CN103685025A - Cross-layer dynamic self-adapting routing method based on LEO satellite network - Google Patents

Abstract

The invention relates to a cross-layer dynamic self-adapting routing method based on an LEO satellite network. According to the cross-layer dynamic self-adapting routing method based on the LEO satellite network, through overall consideration of the transmission delay, the network link utilization rate and the network link quality of a data packet, the link weight is defined, an optimization model is established, and solving is conducted on the optimization model based on an ant colony algorithm. According to concrete realization of the cross-layer dynamic self-adapting routing method, a mobile agent is used for collecting routing information autonomously, a routing decision can be made in the stage of route establishment through the information of a physical layer, and an ant routing probability value is changed dynamically according to the link quality of the satellite network. By the adoption of the cross-layer dynamic self-adapting routing method, under the condition that the transmission delay and the bandwidth utilization rate of the data packet of the satellite network are guaranteed, data packet dropout caused by deterioration of a link state is avoided, and the robustness of the network is effectively guaranteed.

Description

A kind of cross-layer dynamic self-adapting method for routing based on LEO satellite network

Technical field

The present invention relates to a kind of cross-layer dynamic self-adapting method for routing based on LEO satellite network, belong to satellite communication field.

Background technology

The all-IP of Network turns to Low Earth Orbit (low earth orbit, LEO) satellite communication system new chance is provided.LEO satellite network can provide Global coverage, and can under certain quality of service guarantee, provide wider bandwidth and lower data transfer delay.Therefore, by LEO satellite network, carry out the extensive concern that transfer of data has caused the world.How to design method for routing efficiently, reliably and flexibly and be one of difficult point that LEO satellite network need to solve.

According to the producing method of route, satellite network routing mechanism can be divided into two kinds of static routing and dynamic routings.Static routing mechanism is utilized periodicity and the predictability of satellite network operation, carries out the transmission of packet according to the routing iinformation of calculated in advance, has and realizes simply, the advantage that routing cost is little.But static routing mechanism can not adapt to the variation of inter-satellite link, offered load, thereby be difficult to find optimum path may cause the remarkable decline of network performance.Dynamic routing, according to the variation of network traffics and link delay, is adjusted path adaptively, thereby has guaranteed the high efficiency of package forward.

Recent years, more and more extensive for the demand of satellite network transmitting multimedia service.Thus, scholars turn to the research based on internet service application.And, in wireless sensor network, introduced the design philosophy of cross-layer routing, multiple target service quality (Quality of service, the QoS) cross-layer optimization model that is applied to low-track satellite network had not had open source literature report.

Its core concept of cross-layer structure that the present invention carries is to carry out routing decision by collecting the information auxiliary network layer of physical layer.And the present invention guarantees link utilization by setting up multiple target QoS Optimized model when realizing load balancing.

Summary of the invention

Technical problem to be solved by this invention: overcome the deficiencies in the prior art, a kind of cross-layer dynamic self-adapting method for routing based on LEO satellite network has been proposed, by collecting the information auxiliary network layer of physical layer, carry out routing decision, set up multiple target QoS Optimized model and when realizing load balancing, guarantee link utilization

Technical solution of the present invention: a kind of cross-layer dynamic self-adapting method for routing based on LEO satellite network, described step is as follows:

A, according to chain-circuit time delay and the link remaining bandwidth of physical layer, link cost weights are defined, the computing formula of link cost weights cost (i, j) is shown below:

$\cos t(i,j)={\mathrm{\ω}}_1\×\frac{\mathrm{PD}{(i,j)}_{\min }}{\mathrm{PD}(i,j)}+{\mathrm{\ω}}_2\×\frac{\mathrm{band}(i,j)}{\mathrm{band}{(i,j)}_{\max }}$

I, j represents satellite node, PD (i, j) represents the time delay that current LEO satellite network link data bag is propagated, PD (i, j) _minrepresent the minimum value of LEO satellite network link data bag propagation delay; Equally, band (i, j) represents the remaining bandwidth of current LEO satellite network link, band (i, j) _maxrepresent the maximum of LEO satellite network link remaining bandwidth; ω ₁and ω ₂respectively corresponding weights, according to the feature of LEO satellite network, PD (i, j) _minand band (i, j) _maxdetermine accordingly.

B, according to the link data bag propagation delay of link cost weights, physical layer, these four channel condition informations of the remaining bandwidth of link and link error rates, build Optimum Theory model, the theoretical model of structure as shown in the formula:

$\max \underset{(i,j)\∈P(\mathrm{src},\mathrm{des})}{\overset{\mathrm{des}}{\mathrm{\Σ}}}\cos t(i,j)$

Meet $\underset{(i,j)\∈P(\mathrm{src},\mathrm{des})}{\overset{\mathrm{des}}{\mathrm{\Σ}}}\mathrm{delay}(i,j)\≤\mathrm{De}$

$\underset{(i,j)\∈P(\mathrm{src},\mathrm{des})}{\min }\mathrm{bw}(i,j)\≥B$

$\∀\underset{(i,j)\∈P(\mathrm{src},\mathrm{des})}{\mathrm{link}(i,j)}\mathrm{Pe}(i,j)\≤{10}^{-6}$

Two satellite nodes in i and j delegated path P (Src, Des), delay (i, j) has represented the time delay of link (i, j), De is the patient maximum delay of network; Bw (i, j) has represented the bandwidth of link (i, j), and B is the minimum bandwidth constraint in network.P _e(i, j) refers to the error rate of link (i, j), and its value has reflected the quality of satellite network wireless link, P _evalue need to meet P _e≤ 10 ^-6.

Ant group method dynamic collection link data bag propagation delay, link remaining bandwidth and these state informations of link error rates of C, employing mobile agent mode, carry out Path selection decision-making according to this information, so that Optimized model is solved;

D, the path situation obtaining according to the mobile agent decision-making of constantly updating, build pheromones weighted model, by the value of parameter a, b is derived, assurance mobile agent selects link cost to strengthen link all the time, agency is after link, and the pheromones on this link is upgraded by following formula:

y=ax+b a∈[0,1] (9)

E, repetition A, B, C step, constantly update time delay, link remaining bandwidth, link error rates information that link data bag is propagated, with continuous adaptive optimization path.

According to the feature of LEO satellite network, the time delay PD (i, j) that link data bag is propagated _minwith link remaining bandwidth band (i, j) _maxcan determine accordingly, as follows:

A1: 60 °, above and south latitude is decided to be region, polar region with lower area by 60 ° of north latitude, according to the specific minimum value that can calculate chain circuit transmission time delay between satellite orbit of constellation structures;

A2: according to the efficiency of transmission of LEO satellite network link bandwidth value and packet, can calculate the maximum of remaining bandwidth in network.

Described in step C in specific implementation process, used the ant group algorithm in mobile proxy system, by the agency to different, coordinate to solve routing issue, in ant group system, defined the ant of two types, Front ant and Back ant, in network along the path from source node to destination node from a node motion to next node, in mobile process, find available path and also collect some Useful Informations; Mobile agent is designed to Front ant and Back ant, and the routing procedure that finally completes whole network comprises:

C1, every Fixed Time Interval, Front ant sends and shifts to object satellite from source satellite node.Because the topological structure of low-track satellite network has regularity, therefore for source satellite node, the down hop satellite selecteed probability nearer with object satellite distance is larger, for intermediate, satellite node, Front ant is used pseudorandom proportionality principle, be that random selecting and confirming is selected the method combining, to being positioned at the Front ant of node, select satellite node;

C2, Front ant are understood auto-destruct after arriving at object satellite node, destination node can generate backward agency simultaneously, Back ant can copy the two class lists that comprise in Front ant, and the direction while coming along Front ant is returned to source node in opposite direction, when through intermediate, satellite, intermediate, satellite can be upgraded the value in probability tables automatically.Back ant upgrades the routing probability tables in satellite according to the Link State of satellite network after arriving intermediate, satellite;

C3, when packet is from source satellite node sends and arrive intermediate, satellite, the probability tables that intermediate, satellite can be stored according to this satellite forwards packet.

Mobile proxy system described in step C can be a plurality of be multi-agent system, multi-agent system coordinates to solve routing issue by the agency to different.

Accompanying drawing explanation

Fig. 1 is cross-layer model framework of the present invention;

Fig. 2 is the dynamic self-adapting method for routing based on agency of the present invention.

Embodiment

Basic ideas of the present invention are to carry out routing decision by collecting the information auxiliary network layer of physical layer, and when realizing load balancing, guarantee link utilization by setting up multiple target QoS Optimized model.A kind of cross-layer dynamic self-adapting method for routing based on LEO satellite of the present invention, the method comprises the steps:

A, consider on the basis of time delay and bandwidth considerations, link cost weights are defined;

B, in conjunction with the channel condition information of physical layer, build Optimum Theory model;

C, utilize ant group algorithm to solve Optimized model, adopt mobile agent mode dynamic collection network link status information, according to this information, carry out Path selection decision-making;

D, for pheromones weighting parameters in method, set, guarantee that mobile agent selects link cost to strengthen link all the time, effectively avoid the path that Link State worsens, improve the robustness of satellite network data packet transmission.

In said method, steps A is classified port status, and its concrete steps comprise:

Link transmission time delay and bandwidth information are considered, the link cost value of satellite network is defined, shown in (1):

$\cos t(i,j)={\mathrm{\ω}}_1\×\frac{\mathrm{PD}{(i,j)}_{\min }}{\mathrm{PD}(i,j)}+{\mathrm{\ω}}_2\×\frac{\mathrm{band}(i,j)}{\mathrm{band}{(i,j)}_{\max }}---\left(1\right)$

In above formula (1), i, j represents satellite node, PD (i, j) represents the time delay that current LEO satellite network link data bag is propagated, PD (i, j) _minrepresent the minimum value of LEO satellite network link data bag propagation delay; Equally, band (i, j) represents the remaining bandwidth of current LEO satellite network link, band (i, j) _maxrepresent the maximum of LEO satellite network link remaining bandwidth; ω ₁and ω ₂respectively corresponding weights.According to the feature of LEO satellite network, PD (i, j) _minand band (i, j) _maxcan determine accordingly.Method definite in the present invention is as follows:

A1: 60 °, above and south latitude is decided to be region, polar region with lower area by 60 ° of north latitude, according to the specific minimum value that can calculate chain circuit transmission time delay between satellite orbit of constellation structures;

A2: according to the efficiency of transmission of LEO satellite network link bandwidth value and packet, can calculate the maximum of remaining bandwidth in network.

In said method, step B, after having considered the link-state information of physical layer, can carry out link transmission time delay, network link remaining bandwidth and the network link layer error rate comprehensively, builds the Optimum Theory model guaranteeing based on QoS.Cross-layer model framework as shown in Figure 1.The theoretical model building, as shown in the formula shown in (2), is the multi-objective optimization question meeting under three inequality conditions.

$\max \underset{(i,j)\∈P(\mathrm{src},\mathrm{des})}{\overset{\mathrm{des}}{\mathrm{\Σ}}}\cos t(i,j)$

Meet $\underset{(i,j)\∈P(\mathrm{src},\mathrm{des})}{\overset{\mathrm{des}}{\mathrm{\Σ}}}\mathrm{delay}(i,j)\≤\mathrm{De}---\left(2\right)$

$\underset{(i,j)\∈P(\mathrm{src},\mathrm{des})}{\min }\mathrm{bw}(i,j)\≥B$

$\∀\underset{(i,j)\∈P(\mathrm{src},\mathrm{des})}{\mathrm{link}(i,j)}\mathrm{Pe}(i,j)\≤{10}^{-6}$

As shown in above formula (2), two satellite nodes in i and j delegated path P (Src, Des), delay (i, j) has represented the time delay of link (i, j), De is the patient maximum delay of network; Bw (i, j) has represented the bandwidth of link (i, j), and B is the minimum bandwidth constraint in network.P _e(i, j) refers to the error rate of link (i, j), and its value has reflected the quality of satellite network wireless link.P _evalue need to meet P _e≤ 10 ^-6.

In said method, step C, in specific implementation process, has introduced agency's concept.Multi-agent system is one of research direction of artificial intelligence.Multi-agent system coordinates to solve routing issue by the agency to different.Ant, Front ant and the Back ant of two types in ant group system, have been defined.In the present invention, mobile agent is carried out the function similar to ant in ant group system, and in network, it can be along the path from source node to destination node from a node motion to next node.In mobile process, it can find available path and some other Useful Information.As shown in Figure 2, in each satellite node, there are five queues in method key diagram.Queue 5 is responsible for depositing and is sent to the packet that He Cong ground station of ground station sends.Meanwhile, in two tracks, link and two interorbital links also have queue and its binding.P _evalue computing module is responsible for cognitive radio Link State and is calculated the error rate of every link, and result of calculation is as the input of ant group algorithm operation module.Ant group algorithm operation module can calculate the probability that packet is sent to four adjacent satellite, and probable value is stored in the probability tables of satellite node.In institute of the present invention extracting method, there are two class mobile agents: forward direction agency and reverse proxy.Forward direction agency runs on responsible collection routing information reverse proxy in satellite network and is responsible for probability tables to upgrade.Statistical parameter collection module is responsible for the simulation parameters such as end-to-end time delay and receiving terminal throughput to collect.

C1, every Fixed Time Interval Δ t, forward direction is acted on behalf of F _{s → d}from source satellite node s, send and shift to object satellite d.Because the topological structure of low-track satellite network has regularity, therefore for source satellite node, the down hop satellite selecteed probability nearer with object satellite distance is larger.When forward direction agency sends, source satellite node selects the probability of next-hop node to calculate according to formula (3).

${\left(P_{\mathrm{sj}}\right)}_{\mathrm{agent}}=\frac{\frac{1}{{\mathrm{Hop}}_{\mathrm{jd}}}}{\underset{j\∈M}{\mathrm{\Σ}}\frac{1}{{\mathrm{Hop}}_{\mathrm{jd}}}}---\left(3\right)$

${\mathrm{Hop}}_{\mathrm{jd}}=\left\{\begin{array}{cc}{\mathrm{Hop}}_{\mathrm{jd}}& \mathrm{if}(j\≠d)\\ 0.001& \mathrm{else}\end{array}\right.$

In formula (3), (P _sj) _agentrepresent the probability of forward direction agent selection next-hop node j, Hop _jdthe jumping figure minimum value of representative from satellite node j to object satellite d, M is the set of next-hop node.

Each forward direction agency has two class lists.List

preserve forward direction agency No. ID, the intermediate node of process, and list

recorded the moment of forward direction agency through intermediate, satellite.

For intermediate, satellite node i, forward direction agency uses pseudorandom proportionality principle, and random selecting and confirming is selected the method combining, and to being positioned at the forward direction of node i, acts on behalf of k according to following formula selection satellite node j.

${\left(p_{\mathrm{ij}}\right)}_{\mathrm{agent}}=\left\{\begin{array}{cc}\frac{p_{\mathrm{ij}}}{\underset{l\∈{\mathrm{table}}_k}{\mathrm{\Σ}}{p}_{\mathrm{ij}}}& (\mathrm{ifq}\≤q_0)\\ 1/N& (\mathrm{ifq}>q_0)\end{array}\right.---\left(4\right)$

Here q is the random number in interval (0,1), and q ₀it is the constant being positioned in interval (0,1).Q ₀size reflected the importance of utilizing priori to select new node.Table _krepresented that forward direction acts on behalf of the next-hop node set that k can select, N represents table _kthe number of middle element.Table is set in method _kobject be to prevent route loop.As previously mentioned, in institute of the present invention extracting method, the performed function of probability tables is similar to routing table, therefore for (4) formula, p _ijrepresentative data bag is selected the probability of down hop.

C2, forward direction are acted on behalf of F _{s → d}after arriving at object satellite node, understand auto-destruct, destination node can generate the backward B of agency simultaneously _{d → s}.B _{d → s}can copy the two class lists that comprise in forward direction agency, and along list in node return to source node in the other direction.Work as B _{d → s}during through intermediate, satellite, intermediate, satellite can be upgraded the value in probability tables automatically.

In Fig. 2, suppose that backward agency moves to satellite node 2 from satellite node 1.Numbering for 1, four link of satellite node is consistent with the numbering of four inter-satellite link ports.Suppose that node j is the abutment points of satellite node 1, set K meets condition below.

K={j,(Pe) _ij<10 ^-6}∩{j≠2∩j∈M} (5)

l∈K∩cost _il=max{cost _ij,j∈K} (6)

Wherein gather K and represented the node set except node 2 that meets error rate requirement, l has represented the next-hop node of set K link cost maximum.

Packet selects the probable value of down hop satellite node to calculate by following formula (7) and (8).

${\left(P_{i2}\right)}_{\mathrm{data}}(t+1)=\left\{\begin{array}{cc}\mathrm{\&rho;}\&times;{\left(P_{i2}\right)}_{\mathrm{data}}\left(t\right)+(1-\mathrm{\&rho;})& {\left(\mathrm{Pe}\right)}_{i2}<{10}^{-6}\\ \mathrm{\&rho;}\&times;{\left(P_{i2}\right)}_{\mathrm{data}}\left(t\right)& {\left(\mathrm{Pe}\right)}_{i2}\&GreaterEqual;{10}^{-6}\end{array}\right.---\left(7\right)$

${\left(\begin{array}{c}{P}_{\mathrm{ij}}\\ j\&NotEqual;2\end{array}\right)}_{\mathrm{data}}(t+1)=\left\{\begin{array}{cc}\mathrm{\&rho;}\&times;{\left(P_{\mathrm{il}}\right)}_{\mathrm{data}}\left(t\right)+(1-\mathrm{\&rho;})& {\left(\mathrm{Pe}\right)}_{i2}\&GreaterEqual;{10}^{-6}\&cap;j=l\\ \mathrm{\&rho;}\&times;{\left(P_{\mathrm{ij}}\right)}_{\mathrm{data}}\left(t\right)& {\left(\mathrm{Pe}\right)}_{i2}\&GreaterEqual;{10}^{-6}\&cap;j\&Element;K\&cap;{\cos t}_{\mathrm{ij}}\&NotEqual;\max \{{\cos t}_{\mathrm{ik}},k\&Element;K\}\\ \mathrm{\&rho;}\&times;{\left(P_{\mathrm{ij}}\right)}_{\mathrm{data}}\left(t\right)& {\left(\mathrm{Pe}\right)}_{i2}\&GreaterEqual;{10}^{-6}\&cap;j\&NotElement;K\\ \mathrm{\&rho;}\&times;{\left(P_{\mathrm{ij}}\right)}_{\mathrm{data}}\left(t\right)& {\left(\mathrm{Pe}\right)}_{i2}<{10}^{-6}\end{array}\right.---\left(8\right)$

In formula (7) and (8), (P _i2) _datarepresentative data bag selects node 2 as the probability of next-hop node, ${\left(\begin{array}{c}{P}_{\mathrm{ij}}\\ j\&NotEqual;2\end{array}\right)}_{\mathrm{data}}$ Represented that it is not node 2 that packet is selected node j() as the probability of next-hop node, ρ representative information element volatility coefficient.

In said method, the link of step D in order to make packet select all the time probability to be strengthened in transmitting procedure, the present invention sets the span of pheromones weighting parameters in method, thus the robustness of Enhancement Method.In the method, after acting on behalf of k process link (i, j), the pheromones on this link is upgraded by following formula:

y=ax+b a∈[0,1] (9)

Wherein, a and b are constants, by theory, derive and emulation, and when the value of a meets a≤0.5, the link that packet can select probable value to be strengthened all the time.

In sum, in the present invention, by considering transmission delay, network link utilance and the network link quality of packet, definition link metric, builds Optimized model.By ant group algorithm, this Optimized model is solved.In specific implementation, the present invention uses the autonomous collect & route information of mobile agent, at path establishment stage, can utilize the information of physical layer to carry out path decision, according to satellite network link-quality, dynamically changes the size of acting on behalf of routing probable value.Adopt the present invention, can, when guaranteeing satellite network data packet transmission time delay and bandwidth availability ratio, avoid, because Link State worsens the data-bag lost causing, effectively guaranteeing the robustness of network.Meanwhile, the pheromones weighting parameters in method has been carried out to value setting, the link that makes packet select all the time probable value to be strengthened, the robustness of raising satellite network when carrying out path decision.

Claims (4)

1. the cross-layer dynamic self-adapting method for routing based on LEO satellite network, is characterized in that described step is as follows:

A, according to chain-circuit time delay and the link remaining bandwidth of physical layer, link cost weights are defined, the computing formula of link cost weights cost (i, j) is as follows:

$\cos t(i,j)={\mathrm{\&omega;}}_1\&times;\frac{\mathrm{PD}{(i,j)}_{\min }}{\mathrm{PD}(i,j)}+{\mathrm{\&omega;}}_2\&times;\frac{\mathrm{band}(i,j)}{\mathrm{band}{(i,j)}_{\max }}$

I, j represents satellite node, PD (i, j) represents the time delay that current LEO satellite network link data bag is propagated, PD (i, j) _minthe minimum value that represents LEO satellite network link data bag propagation delay, band (i, j) represents the remaining bandwidth of current LEO satellite network link, band (i, j) _maxrepresent the maximum of LEO satellite network link remaining bandwidth, ω ₁and ω ₂respectively corresponding weights, according to the feature of LEO satellite network, PD (i, j) _minand band (i, j) _maxdetermine accordingly;

B, according to the link data bag propagation delay of link cost weights, physical layer, these four channel condition informations of the remaining bandwidth of link and link error rates, build Optimum Theory model, the theoretical model of structure as shown in the formula:

$\max \underset{(i,j)\&Element;P(\mathrm{src},\mathrm{des})}{\overset{\mathrm{des}}{\mathrm{\&Sigma;}}}\cos t(i,j)$

Meet $\underset{(i,j)\&Element;P(\mathrm{src},\mathrm{des})}{\overset{\mathrm{des}}{\mathrm{\&Sigma;}}}\mathrm{delay}(i,j)\&le;\mathrm{De}$

$\underset{(i,j)\&Element;P(\mathrm{src},\mathrm{des})}{\min }\mathrm{bw}(i,j)\&GreaterEqual;B$

$\&ForAll;\underset{(i,j)\&Element;P(\mathrm{src},\mathrm{des})}{\mathrm{link}(i,j)}\mathrm{Pe}(i,j)\&le;{10}^{-6}$

Two satellite nodes in i and j delegated path P (Src, Des), delay (i, j) has represented the time delay of link (i, j), De is the patient maximum delay of network; Bw (i, j) has represented the bandwidth of link (i, j), and B is the minimum bandwidth constraint in network.P _e(i, j) refers to the error rate of link (i, j), and its value has reflected the quality of satellite network wireless link, P _evalue need to meet P _e≤ 10 ^-6.

Ant group method dynamic collection link data bag propagation delay, link remaining bandwidth and these state informations of link error rates of C, employing mobile agent mode, carry out Path selection decision-making according to this information, so that Optimized model is solved;

D, the path situation obtaining according to the mobile agent decision-making of constantly updating, build pheromones weighted model, by the value of parameter a, b is derived, assurance mobile agent selects link cost to strengthen link all the time, agency is after link, and the pheromones on this link is upgraded by following formula:

y=ax+b a∈[0,1] (9)

E, repetition A, B, C step, constantly update time delay, link remaining bandwidth, link error rates information that link data bag is propagated, with continuous adaptive optimization path.

2. method according to claim 1, is characterized in that: according to the feature of LEO satellite network, and the time delay PD (i, j) that link data bag is propagated _minwith link remaining bandwidth band (i, j) _maxcan determine accordingly, as follows:

A1: 60 °, above and south latitude is decided to be region, polar region with lower area by 60 ° of north latitude, goes out the minimum value of chain circuit transmission time delay between satellite orbit according to constellation structures specific calculation;

A2: according to the efficiency of transmission of LEO satellite network link bandwidth value and packet, calculate the maximum of remaining bandwidth in network.

3. method according to claim 1, it is characterized in that: described in step C in specific implementation process, used the ant group algorithm in mobile proxy system, by the agency to different, coordinate to solve routing issue, in ant group system, defined the ant of two types, Front ant and Back ant, in network along the path from source node to destination node from a node motion to next node, in mobile process, find available path and collect some Useful Informations; Mobile agent is designed to Front ant and Back ant, and the routing procedure that finally completes whole network comprises:

C1, every Fixed Time Interval, Front ant sends and shifts to object satellite from source satellite node.Because the topological structure of low-track satellite network has regularity, therefore for source satellite node, the down hop satellite selecteed probability nearer with object satellite distance is larger, for intermediate, satellite node, Front ant is used pseudorandom proportionality principle, be that random selecting and confirming is selected the method combining, to being positioned at the Front ant of node, select satellite node;

C2, Front ant are understood auto-destruct after arriving at object satellite node, destination node can generate backward agency simultaneously, Back ant can copy the two class lists that comprise in Front ant, and the direction while coming along Front ant is returned to source node in opposite direction, when through intermediate, satellite, intermediate, satellite can be upgraded the value in probability tables automatically.Back ant upgrades the routing probability tables in satellite according to the Link State of satellite network after arriving intermediate, satellite;

C3, when packet is from source satellite node sends and arrive intermediate, satellite, the probability tables that intermediate, satellite can be stored according to this satellite forwards packet.

4. method according to claim 3, is characterized in that: described mobile agent be a plurality of be multi-agent system, multi-agent system coordinates to solve routing issue by the agency to different.

Images