CN104244356A - Orientation ant colony route optimization method based on evolution graph full route forecasting - Google Patents

Abstract

The invention relates to an orientation ant colony route optimization method based on evolution graph full route forecasting. The orientation ant colony route optimization method is mainly characterized in that on an application layer, a control center or a GPS provides a node track, link dispatching information or node positions and motion speed and direction information and sends the node track, the link dispatching information or the node positions and the motion speed and direction information to a network layer; on the network layer, an evolution graph model is firstly built by nodes according to information of the application layer, then full route forecasting is carried out, finally the orientation ant colony route optimization method is used for selecting an optimal route, and the data are sent; on a physical layer, link delay, data transmission rate and available bandwidth information is collected. The orientation ant colony route optimization method is reasonable in design, orientation ant colonies are used, slow convergence caused by frequent network topology changes is avoided, a routing algorithm meeting the QoS requirement is provided, and performance indexes are obviously improved compared with a traditional mobile Ad Hoc network DSR and AODV routing algorithm.

Description

A kind of based on evolution diagram system-wide by the directed Ant Routing optimization method predicted

Technical field

The invention belongs to mobile ad hoc network route technology field, especially a kind of based on evolution diagram system-wide by the directed Ant Routing optimization method predicted.

Background technology

Flourish along with wireless communication technology and the Internet, network becomes obtaining information in people's daily life, the requisite infrastructure of exchange of information, and people more and more need the infrastructure providing Internet access service in any place any time.The Internet overall situation building global seamless access is the inexorable trend of future network development, a very important link in the military strategy deployment of Ye Shige great powers in the world.Because earth surface 70% is ocean, terrestrial cellular mobile communication cannot be set up, and satellite communication greatly, is not subject to the innate advantages such as geographical conditions restriction owing to having area coverage, is subject to people's attention in the process of Communication Development always.

By today from 20 century 70s, built up can be completely covering the whole world satellite navigation system only have the global positioning system of the U.S. to unify the glonass system of the former Soviet Union.The Beidou satellite navigation system of China, the GALILEO positioning system of European Union are built.Other country, comprises France, Japan and India all in Development area navigation system.Therefore navigation satellite mechanics of communication is the focus that scholars studies all the time, and this is comprising the routing solution communicated between star.

Due to features such as satellite cost is high, number of nodes is few, resource-constraineds, its routing solution and ground are very different.In the communication of multilayer orbiting satellite system, the link duration between low-orbit satellite is short, and switching frequency is high, for routing algorithm brings difficulty.But joint movements track, along orbital period circulation, artificially can be controlled inter-node link and when setting up and switching, so virtual dynamic network is suggested for description node high-speed motion and has periodic network topology structure.

Virtual dynamic network is thought within very little time period, network topology structure is changeless, artificially the satellite earth cycle of one week can be divided into some time slots, link annexation is changeless in same time slot, different between different time-gap.Therefore, the virtual dynamic network that is made up of several static topological of whole satellite network.The object done like this is applied in satellite network by traditional ground routing algorithm, its drawback is, calculate a fixing routing table for each time slot of each node in advance and be configured on satellite node, its drawback is that aggregate network throughput is low, very flexible during reply emergency.Again because satellite link is short for effective time, switch frequent, adopt existing mobile ad hoc network route technology, will increase the weight of node calculate burden, can not communication node many to quantity.

In order to overcome the shortcoming of above routing algorithm, within 2002, French scholar proposes evolution diagram model at first.The scheduling of this model analysis inter-satellite link switches the network topology structure of the constellation dynamic change caused, on virtual dynamic model basis modeling.The routing algorithm arriving path, the shortest jumping figure, most short time-delay the earliest based on evolution diagram model was proposed in recent years successively.At given node link scheduling scheme, simplation verification is carried out to algorithm, the average throughput obtained, average packet loss ratio, average delay performance are all better than wireless self-networking plan range vector Routing Protocol (Ad hoc on-demand distance vector routing as required, and dynamic source routing protocol (Dynamic Source Routing, DSR) AODV).Within 2013, have scholar by evolution diagram model use in vehicle-mounted mobile MANET, propose the reliable routing strategy based on evolution diagram, comparatively dynamic source routing protocol performance increases.

Along with the development of satellite radio transmit-receive technology, realize the development trend that QoS routing will be Future Satellite communication, and the existing routing algorithm based on evolution diagram model only can go out unique optimum route for different index prediction, the demand of route technology development can not be met.

In addition, in artificial intelligence field, often ant group algorithm is used.Ant group algorithm is a kind of probability type algorithm being used for finding in the drawings path optimizing, and it is proposed in thesis for the doctorate at him in 1992 by Marco Dorigo.Be that the application of the distributed artificial intelligence research of representative becomes a study hotspot in recent years with ant group algorithm, many algorithms coming from bee colony and Ants model design oneself be applied to the research in the field such as operation mode of routing algorithm, coloring problem, vehicle scheduling and enterprise more and more, it is for solve and multi-objective optimization problem provides one machine-processed simply and effectively.Intelligent algorithm is applied in ground moving MANET by current research majority, and not yet propose the solution for virtual dynamic topology, because such network link frequently switches, traditional ant group algorithm exists serious slow convergence problem.Traditional algorithm is applied to node under the environment of network topology the unknown, and each step of ant group shifts the transition probability depending on node calculate and go out, and therefore the computing cost of Nodes is comparatively large, is not suitable for the situation that satellite computational resource is limited.

In sum, it is the static routing table of satellite node configuration in advance that the routing mechanism communicated between existing star depends on ground control centre, and only can realize simple information receiving and transmitting, does not possess dynamic routing function, cannot realize and terrestrial interconnection net real time communication.Between star, how in communication, to provide efficient, dynamic, adaptive route technology to be realize future by problem in the urgent need to address in global satellite system and internet seamless access technology.

Summary of the invention

The object of the invention is to overcome the deficiencies in the prior art, provide a kind of reasonable in design, can communicate between star in provide efficient, dynamic, adaptive routing function based on evolution diagram system-wide by the directed Ant Routing optimization method predicted.

The present invention solves existing technical problem and takes following technical scheme to realize:

Based on evolution diagram system-wide by the directed Ant Routing optimization method predicted, comprise the following steps:

Step 1, in application layer, control centre or GPS global positioning system provide node track, link scheduling information or node location, movement velocity and directional information and are sent to network layer;

Step 2, in network layer, first node sets up evolution diagram model according to application layer message, then carries out system-wide by predicting, finally adopts directed Ant Routing optimization method select optimum route and send data;

Step 3, in physical layer, contribution link time delay, message transmission rate and available bandwidth information.

And, described step 2 system-wide is by predicting that the system-wide of employing based on evolution diagram model is by Forecasting Methodology, this system-wide comprises by the input content of Forecasting Methodology the maximum delay threshold value that evolution diagram model, source node, destination node and system allow, output content is that source node arrives whole route of destination node and stored in routing table, concrete Forecasting Methodology comprises the following steps:

Step (1), by stacked for the summit of access, whether search stack top node has the adjacent vertex of not accessing, and has, carries out step (2); Then do not eject stack top summit, until stack is empty, terminate algorithm;

Step (2), judge that whether the significant instant on limit between stack top joint and adjacent vertex meets and be greater than current time and the condition being less than maximum delay threshold value, if meet, adjacent vertex is stacked, upgrading current time is the moment the earliest that data send along this limit, carries out step (3); If do not meet, mark adjacent vertex, for accessing, continues step (2);

Step (3), judge whether adjacent vertex is destination node, if then whole node in output stack in order.

And the structure of described routing table comprises destination node, route, route break moment and select probability.

And, the directed Ant Routing optimization method of described step 2 is: ant group configuration fixed route, ant group is divided into Front ant and reverse ant two type, Front ant is according to the network communication quality parameter at set route collector node place, and these parameters are reflected in pheromones size are retained in forward node place, pheromones can be passed in time and reduce; Finally, reverse ant is collected the residual, information element at forward node place according to set route and feeds back to summit, source, and summit, source calculates the select probability of the route that reverse ant is carried according to residual, information element summation.

And the specific implementation step of described directed Ant Routing optimization method comprises:

Step (1): the Front ant carrying this routing iinformation after summit, source equiprobability selects a route from routing table by the release of certain hour interval;

Step (2): Front ant often arrives a network node, first judges whether to have exceeded the existence life-span, then judges whether current ink degree of Congestion satisfies condition; When above-mentioned two conditions all meet, this Nodes discharge a certain amount of pheromones and the routing forwarding of being carried according to Front ant by present node to next node;

Step (3): each node, when receiving Front ant, first judges it oneself is whether the destination node of Front ant; If not, then in the data structure of Front ant, write the moment arriving present node, the pheromones of encoding stored in this Front ant and discharging in local information element table, then Front ant is forwarded to next node; If self be the destination node of Front ant, then copy the route of Front ant, generate the reverse ant of identical ant numbering and discharge;

Step (4): oppositely ant returns source node along identical route, often an arrival forward node place collects identical ant and numbers the pheromones residual volume stayed, stored in reverse ant packet;

Step (5): after source node receives reverse ant, according to the pheromones calculation of total Route Selection probability carried, restores in routing table, and finally, node sends data according to the route that route select probability in routing table is maximum.

And described Front ant is a packet, comprising: ant numbering, ant type, date of birth, life-span and maximum link degree of Congestion threshold value and routing iinformation; Each routing node of described Front ant comprises information to be had: node serial number, the time of advent, transmitting time, current bandwidth and current data rate; Described reverse ant is a packet, comprising: ant numbering, ant type, date of birth, life-span and routing iinformation; Each routing node of reverse ant comprises information to be had: node serial number and node residual, information element.

And described link congestion degree is calculated by following formula:

η _i,j＝r _i,j/c _i,j

Wherein, η _i,jthe congestion ratio of link between node i and j, r _i,jthe data rate of link between node i and j, c _i,jit is the capacity of link between node i and j.

And described pheromones is that Front ant often discharges through a forward node, and the size of pheromones reflects network communication quality, and pheromone release amount is calculated by following formula:

$t_k^{\mathrm{ij}}=\frac{1}{{\left(d_{\mathrm{ik}}\right)}^{\mathrm{\α}}{\mathrm{\Σ}}_i^k{\left({\mathrm{\η}}_{m,m+1}\right)}^{\mathrm{\β}}}$

In formula, ij represents the route of source node i to destination node j; represent therefrom in the pheromones of forward node k place release, d _ikrepresent the time delay from i to k, η _{m, m+1}represent the link congestion degree between node m and m+1; α and β is all regulation coefficient, and what be used for regulating time delay and link congestion degree two characteristics on pheromones affects size, α and β all gets 0.5.

And described residual, information element is the pheromones of Nodes through certain hour remaining after evaporation, this residual, information element is upgraded and is calculated by following formula:

$T^{\mathrm{ij}}\←{\mathrm{\μT}}^{\mathrm{ij}}+(1-\mathrm{\μ})t_k^{\mathrm{ij}}$

In formula, T ^ijpheromones residual volume, be Pheromone update amount, μ is evaporation coefficient, and this evaporation coefficient gets empirical value 0.5 by emulation.

And described Route Selection probability is calculated by following formula:

$p_n^{\mathrm{ij}}=\frac{T_n^{\mathrm{ij}}}{{\mathrm{\Σ}}_{k=1}^n{T}_k^{\mathrm{ij}}}$

In formula, it is the select probability of n-th route from node i to destination node j; the pheromones total amount of n-th route from node i to destination node j, comprehensive to the amount of information of all routes of destination node j from node i.

Advantage of the present invention and good effect are:

The present invention is reasonable in design, it proposes system-wide by prediction algorithm according to evolution diagram model, many routes in virtual dynamic network between two nodes can be calculated, route re-computation need not be carried out when carrying out routing optimality afterwards, the computing cost of follow-up routing optimality part will be greatly reduced, meet the routing algorithm of qos requirement, more traditional mobile ad hoc network DSR, AODV routing algorithm of performance index is significantly improved.Simultaneously, the directed Ant Routing optimized algorithm that the present invention proposes is the Routing Optimization Algorithm proposed based on the predictability of route, ant group more efficiently can must obtain network communication quality on the basis of learning path, tackle network emergency case more fast, avoid the shortcoming of original ant group algorithm Slow converge.

Accompanying drawing explanation

Fig. 1 is the ant group optimization routing policy structure based on evolution diagram model;

Fig. 2 is evolution diagram model schematic;

Fig. 3 is based on evolution diagram route deep search algorithm flow chart;

Fig. 4 is neighbors searching algorithm flow chart;

Fig. 5 is Front ant data structure schematic diagram;

Fig. 6 is reverse ant data structure schematic diagram;

Fig. 7 is average packet forward rate comparison diagram;

Fig. 8 is average route replies rate comparison diagram;

Fig. 9 is average route discovery time comparison diagram;

Figure 10 is average end-to-end time delay comparison diagram.

Embodiment

Below in conjunction with accompanying drawing, the embodiment of the present invention is further described.

Based on evolution diagram system-wide by the directed Ant Routing optimization method predicted, as shown in Figure 1, through the application layer in the virtual dynamic network being applied to node high-speed motion, network layer and physical layer.Specifically comprise the following steps:

Step 1, in application layer, control centre or GPS global positioning system provide node track, link scheduling information or node location, movement velocity and directional information and are sent to network layer.

Information transmission is divided in real time and non real-time two kinds of modes, real-time mode provides real-time position, movement velocity and directional information by global positioning system, node can carry out track according to road conditions, speed, direction and position, and this mode requires that node has stronger computing capability; Non-real time is before system cloud gray model, reserve node in advance tracks and Link Schedule Time, and this mode requires low to the computing capability of node.

Step 2, in network layer, first node sets up evolution diagram model according to application layer message, then carries out system-wide by predicting, adopts directed Ant Routing optimization method select optimum route and send data.

Data comprise user data and directed ant group, and wherein directed ant group is responsible for providing network communication quality to supervise information, is the pith of routing optimization method of the present invention.The routing strategy of this layer is the probability of survival calculating all routes according to directed ant group feedack element, and the route that probability of survival is the highest is optimum route, uses current optimum route when sending data.

The present invention use evolution diagram model as shown in Figure 2, vertex representation network node in figure, while represent the communication link set up between node, the effective time slot of numeral link that limit marks.Such as, in figure, the significant instant on AB limit is the 3rd time slot, and the significant instant on AC limit is the 2nd, 3 time slot.For simplified illustration, initial time is set to 0, time-gap number all calculates from initial time 0.Such as, if slot length is 3 seconds, then time slot 1 represents 0 to 3 second, and time slot 2 represents 3 to 6 seconds, by that analogy.

The present invention carrying out system-wide by when predicting, adopts system-wide based on evolution diagram model by Forecasting Methodology, as shown in Figure 3.This system-wide comprises by the input content of Forecasting Methodology the maximum delay threshold value (value of being specified by type of service) that evolution diagram model, source node, destination node and system allow; Output content is whole routes that source node arrives destination node, stored in routing table.Concrete Forecasting Methodology comprises the following steps:

Step (1), by stacked for the summit of access, whether search stack top node has the adjacent vertex of not accessing, and has, carries out step (2); Then do not eject stack top summit, until stack is empty, terminate algorithm.

Step (2), judge that whether the significant instant on limit between stack top joint and adjacent vertex meets and be greater than current time and the condition being less than maximum delay threshold value, if meet, adjacent vertex is stacked, upgrading current time is the moment the earliest that data send along this limit, carries out step (3); If do not meet, mark adjacent vertex, for accessing, continues step (2).

Step (3), judge whether adjacent vertex is destination node, if then whole node in output stack in order.

This system-wide by Forecasting Methodology in traversing graph be connected summit time premised on Temporal orderliness condition, the validity of opposite side differentiates.Meanwhile, calculate chain-circuit time delay, travel through out the effective routing meeting and postpone threshold value.

Wherein each node has a local routing table, and router-table structure comprises destination node, route, route break moment and select probability, as shown in Figure 4.Namely the route break moment is inefficacy moment on the last item limit in route, can draw by the effective time slot according to limit in evolution diagram.

The route that the present invention adopts directed Ant Routing optimization method select probability maximum, according to network communication quality real-time determine optimum route.The main contents of directed Ant Routing optimization method comprise to ant group configuration fixed route, and ant group does not need at forward node place to calculate transition probability thus the computation burden reducing node; Secondly ant group is divided into Front ant and reverse ant two type, Front ant is according to the network communication quality parameter at set route collector node place, and these parameters are reflected in pheromones size are retained in forward node place, pheromones can be passed in time and reduce; Finally, reverse ant is collected the pheromones at forward node place according to set route and feeds back to summit, source, and summit, source calculates the select probability of the route that reverse ant is carried according to pheromones summation.

Comprising the following steps of directed Ant Routing optimization method:

Step (1): the Front ant carrying this routing iinformation after summit, source equiprobability selects a route from routing table by the release of certain hour interval.

Step (2): Front ant often arrives a network node, first judges whether to have exceeded the existence life-span, namely postpones threshold value.Then, judge whether current ink degree of Congestion satisfies condition.When above-mentioned two conditions all meet, this Nodes discharge a certain amount of pheromones and the routing forwarding of being carried according to Front ant by present node to next node.

Link congestion degree, calculates by following formula:

η _i,j＝r _i,j/c _i,j (1)

Wherein, η _i,jthe congestion ratio of link between node i and j, r _i,jthe data rate of link between node i and j, c _i,jbeing the capacity of link between node i and j, is also bandwidth.R _i,jand c _i,junit are all bps.

Pheromone release amount, calculates by following formula:

$t_k^{\mathrm{ij}}=\frac{1}{{\left(d_{\mathrm{ik}}\right)}^{\mathrm{\α}}{\mathrm{\Σ}}_i^k{\left({\mathrm{\η}}_{m,m+1}\right)}^{\mathrm{\β}}}---\left(2\right)$

Ij represents the route of source node i to destination node j; represent therefrom in the pheromones of forward node k place release.D _ikrepresent the time delay from i to k, η _{m, m+1}represent the link congestion degree between node m and m+1.α and β is all regulation coefficient, and what be used for regulating time delay and link congestion degree two characteristics on pheromones affects size, through Computer Simulation, all gets 0.5 in the present embodiment.

Step (3): each node, when receiving Front ant, first judges it oneself is whether the destination node of Front ant.If not, then in the data structure of Front ant, write the moment arriving present node, the pheromones of encoding stored in this Front ant and discharging in local information element table, then Front ant is forwarded to next node; If self be the destination node of Front ant, then copy the route of Front ant, generate the reverse ant of identical ant numbering and discharge.

Step (4): oppositely ant returns source node along identical route, often an arrival forward node place collects identical ant and numbers the pheromones residual volume stayed, stored in reverse ant packet.

Pheromones residual volume is remaining pheromones after pheromones is volatilized in time.Node is shown along with passage of time upgrades local information element, and pheromones is evaporated in time by a certain percentage.Pheromone update, calculates by following formula:

$T^{\mathrm{ij}}\←{\mathrm{\μT}}^{\mathrm{ij}}+(1-\mathrm{\μ})t_k^{\mathrm{ij}}---\left(3\right)$

Wherein, T ^ijpheromones residual volume, be Pheromone update amount, μ is evaporation coefficient, gets empirical value 0.5 in this algorithm by emulation.

Step (5): after source node receives reverse ant, according to the pheromones calculation of total Route Selection probability carried, restores in routing table, as last row in Fig. 4 routing table.Finally, node sends data according to the route that route select probability in routing table is maximum.

Route Selection probability, calculates by following formula:

$p_n^{\mathrm{ij}}=\frac{T_n^{\mathrm{ij}}}{{\mathrm{\Σ}}_{k=1}^n{T}_k^{\mathrm{ij}}}---\left(4\right)$

it is the select probability of n-th route from node i to destination node j; the pheromones total amount of n-th route from node i to destination node j, comprehensive to the amount of information of all routes of destination node j from node i.

It is a brief datagram that the present invention defines Front ant and reverse ant, and the life span of ant is once exceed the life-span, and ant datagram can be destroyed by node.Datagram includes head and routing iinformation two parts.

Front ant header fields has: ant numbering, ant type, date of birth, life-span and maximum link degree of Congestion threshold value.Routing iinformation comprises the forward node of approach, and each node comprises information to be had: node serial number, the time of advent, transmitting time, current bandwidth and current data rate, as shown in Figure 5.

Reverse ant header fields has: ant numbering, ant type, date of birth, life-span.Routing iinformation comprises: node serial number and node residual, information element.As shown in Figure 6.

Step 3, in physical layer, the information such as contribution link time delay, message transmission rate, available bandwidth.

Physical layer comprises data communication section and divides and network communication quality monitoring part, the routing strategy respectively in map network layer and directed ant group.Network communication quality monitoring part will provide the network quality parameters such as channel width, channel data rates, chain-circuit time delay, queue time delay for the directed ant group of network layer, and carry all nodes in these Information Communications to network by directed ant group.

By routing algorithm of the present invention and the most frequently used two kinds of routing algorithm wireless self-networkings of existing mobile ad hoc network, plan range vector Routing Protocol and dynamic source routing protocol contrast as required.Algorithm average packet forward rate of the present invention is higher than 90%, and average route replies rate, about 80%, apparently higher than other two kinds of algorithms, and has stable performance in larger data speed range.The average route discovery time of algorithm of the present invention is within 2 seconds, and plan range vector Routing Protocol is fair as required with wireless self-networking, is better than dynamic source routing protocol.This is because starting to set up evolution diagram model and system-wide is consuming time many by predicting, improve average route discovery time, and later stage route discovery time is reducing greatly.The average end-to-end time delay of algorithm of the present invention is lower than wireless self-networking plan range vector Routing Protocol as required, higher than dynamic source routing protocol, this is because the average packet forward rate of dynamic source routing protocol is very low, only a few packets is comparatively fast forwarded to apart near node, so end-to-end time delay seems lower.Consider packet average arrival rate and average end-to-end time delay, routing algorithm performance of the present invention is more better than other two kinds of algorithms, thus illustrates that algorithm of the present invention can provide the routing function of more dominance energy in virtual dynamic network.

The present invention adopts evolution diagram model system-wide by the directed Ant Routing optimization method predicted, can go out specify whole effective routings in end-to-end time delay, for satellite network multicast provides route implementation method based on evolution diagram model prediction.

The present invention can system-wide by after predicting by directed ant group real time monitoring network communication quality, sensor selection problem is helped to go out current optimal path and send data according to the feedback information that ant group provides, thus the Satellite Network Routing Algorithms of quality of service guarantee (Quality of Service, QoS) is provided.Directed ant group reduces node section computation burden, avoids traditional ant group algorithm when being applied in virtual dynamic network, is frequently switched the slow convergence problem brought by link.The present invention devises computational methods and the ant group structure of directed ant group Pheromone update according to algorithm requirements, network communication quality Parameter Switch is become pheromones, by pheromones size reflection network communication quality, thus simplify directed ant group structure, improve the flexibility of network reply emergency.

Contrast effect figure as can be seen from Fig. 7 to Figure 10, average packet forward rate of the present invention and average route replies rate are far away higher than than existing method.Average route discovery time of the present invention is far below dynamic agreement, and average end-to-end time delay of the present invention is much smaller than plan range vector Routing Protocol as required.

It is emphasized that; embodiment of the present invention is illustrative; instead of it is determinate; therefore the present invention includes the embodiment be not limited to described in embodiment; every other execution modes drawn by those skilled in the art's technical scheme according to the present invention, belong to the scope of protection of the invention equally.

Claims (10)

1. based on evolution diagram system-wide by the directed Ant Routing optimization method predicted, it is characterized in that comprising the following steps:

Step 1, in application layer, control centre or GPS global positioning system provide node track, link scheduling information or node location, movement velocity and directional information and are sent to network layer;

Step 2, in network layer, first node sets up evolution diagram model according to application layer message, then carries out system-wide by predicting, finally adopts directed Ant Routing optimization method select optimum route and send data;

Step 3, in physical layer, contribution link time delay, message transmission rate and available bandwidth information.

2. according to claim 1 a kind of based on evolution diagram system-wide by the directed Ant Routing optimization method predicted, it is characterized in that: described step 2 system-wide is by predicting that the system-wide of employing based on evolution diagram model is by Forecasting Methodology, this system-wide comprises by the input content of Forecasting Methodology the maximum delay threshold value that evolution diagram model, source node, destination node and system allow, output content is that source node arrives whole route of destination node and stored in routing table, concrete Forecasting Methodology comprises the following steps:

Step (1), by stacked for the summit of access, whether search stack top node has the adjacent vertex of not accessing, and has, carries out step (2); Then do not eject stack top summit, until stack is empty, terminate algorithm;

Step (2), judge that whether the significant instant on limit between stack top joint and adjacent vertex meets and be greater than current time and the condition being less than maximum delay threshold value, if meet, adjacent vertex is stacked, upgrading current time is the moment the earliest that data send along this limit, carries out step (3); If do not meet, mark adjacent vertex, for accessing, continues step (2);

Step (3), judge whether adjacent vertex is destination node, if then whole node in output stack in order.

3. according to claim 2 a kind of based on evolution diagram system-wide by the directed Ant Routing optimization method predicted, it is characterized in that: the structure of described routing table comprises destination node, route, route break moment and select probability.

4. according to claim 1 a kind of based on evolution diagram system-wide by the directed Ant Routing optimization method predicted, it is characterized in that: the directed Ant Routing optimization method of described step 2 is: ant group configuration fixed route, ant group is divided into Front ant and reverse ant two type, Front ant is according to the network communication quality parameter at set route collector node place, and these parameters are reflected in pheromones size are retained in forward node place, pheromones can be passed in time and reduce; Finally, reverse ant is collected the residual, information element at forward node place according to set route and feeds back to summit, source, and summit, source calculates the select probability of the route that reverse ant is carried according to residual, information element summation.

5. according to claim 4 a kind of based on evolution diagram system-wide by the directed Ant Routing optimization method predicted, it is characterized in that: the specific implementation step of described directed Ant Routing optimization method comprises:

Step (1): the Front ant carrying this routing iinformation after summit, source equiprobability selects a route from routing table by the release of certain hour interval;

Step (2): Front ant often arrives a network node, first judges whether to have exceeded the existence life-span, then judges whether current ink degree of Congestion satisfies condition; When above-mentioned two conditions all meet, this Nodes discharge a certain amount of pheromones and the routing forwarding of being carried according to Front ant by present node to next node;

Step (3): each node, when receiving Front ant, first judges it oneself is whether the destination node of Front ant; If not, then in the data structure of Front ant, write the moment arriving present node, the pheromones of encoding stored in this Front ant and discharging in local information element table, then Front ant is forwarded to next node; If self be the destination node of Front ant, then copy the route of Front ant, generate the reverse ant of identical ant numbering and discharge;

Step (4): oppositely ant returns source node along identical route, often an arrival forward node place collects identical ant and numbers the pheromones residual volume stayed, stored in reverse ant packet;

Step (5): after source node receives reverse ant, according to the pheromones calculation of total Route Selection probability carried, restores in routing table, and finally, node sends data according to the route that route select probability in routing table is maximum.

6. a kind of according to claim 4 or 5 based on evolution diagram system-wide by the directed Ant Routing optimization method predicted, it is characterized in that: described Front ant is a packet, comprising: ant numbering, ant type, date of birth, life-span and maximum link degree of Congestion threshold value and routing iinformation; Each routing node of described Front ant comprises information to be had: node serial number, the time of advent, transmitting time, current bandwidth and current data rate; Described reverse ant is a packet, comprising: ant numbering, ant type, date of birth, life-span and routing iinformation; Each routing node of reverse ant comprises information to be had: node serial number and node residual, information element.

7. according to claim 5 a kind of based on evolution diagram system-wide by the directed Ant Routing optimization method predicted, it is characterized in that: described link congestion degree is calculated by following formula:

η _i,j＝r _i,j/c _i,j

Wherein, η _i,jthe congestion ratio of link between node i and j, r _i,jthe data rate of link between node i and j, c _i,jit is the capacity of link between node i and j.

8. according to claim 5 a kind of based on evolution diagram system-wide by the directed Ant Routing optimization method predicted, it is characterized in that: described pheromones is that Front ant often discharges through a forward node, the size of pheromones reflects network communication quality, and pheromone release amount is calculated by following formula:

$t_k^{\mathrm{ij}}=\frac{1}{{\left(d_{\mathrm{ik}}\right)}^{\mathrm{\α}}{\mathrm{\Σ}}_i^k{\left({\mathrm{\η}}_{m,m+1}\right)}^{\mathrm{\β}}}$

In formula, ij represents the route of source node i to destination node j; represent therefrom in the pheromones of forward node k place release, d _ikrepresent the time delay from i to k, η _{m, m+1}represent the link congestion degree between node m and m+1; α and β is all regulation coefficient, and what be used for regulating time delay and link congestion degree two characteristics on pheromones affects size, α and β all gets 0.5.

9. according to claim 5 a kind of based on evolution diagram system-wide by the directed Ant Routing optimization method predicted, it is characterized in that: described residual, information element is the pheromones of Nodes through certain hour remaining after evaporation, this residual, information element is upgraded and is calculated by following formula:

$T^{\mathrm{ij}}\←{\mathrm{\μT}}^{\mathrm{ij}}+(1-\mathrm{\μ})t_k^{\mathrm{ij}}$

In formula, T ^ijpheromones residual volume, be Pheromone update amount, μ is evaporation coefficient, and this evaporation coefficient gets empirical value 0.5 by emulation.

10. according to claim 5 a kind of based on evolution diagram system-wide by the directed Ant Routing optimization method predicted, it is characterized in that: described Route Selection probability is calculated by following formula:

$p_n^{\mathrm{ij}}=\frac{T_n^{\mathrm{ij}}}{{\mathrm{\Σ}}_{k=1}^n{T}_k^{\mathrm{ij}}}$

In formula, it is the select probability of n-th route from node i to destination node j; the pheromones total amount of n-th route from node i to destination node j, comprehensive to the amount of information of all routes of destination node j from node i.