on joint secrecy degree and power consumption.

The technical idea of the present invention is as follows: firstly, it is considered that in a heterogeneous wireless network, the AP and the BS achieve certain economic benefits by offloading data to the MU to minimize power. In the present invention, by analyzing the characteristics of the problem, the problem is converted into a bottom sub-problem and a top-level problem. Transform the underlying subproblem into an easily solvable convex optimization problem through an equivalent transformation. Afterwards, by analyzing the obtained equivalent problem, assuming that the control variable in the problem has been given, and then according to the monotonicity of the first derivative of the objective function and the binary search method, to obtain the current range of the given control variable The optimal solution of the underlying subproblem of . Afterwards, through the analysis of the problem, the control variable ranges in different situations are listed, and then the different control variable ranges are substituted into the obtained optimal solution. Finally, the optimal solution to the underlying sub-problem is obtained. Finally, the top-level problem is solved by the linear search method, and the optimal solution of the top-level problem is obtained. Finally, an optimal solution of uplink dual connection data offloading method based on joint security degree and power consumption optimization is proposed.

The beneficial effects of the present invention are mainly manifested in: 1. For the overall system, the dual connection technology greatly improves the utilization rate of wireless resources; The total power loss under dual connections not only obtains better uplink data traffic services, ensures the confidentiality requirements of MU, but also improves energy utilization.

Description of drawings

Fig. 1 is a schematic diagram of a scene of a user MU, a macro base station BS, a small base station AP and an eavesdropper in a wireless network.

Detailed ways

The present invention will be described in further detail below in conjunction with the accompanying drawings.

Referring to Figure 1, an uplink dual-connection data offloading method based on joint security degree and power consumption optimization, the implementation of this method can minimize the total energy consumption of the MU under the premise of meeting the appropriate security requirements of the MU, and improve the efficiency of the entire system. Wireless resource utilization and energy utilization, the present invention can be applied to a wireless network, as shown in FIG. 1 . The optimization method designed for the problem mainly includes the following steps:

(1) There is a mobile user MU under the coverage of the base station BS, and a small cell auxiliary network access point AP is deployed to provide data offloading services for the MU through "dual connectivity";

In a wireless network, the optimization problem of minimizing the total power consumption of a MU while satisfying data confidentiality requirements and energy efficiency is described as a non-convex optimization problem P1 shown below, which is expressed as follows:

min p _iA +p _iB

limitation factor:

x _iA ≥ 0

x _iB ≥ 0

Control variables: (x _iA ,p _iA ) and (x _iB ,p _iB )

In the P1 problem, x _iB represents the maximum data flow required by the MU on the BS side, p _iB represents the energy consumed by the MU on the BS side; x _iA represents the maximum data flow required by the MU on the AP side, and p _iA represents the energy consumed by the MU on the AP side Energy consumed by MU; P _out is a function of p _iA and x _iA , expressed as P _out (p _iA , x _iA ), formula (1-5) is obtained through Shannon's theorem;

The meaning of each variable in the question is explained as follows:

p _iA : energy consumed by the MU on the AP side/W;

p _iB : energy consumed by the MU on the BS side/W;

x _iB : the maximum data flow required by the MU on the BS side;

x _iA : the maximum data flow required by the MU on the AP side;

W _B : channel bandwidth/HZ from MU to BS;

W _A : channel bandwidth/HZ from MU to AP;

g _iA : channel gain from MU to AP;

g _iB : channel gain from MU to BS;

g _iE : channel gain from MU to eavesdropper;

n _A : background noise power from MU to AP/W;

n _B : background noise power from MU to BS/W;

n _E : background noise power from MU to eavesdropper/W;

The maximum confidential data throughput that can be obtained from MU to AP;

P _out : probability of confidentiality overflow when AP provides data offload service to MU The maximum energy consumption from MU to AP/W;

The maximum energy consumption from MU to BS/W;

The upper bound of the confidentiality overflow probability of MU;

∈ _i : Confidentiality overflow probability of MU;

α _i : the average value of channel gain from MU to eavesdropper;

(2) Through the analysis of the P1 problem, the P1 problem is decomposed into a bottom sub-problem P1-Sub and a top-level problem P1-Top for optimal solution. The bottom sub-problem P1-Sub is as follows:

V(∈ _i )=minp _iA +p _iB

Restrictions: P _out (p _iA , x _iA )＝ε _i (2-1)

x _iA ≥ 0

x _iB ≥ 0

Control variables: (x _iA ,p _iA ) and (x _iB ,p _iB )

The top-level problem P1-Top looks like this:

min V(∈ _i )

limitation factor:

Control variable: ∈ _i

In the process of optimizing the solution of the P1 problem, the underlying sub-problem P1-Sub is first optimized and solved step by step;

(3) The expression of the probability function P _out (p _iA , x _iA ) of confidentiality overflow is as follows:

in the above formula Indicates the maximum confidential data throughput that can be obtained from MU to AP, and its expression is as follows:

Substitute formula (3-2) into (3-1) to get the expression of P _out (p _iA , x _iA ) as follows:

define an auxiliary Indicates the effective channel power gain from MU to AP, and its expression is as follows:

Combining formula (3-4) to get the expression of P _out (p _iA , x _iA ) as follows:

(4) Through the simultaneous analysis of (1-1) and (3-5), the restricted expression of (1-1) is obtained as follows:

Define a new variable θ _iA to quantify the impact of confidentiality requirements, the expression of θ _iA is as follows:

By further transforming (4-1), the equivalent expression of (1-1) is obtained as follows:

In the optimization scheme of the P1 problem, the above formula is a strict constraint of the problem, and in the problem analysis, the split data flow rate of the MU satisfies the following expression:

Through the analysis of (2-2) and (4-4), the following expressions are obtained:

Therefore, by combining (2-5) and (4-5), the expression is as follows:

(5) The equivalent transformation of the P1-Sub problem, substituting (4-5), (4-6) and the above are related to the P1-Sub problem, and the P2 problem is expressed as follows:

limitation factor:

Control variable: p _iA

Perform equivalent transformation on (5-1), and get the following expression:

Similarly, (5-2) is also converted equivalently, and the expression is as follows:

Through (5-3) and (5-4), the P2 problem is equivalently transformed into a P2-E problem, and "E" means equivalence, as follows:

Restrictions: Conditions (1-3)

Conditions (5-3)

Conditions (5-4)

Control variable: p _iA

The constraints (5-3) and (5-4) in the P2-E problem are both linearly related to p _iA , so in the parameter setting, the three constraints (5-3), (5-4), ( 1-3) Generate a feasible interval about p _iA , namely

(6) The P2-E problem is regarded as a convex optimization problem, and the first-order derivation is performed on the objective function in P2-E, and the expression of the first-order derivative is obtained as follows:

know by analysis is an increasing function about p _iA ;

(7) In the given and In the case of , the algorithm SolP2E to solve the problem according to the monotonicity of the first derivative of the objective function in the P2-E problem is as follows;

Step 7.1: Set the tolerance value of the calculation error to γ, flag=1;

Step 7.2: If established, then Go to step 7.6; if established, then Go to step 7.6, otherwise go to step 7.3;

Step 7.3 Setup

Step 7.4: When flag=1, get if established, then Set flag=0 at the same time, execute step 7.6;

Step 7.5: If established when satisfied when, update Return to step 7.4; when satisfied when, update Return to step 7.4;

Step 7.6: end the loop;

Step 7.7: Output the current optimal solution of the P2-E problem as

(8) In the algorithm SolP2E, it is the upper bound of the given p _iA with the nether The case is calculated, so the upper bound of p _iA with the nether To solve it, it is necessary to consider various situations and Define two new parameters K and L whose expressions are as follows:

Through parameters K and L, (5-3) and (5-4) are transformed into the following expressions:

Based on the above two formulas (8-3) and (8-4) to get and It is necessary to consider K and L in different situations. First, through the analysis (8-3), two different situations are obtained, namely CaseⅠ: KL≥1; CaseⅡ: KL<1;

Case I is the situation under KL ≥ 1. In this case, W _B log ₂ (1+piBmaxgiBnB ≥ Rireq is satisfied, which means that BS can meet all traffic demands of MU, and AP does not need data offloading; on the contrary, Case II is in KL In the case of <1, the BS cannot meet all the data flow requirements of the MU, therefore, the P2-E problem may not be feasible in the case of Case II;

If KL≥1, it is Case I. Through analysis, two sub-cases are obtained, as follows:

CaseⅠ.1: when when, get

CaseⅠ.2: when when, get

If KL<1, it is Case II. Five sub-cases are obtained through analysis, as follows:

CaseⅡ.1: When The P2-E problem is infeasible;

CaseⅡ.2a: when and when, get

CaseⅡ.2b: when and When , the P2-E problem is infeasible;

CaseⅡ.3a: when and when, get

CaseⅡ.3b: when and When , the P2-E problem is infeasible;

Eventually the following situations are obtained:

Case I.1

case

I.2

Case II.1 The P2-E problem is infeasible;

Case II.2a( and ):

Case II.2b( and ): P2-E problem is infeasible; CaseⅡ.3a( and ):

Case II.3b( and ):

The P2-E problem is infeasible;

Obtained through the above process and Substitute into the algorithm SolP2E to get the optimal solution The optimal solution obtained by Get the other three optimal solutions corresponding to the P2-E problem as follows:

The above is the optimal solution of the P2-E problem, that is, the optimal solution of the energy consumed by the MU on the AP side in the P1-Sub problem Optimal solution for MU's data distribution requirements on the AP side The Optimal Solution of Energy Consumption by MU on the BS Side Optimal Solution of MU's Data Requirements on BS Side

(9) The optimal solution of the top-level problem P1-Top. Through the analysis of the bottom-level sub-problems, the top-level problem P1-Top is expressed as follows:

limitation factor:

Control variable: ∈ _i

The algorithm SolP1Top to solve P1-Top according to the linear search method of ∈ _i in the feasible range is as follows:

Step 9.1: Set the current optimal solution CBS as an empty set, the current optimal energy consumption value CBV = ∞, and set the initial value of ∈ _i to Δ, and the step size to Δ;

Step 9.2: If ∈ _i satisfies Then go to step 9.3; otherwise go to step 9.6;

Step 9.3: Bring ∈ _i into the objective function of the top-level problem P1-Top, and judge whether the obtained V(∈ _i ) is less than the current optimal energy consumption value CBV;

Step 9.4: If V(∈ _i )≥CBV is established, then update ∈i=∈ _i +Δ, return to step 9.2;

Step 9.5: If V(∈ _i )<CBV holds, then update the current optimal solution The current optimal energy consumption value is V ^* (∈ _i ), while updating ∈ _i =∈ _i +Δ, return to step 9.2;

Step 9.6: end the loop;

Step 9.7: If the current optimal energy consumption value CBV is ∞, then the P1 problem is not feasible, otherwise output the current optimal solution The current optimal energy consumption value is V ^* (∈ _i );

(10) Through the hierarchical solution to the P1 problem, the optimal solution of the energy consumed by the MU on the AP side is obtained Optimal solution for MU's data distribution requirements on the AP side The Optimal Solution of Energy Consumption by MU on the BS Side Optimal Solution of MU's Data Requirements on BS Side Optimal Solution of MU's Secrecy Degree The optimal energy consumption value of MU is V ^* (∈ _i ).

In this embodiment, FIG. 1 is a system in which a wireless network considered in the present invention includes a macro base station BS, a user MU, a small base station AP, and an eavesdropper. In this system, the main consideration does not include interference, but it will consider 1. The channel environment between the user MU and the small base station AP, the channel environment between the user MU and the macro base station BS, and the communication between the user MU and the eavesdropper Channel environment; 2. Confidentiality requirement of user MU; 3. Total energy consumption of user MU in different situations. In order to make MU achieve a goal of meeting the confidentiality requirement while minimizing energy consumption, an invention is proposed to solve this problem.

This implementation case focuses on minimizing the total energy consumption of the MU under dual connectivity under the premise of meeting the appropriate confidentiality requirements of the user MU, jointly optimizing the degree of confidentiality and power consumption, and improving the utilization rate of wireless resources and energy utilization. In the implementation process of the present invention, it benefits from the effective transformation of the problem and the reduction of the computational complexity of the optimization algorithm.

Claims (1)

1. a kind of uplink dual link data distribution method based on joint privacy degrees and power consumption optimization, which is characterized in that It the described method comprises the following steps：

(1) there are one mobile subscriber MU under the coverage area of base station BS, while deploying a cellulor auxiliary network insertion Point AP provides data distribution service by " dual link " for MU；

In the wireless network, the general power that MU is minimized in the case where meeting data security requirement and energy efficiency disappears The optimization problem of consumption describes the nonconvex property optimization problem P1 problems being as follows, and the problem representation is as follows：

min p_iA+p_iB

Restrictive condition：

x_iA≥0

x_iB≥0

Control variable：(x_iA, p_iA) and (x_iB, p_iB)

In P1 problems, x_iBIndicate the attainable maximum data demand volume of BS side MU institutes, p_iBIndicate the energy of BS side MU consumption； x_iAIndicate the attainable maximum data demand volume of AP side MU institutes, p_iAIndicate the energy of AP side MU consumption；P_outIt is about p_iAWith x_iAFunction, be expressed as P_out(p_iA, x_iA), formula (1-5) is obtained by Shannon's theorems；

The meaning of each variable in problem is described as follows：

p_iA：Energy/W of the sides AP MU consumption；

p_iB：Energy/W of the sides BS MU consumption；

x_iB：The attainable maximum data demand volume of the sides BS MU institutes；

x_iA：The attainable maximum data demand volume of the sides AP MU institutes；

W_B：Channel width/HZ of MU to BS；

W_A：Channel width/HZ of MU to AP；

g_iA：The channel gain of MU to AP；

g_iB：The channel gain of MU to BS；

g_iE：Channel gains of the MU to listener-in；

n_A：Background Noise Power/W of MU to AP；

n_B：Background Noise Power/W of MU to BS；

n_E：Background Noise Power/Ws of the MU to listener-in；

The maximum private data handling capacity that MU to AP can be obtained；

P_out：The probability that confidentiality of the AP when providing data distribution service to MU is overflowed

Maximum consumption energy/W of MU to AP；

Maximum consumption energy/W of MU to BS；

The upper bound of the confidentiality overflow probability of MU；

∈_i：The confidentiality overflow probability of MU；

α_i：Average values of the MU to listener-in's channel gain；

(2) it is an a bottom subproblem P1-Sub and top layer problem P1- by P1 PROBLEM DECOMPOSITIONs by the analysis to P1 problems Top optimizes solution, and bottom subproblem P1-Sub therein is as follows：

V(∈_i)=min P_iA+P_iB

Restrictive condition：P_out(p_iA, x_iA)=∈_i (2-1)

x_iA≥0

x_iB≥0

Control variable：(x_iA, P_iA) and (x_iB, P_iB)

Top layer problem P1-Top is as follows：

min V(∈_i)

Restrictive condition：

Control variable：∈_i

During the Optimization Solution of P1 problems, Optimization Solution gradually first is carried out to bottom subproblem P1-Sub；

(3) the probability function P that confidentiality is overflowed_out(p_iA, x_iA) expression formula is as follows：

In above formulaIndicate that the maximum private data handling capacity that MU to AP can be obtained, expression formula are as follows：

Formula (3-2) substitution (3-1) is obtained into P_out(p_iA, x_iA) expression formula is as follows：

Define an auxiliary quantityIndicate that the efficient channel power gain of MU to AP, expression formula are as follows：

Convolution (3-4) obtains P_out(p_iA, x_iA) expression formula is as follows：

(4) by carrying out simultaneous analysis to (1-1) and (3-5), the limiting expression formula for obtaining (1-1) is as follows：

Define a new variable θ_iATo quantify the influence of confidentiality demand, θ_iAExpression formula it is as follows：

By the further conversion to (4-1), the equivalent expression for obtaining (1-1) is as follows：

And in the optimization scheme of P1 problems, above formula is a hard constraints of problem, and in case study, the shunting of MU Data traffic rate meets following expression：

Following expression is obtained by the analysis of (2-2) and (4-4)：

Therefore, by simultaneous (2-5) and (4-5), it is as follows to obtain expression formula：

(5) equivalent conversion of P1-Sub problems, substitute into (4-5), (4-6) or more is respectively related to P1-Sub problems, obtains P2 Problem representation is as follows：

Restrictive condition：

Control variable：P_iA

Equivalent conversion is carried out to (5-1), it is as follows to obtain expression formula：

Equivalent conversion is also equally carried out to (5-2), it is as follows to obtain expression formula：

Convert P2 problems progress equivalence to P2-E problems by (5-3) and (5-4), what " E " was indicated be it is of equal value, it is as follows：

Restrictive condition：Condition (1-3)

Condition (5-3)

Condition (5-4)

Control variable：P_iA

Restrictive condition (5-3) and (5-4) in P2-E problems all with P_iAIt is linear, so in parameter setting, three limits Condition (5-3) processed, (5-4), (1-3) produce one about P_iAFeasible section, i.e.,

(6) P2-E problems regard a convexity optimization problem as, carry out first derivation to the object function in P2-E, obtain one Order derivative expression formula is as follows：

Known by analysisIt is about P_iAIncreasing function；

(7) givenWithIn the case of, it is solved according to the monotonicity of the first derivative of object function in P2-E problems The algorithm SolP2E of the problem is as follows；

Step 7.1：It is arranged and calculates the tolerance value of error as γ, flag=1；

Step 7.2：Such as bodyIt sets up, thenExecute step 7.6；If It sets up, thenStep 7.6 is executed, it is no to then follow the steps 7.3；

Step 7.3 is arranged

Step 7.4：As flag=1, obtainIfIt sets up, then Flag=0 is set simultaneously, executes step 7.6；

Step 7.5：IfIt sets up, works as satisfactionWhen, updateIt returns Step 7.4；Work as satisfactionWhen, updateReturn to step 7.4；

Step 7.6：End loop；

Step 7.7：Output P2-E problems current optimal solution be

(8) it is in given p in algorithm SolP2E_iAThe upper boundWith lower boundIn the case of calculate, so right p_iAThe upper boundWith lower boundSolved, need to consider it is a variety of in the case ofWithDefine two it is new Parameter K and L, expression formula are as follows：

By parameter K and L, (5-3) and (5-4) is converted into following expression：

It is obtained based on above (8-3) and (8-4) two formulaWithIt needs to consider the K and L under different situations, it is logical first Analysis (8-3) is crossed, two different situations, i.e. Case I are obtained：KL≥1；

Case II：KL < 1；

Case I are that the situation under KL >=1 meets W in this case_B log₂(1+piBmaxgiBnB >=Rireq is indicated BS can meet whole flow demands of MU, do not need AP and carry out data distribution；On the contrary, Case II are in the case of KL < 1 time, BS cannot meet the data traffic demand of MU wholes, and therefore, P2-E problems may be infeasible in Case II；

If KL >=1, as Case I obtain two sub-cases by analysis, as follows：

Case I.1：WhenWhen, it obtains

Case I.2：WhenWhen, it obtains

If KL < 1, as Case II obtain five sub-cases by analysis, as follows：

Case II.1：WhenP2-E problems are infeasible；

Case II.2a：WhenAndWhen, it obtains

Case II.2b：WhenAndWhen, P2-E problems are infeasible；

Case II.3a：WhenAndWhen, it obtains

Case II.3b：WhenAndWhen,

P2-E problems are infeasible；

Finally obtain following situation：

Case I.

Case

I.

Case II.P2-E problems are infeasible；

Case II.

Case II.P2-E problems are infeasible；

Case II.

Case II.

P2-E problems are infeasible；

It is obtained by above procedureWithIt substitutes into algorithm SolP2E and obtains optimal solutionPass through obtained optimal solutionObtain other corresponding three optimal solutions of P2-E problems It is as follows：

It is the optimal solution of P2-E problems above, as in P1-Sub problems, MU consumes the optimal solution of energy in the sides APMU exists AP side datas shunt demand optimal solutionMU consumes the optimal solution of energy in the sides BSMU is optimal BS side data demands Solution

(9) Optimization Solution of top layer problem P1-Top, by the analysis to bottom subproblem, top layer problem P1-Top indicates as follows It is shown：

Restrictive condition：

Control variable：∈_i

According to ∈_iLinear search method in feasible region is as follows come the algorithm SolP1Top for solving P1-Top：

Step 9.1：It is empty set, current optimum energy consumption value CBV=∞, while ∈ is arranged that current optimal solution CBS, which is arranged,_iJust Value is Δ, and step-length is also Δ；

Step 9.2：If ∈_iMeetThen follow the steps 9.3；It is no to then follow the steps 9.6；

Step 9.3：By ∈_iIt brings into the object function of top layer problem P1-Top, judges obtained V (∈_i) whether be less than currently Optimum energy consumption value CBV；

Step 9.4：If V (∈_i) >=CBV is set up, then updating ∈_i=∈_i+ Δ, return to step 9.2；

Step 9.5：If V (∈_i) < CBV establishments, then updating current optimal solutionCurrent optimum energy consumption value For V^*(∈_i), while updating ∈_i=∈_i+ Δ, return to step 9.2；

Step 9.6：End loop；

Step 9.7：If current optimum energy consumption value CBV is ∞, P1 problems are infeasible, otherwise export current optimal solutionCurrent optimum energy consumption value is V^*(∈_i)；

(10) it by the hierarchical solving to P1 problems, obtains MU and consumes the optimal solution of energy in the sides APMU is in AP side datas point Stream demand optimal solutionMU consumes the optimal solution of energy in the sides BSOptimal solutions of the MU in BS side data demandsMU's Confidentiality degree optimal solutionThe optimum energy consumption value of MU is V^*(∈_i)。