CN106454994A - Power control technology based on dynamic logical topology - Google Patents

Abstract

For the problem that coordination optimization management of network energy having dynamic logical topology is difficultly realized, elastic optimization management of ad-hoc network energy is realized by establishing a power optimization model and a network resource overhead optimization mechanism.

Description

A kind of power control techniques based on dynamic logic topology

Technical field

The present invention relates to intelligent grid field, more particularly to communication network, and optimum theory.

Background technology

At present, wireless communication technology is just advanced at an unprecedented rate, from analog circuit exchange system to numeral Circuit switching system, from the wireless phone network of limited coverage area to global mobile communication network, current forth generation movement is logical The theoretic transmision peak of letter system (4G) creates huge wealth up to 100Mbps, the mankind that develop into of wireless communication technology Rich.According to the difference of network structure, cordless communication network can be divided into has infrastructure networks and foundation-free facility network.Wide at present The cell mobile communication systems of general application belongs to infrastructure networks, and terminal is connected with base station, by core net control terminal Information transfer.Foundation-free facility network is exactly wireless self-organization network (AdHoc network), and wireless Mesh netword be exactly by Ad Hoc network is developed, and cognitive radio Mesh network (Cognitive Wireless Mesh Network, CWMN) is just It is that cognitive radio and broadband wireless Mesh network combine, a kind of Wideband with cognitive competence can be set up wireless Network.

In network, the end-to-end route of dynamic change is one section of transmission path comprising multi-hop, and uses shape according to frequency spectrum State, being particularly likely that using channel for each jump is quite different, and on the other hand, the arrival due to primary user PU may result in Frequent switching channel, will take into full account how many channel that can access and possibility in transmission path during frequency spectrum switching By using frequency range and be likely to result in frequency spectrum using change PU number, this frequency spectrum perception technology undoubtedly to network design High requirement is proposed, because once the perception to available frequency band occurs in that deviation, the transmission of user node is easy to primary Family and neighbor node are interfered, and adopt power control techniques solve this problem well, by adjusting cognitive using The transmission power at family makes which not affect normally receiving for primary user's signal in the interference summation of primary user's receiving terminal, it is ensured that main User's proper communication on the premise of not heavily disturbed, realizes the sharing frequency spectrum resource of primary user and cognitive user, frequency spectrum Space schematic diagram is as shown in Figure 1；On the other hand, Power Control can also realize the resource rational utilization between cognitive user, subtract Warfare between multiple cognitive user nodes in little dynamic network.

Therefore, cognitive user not only can be reduced by the through-put power of each cognitive user of power control techniques reasonable distribution Interference to neighbor node, can also realize maximally utilizing for Internet resources.

Content of the invention

The technical problem to be solved is：Optimize machine by setting up power optimization model and network resource overhead System, realizes the elastic optimum management of self-organizing network energy.

The present invention is comprised the following steps by solving the technical scheme that above-mentioned technical problem is adopted, as shown in Figure 2：

A, set up power control mechanism；

B, network resource overhead optimization mechanism is set up, ease up including signaling consumption optimization, the number of dropped packets for optimizing single session Deposit expense optimization.

In step A, power control mechanism is：A.MN monitors the outer busy tone BT of band and whether there is, if its presence, goes to Step b, on the contrary then go to step g；B. using powerSend RTS；C. destination node receive RTS, and judge be No have the outer busy tone BT of band, if there is the outer busy tone BT of band, abandoning RTS, otherwise then going to step d；D.MN using power P= P_maxCTS is sent, and judges network capacityWhether setting up, step e is gone to if setting up, otherwise then go to step Rapid b, whereinEqual transmission path number for needed for packet is single-hop average transmission distance, C_fFor the capacity of channel f, p_fFor The use probability of channel f, it is path number that F is channel number, z；E. destination node receives CTS, and uses power Send traffic packets；F.MN uses power P=P_maxRTS is sent, and step c is gone to, wherein MN is mobile terminal.

In step A, Power Control Optimized model is specially：V is the node set of network, and E is line set, λ^sdFor source Business demand between node s and destination node d, T=[λ^sd] it is business demand set between source node s and destination node d, C is link average size, and W is radio frequency set, C_epFor average routing capacity, T_iSupport most for the transmitting terminal of node i Big number of path, R_iFor the maximum path number that the receiving terminal of node i is supported, P_trFor the energy expenditure of transponder, P_osFor handing over Change the energy expenditure of unit, P_epEnergy expenditure in every Gbps business streaming, PC is the energy expenditure of network, N '_ij,wFor Under upper logical topology environment between source node i and destination node j usage frequency w path number,Be upper one Under logical topology environment between source node i and destination node j usage frequency w path number, and source node i and destination node There is physics direct connected link (m, n) between j,For running on link l_ijOn service traffics, N_ijFor source node i and purpose section Path number between point j, i, j are logical topology node, N_ij,wIt is usage frequency w between source node i and destination node j Path number,It is usage frequency w and the path number through (m, n) between source node i and destination node j, RO is logic Topological transformation expense, it is frequency sets that E is physics direct connected link set, W.

min PC+α×RO

In step B, specially：Signaling consumption is optimized for：

C_u(α)=2s_uh_MN-HAN_h

C_u(β)=2s_uh_MN-HAN_h+2s_lh_FA-HAN_h

C_u(γ)=2s_uh_MN-κN_h+2s_uh_FA-FAN_h+2s_lh_FA-κN_hN_g

C_u(θ)=2s_uh_MN-μN_h+2s_lh_LN-ξN_h

C_u(ω)=2s_uh_nη-εN_h+2s_uh_nη-εN_h

C_u(Δ)=2s_uh_nη-pηN_h+2s_uh_nη-εN_h

C_u(ω-δ)=2s_uh_pη-εN_h+2s_lh_nη-εN_h

C_u() is the login cost of certain network behavior, wherein t_sFor the average Connection Time of single session, t_ΓFor packet In the mean residence time of network, T_adFor the mean transit delay between agent node, N_hAveragely cut for Internet in single session Change number of times, N_gFor the remaining root node number in δ domain, s_uFor the average bag size of signaling information, s is to transmit in label switched path Packet mean size, h_x-yFor the average number of hops between destination node d in wire link, B_wFor the bandwidth of wire link, B_w1 For the bandwidth of wireless link, L_w1For the time delay of wireless link, L_wFor the time delay of wire link, P_t() is path delay function, λ_d For the weight coefficient of downlink packet, T_interFor the continuous time interval for reaching data parlor, T_cFor the flat of single switching All the time, it is multi-tag label switched domain that MN is mobile node, δ, and ξ is the next Network Access Point of mobile node, and κ is δ domain Root node, HA is the home agent node of user, and it is Mobile Access Gateway that LER is edge router, η, and ε is connect for local movement Enter end, ω be proxy mobile IPv 6, Δ be based on ω, the LN being switched fast for linkage node, γ be based on the shifting being switched fast Dynamic agent node, FA is that Foreign Agent node, θ is the δ path based on low-power consumption link, and n and p is that gateway identification, n η and p η divides Not Wei n-th and p Mobile Access Gateway, mobile node is transferred to δ domain from proxy mobile IPv 6 domain, and μ is edge router between domain, " " is that network or node are transferred to another Autonomous Domain, or the mark for being transferred to another state by a certain state by a certain Autonomous Domain Know, it is predecessor LER, ξ for current LER, τ for carrying out being switched fast δ domain for mobile δ, σ that α is mobile IP, β.

In step B, optimize the number of dropped packets of single session, wherein P_loss() is certain network for producing in single session The number of dropped packets of behavior,

In step B, caching expense optimization is specially：

B_size() is the nodal cache expense that certain network behavior is produced

Description of the drawings

Fig. 1 spectrum space schematic diagram

Power control procedures schematic diagram of the Fig. 2 based on dynamic logic topology

Specific embodiment

For reaching above-mentioned purpose, technical scheme is as follows：

The first step, sets up power optimization model, and power control mechanism is：A.MN monitors the outer busy tone BT of band and whether there is, if Which is present, then go to step b, otherwise then go to step g；B. using powerSend RTS；C. destination node is received RTS, and judge whether to carry outer busy tone BT, if there is the outer busy tone BT of band, RTS is abandoned, otherwise then goes to step d；d.MN Using power P=P_maxCTS is sent, and judges network capacityWhether setting up, step e is gone to if setting up, instead Then go to step b, whereinEqual transmission path number for needed for packet is single-hop average transmission distance, C_fFor channel f's Capacity, p_fFor the use probability of channel f, it is path number that F is channel number, z；E. destination node receives CTS, and uses powerSend traffic packets；F.MN uses power P=P_maxRTS is sent, and step c is gone to, wherein MN is mobile whole End.

3rd step, Power Control Optimized model is specially：V is the node set of network, and E is line set, λ^sdFor source node s Business demand between destination node d, T=[λ^sd] be business demand set between source node s and destination node d, C be Road average size, W is radio frequency set, C_epFor average routing capacity, T_iFor the maximum path that the transmitting terminal of node i is supported Number, R_iFor the maximum path number that the receiving terminal of node i is supported, P_trFor the energy expenditure of transponder, P_osFor crosspoint Energy expenditure, P_epEnergy expenditure in every Gbps business streaming, PC is the energy expenditure of network, N '_ij,wBe upper one Under logical topology environment between source node i and destination node j usage frequency w path number,It is to open up in a upper logic The path number of usage frequency w between source node i and destination node j under environment is flutterred, and between source node i and destination node j There is physics direct connected link (m, n),For running on link l_ijOn service traffics, N_ijFor source node i and destination node j it Between path number, i, j be logical topology node, N_ij,wIt is the path of usage frequency w between source node i and destination node j Number,It is usage frequency w and the path number through (m, n) between source node i and destination node j, RO is logical topology Conversion expense, it is frequency sets that E is physics direct connected link set, W.

min PC+α×RO

3rd step, sets up network resource overhead optimization mechanism, including signaling consumption optimization, the number of dropped packets of the single session of optimization With caching expense optimization, signaling consumption is optimized for：

C_u() is the login cost of certain network behavior, wherein t_sFor the average Connection Time of single session, t_ΓExist for packet The mean residence time of network, T_adFor the mean transit delay between agent node, N_hAveragely switch for Internet in single session Number of times, N_gFor the remaining root node number in δ domain, s_uFor the average bag size of signaling information, s is transmission in label switched path Packet mean size, h_x-yFor the average number of hops between destination node d in wire link, B_wFor the bandwidth of wire link, B_w1For The bandwidth of wireless link, L_w1For the time delay of wireless link, L_wFor the time delay of wire link, P_t() is path delay function, λ_dFor The weight coefficient of downlink packet, T_interFor the continuous time interval for reaching data parlor, T_cFor the average of single switching Time, it is multi-tag label switched domain that MN is mobile node, δ, and ξ is the next Network Access Point of mobile node, and κ is δ domain Root node, HA is the home agent node of user, and it is Mobile Access Gateway that LER is edge router, η, and ε is accessed for locally mobile End, ω be proxy mobile IPv 6, Δ be based on ω, the LN being switched fast for linkage node, γ be based on the movement being switched fast Agent node, FA is that Foreign Agent node, θ is the δ path based on low-power consumption link, and n and p is gateway identification, n η and p η difference For n-th and p Mobile Access Gateway, mobile node is transferred to δ domain from proxy mobile IPv 6 domain,_μFor edge router between domain, " " is that network or node are transferred to another Autonomous Domain, or the mark for being transferred to another state by a certain state by a certain Autonomous Domain Know, it is predecessor LER, ξ for current LER, τ for carrying out being switched fast δ domain for mobile δ, σ that α is mobile IP, β.

4th step, optimizes the number of dropped packets of single session, wherein P_loss() is certain network behavior for producing in single session Number of dropped packets,

5th step, caching expense optimization is specially：

B_size() is the nodal cache expense that certain network behavior is produced

The present invention proposes a kind of power control techniques based on dynamic logic topology, by by setting up power optimization mould Type and network resource overhead optimization mechanism, realize the elastic optimum management of self-organizing network energy.

Claims (6)

1. a kind of power control techniques based on dynamic logic topology, excellent by setting up power optimization model and network resource overhead Change mechanism, realizes the elastic optimum management of self-organizing network energy, comprises the steps：

A, set up power control mechanism；

B, network resource overhead optimization mechanism is set up, including signaling consumption optimization, optimize the number of dropped packets of single session and caching is opened Pin optimizes.

2. method according to claim 1, for step A it is characterized in that：Power control mechanism is：It is outer that a.MN monitors band Busy tone BT whether there is, if its presence, goes to step b, otherwise then go to step g；B. using powerSend RTS；C. destination node receives RTS, and judges whether to carry outer busy tone BT, if there is the outer busy tone BT of band, abandons RTS, instead Then go to step d；D.MN uses power P=P_maxCTS is sent, and judges network capacityWhether set up, if Set up and step e is then gone to, otherwise step b is then gone to, whereinEqual transmission path number for needed for packet is that single-hop is average Transmission range, C_fFor the capacity of channel f, p_fFor the use probability of channel f, it is path number that F is channel number, z；E. the section of mesh Point receives CTS, and uses powerSend traffic packets；F.MN uses power P=P_maxRTS is sent, and goes to step Rapid c, wherein MN are mobile terminal.

3. method according to claim 1, for step A it is characterized in that：Power Control Optimized model is specially：V is The node set of network, E is line set, λ^sdFor the business demand between source node s and destination node d, T=[λ^sd] save for source Business demand set of the point between s and destination node d, it is radio frequency set that C is link average size, W, C_epFor average road By capacity, T_iFor the maximum path number that the transmitting terminal of node i is supported, R_iFor the maximum path number that the receiving terminal of node i is supported, P_tr For the energy expenditure of transponder, P_osFor the energy expenditure of crosspoint, P_epEnergy in every Gbps business streaming Consume, PC is the energy expenditure of network, N '_ij,wIt is to make between source node i and destination node j under upper logical topology environment With the path number of frequency w,Be under upper logical topology environment between source node i and destination node j usage frequency w Path number, and between source node i and destination node j, there is physics direct connected link (m, n),For running on link l_ijOn Service traffics, N_ijFor the path number between source node i and destination node j, i, j are logical topology node, N_ij,wBe in source The path number of usage frequency w between node i and destination node j,It is usage frequency between source node i and destination node j W and the path number through (m, n), RO converts expense for logical topology, and it is frequency sets that E is physics direct connected link set, W,

min PC+α×RO

$\begin{array}{ccc}s.t.& \underset{w}{\Σ}{N}_{ij,w}=N_{ij}& \∀i,j\∈V\end{array}$

$\begin{array}{c}PC=P_{os}\×\underset{w}{\mathrm{\Σ}}\underset{i}{\mathrm{\Σ}}\underset{j,i\≠i}{\mathrm{\Σ}}(\underset{(m,n)}{\mathrm{\Σ}}{P}_{mn}^{ij,w}+N_{ij,w})+2\×P_{tr}\×\underset{i}{\Σ}\underset{j,j\≠i}{\mathrm{\Σ}}{N}_{ij}+P_{ep}\\ \×\underset{i}{\mathrm{\Σ}}\underset{j}{\mathrm{\Σ}}\underset{s,i\≠s}{\mathrm{\Σ}}\underset{d}{\mathrm{\Σ}}{f}_{ij}^{sd}\end{array}$

$RO=\underset{i}{\Σ}\underset{j}{\Σ}\underset{m}{\Σ}\underset{n}{\Σ}\underset{w}{\Σ}|P_{mn}^{ij,w}-P_{mn}^{\′ij,w}|$

$\underset{j}{\Σ}{f}_{ij}^{sd}-\underset{j}{\Σ}{f}_{ji}^{sd}=\left\{\begin{array}{cc}{\mathrm{\λ}}^{sd},& ifi=s\\ -{\mathrm{\λ}}^{sd},& ifi=d\\ 0,& otherwise\end{array}\right.,\∀i,s,d\∈V$

$\underset{n}{\Σ}{P}_{mn}^{ij,w}-\underset{n}{\Σ}{P}_{nm}^{ij,w}=\left\{\begin{array}{cc}{N}_{ij,w},& ifm=i\\ -N_{ij,w},& ifm=j\\ 0,& otherwise\end{array}\right.,\∀i,j,m\∈V,\∀w\∈W$

$\underset{s}{\Σ}\underset{d}{\Σ}{f}_{ij}^{sd}\≤C\×N_{ij},\∀i,j\∈V$

$\underset{j}{\mathrm{\Σ}}\underset{s,i\≠s}{\mathrm{\Σ}}\underset{d}{\mathrm{\Σ}}{f}_{ij}^{sd}\≤C_{ep},\∀i\∈V$

$\underset{w}{\mathrm{\Σ}}\underset{j}{\mathrm{\Σ}}{N}_{ij,w}\≤T_i,\∀i\∈V$

$\underset{w}{\mathrm{\Σ}}\underset{j}{\mathrm{\Σ}}{N}_{ji,w}\≤R_i,\∀i\∈V$

$\begin{array}{cc}\underset{i}{\Σ}\underset{j}{\Σ}{P}_{mn}^{ij,w}\≤1& \∀(m,n)\∈E,\∀w\∈W\end{array}.$

4. method according to claim 1, for step B it is characterized in that：Signaling consumption is optimized for：

C_u(α)=2s_uh_MN-HAN_h

C_u(β)=2s_uh_MN-HAN_h+2s_lh_FA-HAN_h

C_u(γ)=2s_uh_MN-κN_h+2s_uh_FA-FAN_h+2s_lh_FA-κN_hN_g

C_u(θ)=2s_uh_MN-μN_h+2s_lh_LN-ξN_h

C_u(ω)=2s_uh_nη-εN_h+2s_uh_nη-εN_h

C_u(Δ)=2s_uh_nη-pηN_h+2s_uh_nη-εN_h

C_u(ω-δ)=2s_uh_pη-εN_h+2s_lh_nη-εN_h

C_u() is the login cost of certain network behavior, wherein t_sFor the average Connection Time of single session, t_ΓFor packet in net The mean residence time of network, T_adFor the mean transit delay between agent node, N_hAveragely switch for Internet in single session secondary Number, N_gFor the remaining root node number in δ domain, s_uFor the average bag size of signaling information, s is the number for transmitting in label switched path According to bag mean size, h_x-yFor the average number of hops between destination node d in wire link, B_wFor the bandwidth of wire link, B_w1For no The bandwidth of wired link, L_w1For the time delay of wireless link, L_wFor the time delay of wire link, P_t() is path delay function, λ_dFor under The weight coefficient of downlink packet, T_interFor the continuous time interval for reaching data parlor, T_cMean time for single switching Between, it is multi-tag label switched domain that MN is mobile node, δ, and ξ is the next Network Access Point of mobile node, and κ is the root in δ domain Node, HA is the home agent node of user, and it is Mobile Access Gateway that LER is edge router, η, and ε is accessed for locally mobile End, ω be proxy mobile IPv 6, Δ be based on ω, the LN being switched fast for linkage node, γ be based on the movement being switched fast Agent node, FA is that Foreign Agent node, θ is the δ path based on low-power consumption link, and n and p is gateway identification, n η and p η difference For n-th and p Mobile Access Gateway, mobile node is transferred to δ domain from proxy mobile IPv 6 domain, and μ is edge router between domain, " " is that network or node are transferred to another Autonomous Domain, or the mark for being transferred to another state by a certain state by a certain Autonomous Domain Know, it is predecessor LER, ξ for current LER, τ for carrying out being switched fast δ domain for mobile δ, σ that α is mobile IP, β.

5. method according to claim 1, for step B it is characterized in that：Optimize the number of dropped packets of single session, wherein P_loss() is the number of dropped packets of certain network behavior for producing in single session,

$P_{loss}\left(\mathrm{\α}\right)=\[\left(\frac{1}{2}{T}_{ad}\right)+T_c\mathrm{\α}\]{\mathrm{\λ}}_d{N}_h$

$P_{loss}\left(\mathrm{\β}\right)=\[\left(\frac{1}{2}{T}_{ad}\right)+T_c\left(\mathrm{\β}\right)\]{\mathrm{\λ}}_d{N}_h$

P_loss(τ)=t (s_uh_MN-FA)λ_dN_h

P_loss(θ)=t (s_uh_MN-ξ)λ_dN_h

$P_{loss}\left(\mathrm{\ω}\right)=\[\left(\frac{1}{2}{T}_{ad}\right)+T_c\left(\mathrm{\ω}\right)\]{\mathrm{\λ}}_d{N}_h$

P_loss(Δ)=t (s_uh_MN-pη)λ_dN_h

$P_{loss}(\mathrm{\ω}-\mathrm{\δ})=\[\left(\frac{1}{2}{T}_{ad}\right)+T_c(\mathrm{\ω}-\mathrm{\δ})\]{\mathrm{\λ}}_d{N}_h$

6. method according to claim 1, for step B it is characterized in that：Caching expense optimization is specially：

B_size() is the nodal cache expense that certain network behavior is produced

$B_{size}\left(\mathrm{\τ}\right)=\left(\frac{1}{2}{T}_{ad}\right)+t(s_u,h_{MN-FA}+h_{FA-FA}){\mathrm{\λ}}_d$

$B_{size}\left(\mathrm{\θ}\right)=\left(\frac{1}{2}{T}_{ad}\right)+t(s_u,h_{LN-\mathrm{\σ}}+h_{\mathrm{\σ}-LN}){\mathrm{\λ}}_d$

$B_{size}\left(\mathrm{\Δ}\right)=\left(\frac{1}{2}{T}_{ad}\right)+t(s_u,h_{MN-p\mathrm{\η}}+h_{p\mathrm{\η}-n\mathrm{\η}}){\mathrm{\λ}}_d.$

$t(s,h_{x-y})=c+h_{x-y}(\frac{s}{B_w}+L_w)+P_t(h_{x-y}+1)$