CN110365585B - Route cutting optimization method based on multi-cost index - Google Patents

Route cutting optimization method based on multi-cost index Download PDF

Info

Publication number
CN110365585B
CN110365585B CN201810254608.5A CN201810254608A CN110365585B CN 110365585 B CN110365585 B CN 110365585B CN 201810254608 A CN201810254608 A CN 201810254608A CN 110365585 B CN110365585 B CN 110365585B
Authority
CN
China
Prior art keywords
node
cost index
cost
route
network
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN201810254608.5A
Other languages
Chinese (zh)
Other versions
CN110365585A (en
Inventor
黄传河
覃匡宇
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Wuhan University WHU
Original Assignee
Wuhan University WHU
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Wuhan University WHU filed Critical Wuhan University WHU
Priority to CN201810254608.5A priority Critical patent/CN110365585B/en
Publication of CN110365585A publication Critical patent/CN110365585A/en
Application granted granted Critical
Publication of CN110365585B publication Critical patent/CN110365585B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L45/00—Routing or path finding of packets in data switching networks
    • H04L45/14—Routing performance; Theoretical aspects

Abstract

The invention provides a route cutting optimization method based on multi-cost indexes. Constructing a multi-dimensional space through cost index vectors and performing variable length segmentation on each dimensional axis according to a stepping sequence established by precision; selecting an intermediate node from the nodes routed from the source node to the target node, and calculating a cost index vector from the source node to a neighbor node of the intermediate node; comparing any two cost index vectors in all cost index vectors from a source node to a neighbor node of an intermediate node to realize one-time routing cutting; performing secondary cutting in a multi-dimensional space according to the residual cost index vectors from the source node to the neighbor nodes of the intermediate node; repeatedly executing the first cutting and the second cutting to search the route from the source node to the target node and carrying out optimization selection; and repeatedly executing the searching to the route from the source node to the target node and optimizing for multiple iterations. Compared with the prior art, the method is simple and feasible, and can process the high-dimensional cost index in polynomial time.

Description

Route cutting optimization method based on multi-cost index
Technical Field
The invention relates to the technical field of wireless network communication, in particular to a route cutting optimization method based on multi-cost indexes.
Background
In the calculation of the shortest path, each link has a cost value, and the routing decision module completes routing by finding a path with the least total cost at the destination. In the RIP protocol, the hop count is used as a basis for routing, and the cost per link is 1. In the OSPF protocol, default is to use the baseline bandwidth divided by the interface bandwidth as the cost indicator for each egress link.
In recent years, Software Defined Networking (SDN) has been greatly developed. The software defined network implements centralized management of control logic by separating the control plane from the data plane. In software defined networks, forwarding of data is decided by an SDN controller. Since the SDN controller has a global view, paths can be planned very accurately for each flow. When the routing decision is made, for a path between two points, the cost of the path is the sum of the costs of each link. The controller selects the best path from the source node to the target node based on the cost values of the paths in the network.
The current controller uses a single index to measure the cost of the link, but in practical use, a really good path cannot be simply measured by using a single index, multiple factors including time delay, bandwidth, hop count, rent and the like need to be considered simultaneously, and at this time, two or more cost indexes should be considered for each link. The Dijkstra algorithm may be used to compute when the optimal path is selected at a single cost. But for a path with multiple indexes, no good method exists at present. The main problem is that the search space is too large, considering a fully connected network of n nodes, the path between any two points can have (n-1)! In short, the calculation of the optimal path cannot be completed in a short time. This problem is called Multi-Object Optimal Path, MOOP, and related studies indicate that it is an NP-complete problem requiring exponential time to obtain a pareto Optimal solution set. Therefore, the current SDN controller only supports routing with a single cost index, and there is no routing scheme with multiple cost indexes.
Disclosure of Invention
In order to solve the problems, the invention provides a route cutting optimization method based on multi-cost indexes, and the technical scheme adopted by the invention is as follows:
step 1: constructing a multi-dimensional space through the cost index vector, and performing variable length segmentation on each dimensional axis of the multi-dimensional space according to a stepping sequence established by precision;
step 2: selecting a source node and a target node from network nodes, selecting an intermediate node from the nodes routed from the source node to the target node, and calculating cost index vectors of neighbor nodes from the source node to the intermediate node according to the cost index vectors from the source node to the intermediate node;
and step 3: all cost index vectors from a source node to a neighbor node of an intermediate node are traversed, and any two cost index vectors are compared to achieve one-time routing cutting to obtain residual cost index vectors and residual routing sets from the node to the neighbor node of the intermediate node;
and 4, step 4: performing secondary cutting on the remaining routes from the source node to the neighbor nodes of the intermediate node in the multi-dimensional space according to the remaining cost index vectors from the source node to the neighbor nodes of the intermediate node;
and 5: repeating the steps 3 to 4 until the route from the source node to the target node is searched and carrying out optimization selection;
step 6: step 5 is repeatedly performed for a plurality of iterations.
Preferably, the cost index vector in step 1 is:
Wxy,i＝(wxy,i,1,wxy,i,2,...,wxy,i,K)x∈[1,N],y∈[1,N],i∈[1,|Pxy|],x≠y
wherein, Wxy,iCost index vector, P, for the ith route from network node x to network node yxyFor the set of routes from network node x to network node y, | PxyI is the number of the routes from the network node x to the network node y, namely the number of the cost index vectors, N is the number of the nodes in the network, K is the number of the indexes in the network, and w isxy,i,kk∈[1,K]A kth cost index of a cost index vector of an ith route from the network node x to the network node y;
constructing a multidimensional space H according to the number K of indexes in the networkKThe dimension of the compound is K;
the lower bound of the j (j is more than or equal to 1 and less than or equal to K) dimension cost in all paths is set as LBjThe upper bound is UBjPrecision e, and there is one MjE.z + (positive integer) such that
Figure BDA0001608682830000021
Then
The step sequence established according to the precision e in the step 1 is as follows:
Figure BDA0001608682830000022
wherein, the mth element in the stepping sequence is (1+ e)m-1m∈[1,Mj]；
The pair of multidimensional spaces H in step 1KIs segmented into:
Figure BDA0001608682830000031
wherein, the nth segment on the j dimension axis is [ (1+ e)n-1,(1+e)n]n∈[1,Mj]Multidimensional space HKThe K axes are subjected to variable length segmentation according to the stepping sequence;
preferably, in step 2, network node s (s e [1, N ]) is selected from the network nodes as a source node, network node t (t e [1, N ]) is selected as a target node, and N is the number of nodes in the network;
in step 2, the network node u (u belongs to [1, N ]) is selected as the intermediate node, and the cost index vector from the source node to the intermediate node in step 2
Wsu,l＝(wsu,l,1,wsu,l,2,...,wsu,l,K)s∈[1,N],u∈[1,N],l∈[1,|Psu|],s≠u
Wherein, Wsu,lCost index vector, P, for the ith route from source node to intermediate nodesuIs a set of routes, | P, from the source node to the intermediate nodexyI is the number of routes from the source node to the intermediate node, i.e. the number of cost index vectors, N is the number of nodes in the network, K is the number of indexes in the network, wsu,l,k k∈[1,K]The kth cost index of the cost index vector of the ith route from the source node to the intermediate node in the network;
in step 2, the cost index vector from the intermediate node to the neighbor node of the intermediate node is as follows:
Wuv,o＝(wuv,o,1,wuv,o,2,...,wuv,o,K)u∈[1,N],v∈[1,N],o∈[1,|Puv|],u≠v
wherein, Wuv,oCost index vector P of the o-th route from the intermediate node to the neighbor node of the intermediate nodesuIs a set of routes, | P, from intermediate node to intermediate node's neighbor nodexyI is the neighbor from the intermediate node to the intermediate nodeThe number of nodes routing, i.e. the number of cost indicator vectors, N is the number of nodes in the network, K is the number of indicators in the network, wuv,o,kk∈[1,K]The kth cost index of the cost index vector of the o-th route from the intermediate node to the neighbor node of the intermediate node;
step 2, the cost index vector from the source node to the neighbor node of the intermediate node is as follows:
Wsv,q＝Wsu,l+Wuv,o
wherein, Wsv,q(q∈[1,|Psv|]) Cost index vector P of q route from source node to neighbor node of intermediate nodesvIs a set of routes, | P, from the source node to the neighbor nodes of the intermediate nodesvI is the number of the routes from the source node to the neighbor nodes of the intermediate node, namely the number of the cost index vectors, and PsvAny route in the network does not contain intermediate nodes, namely network nodes u (u belongs to [1, N ]])；
Preferably, the cost index vectors of the neighboring nodes from the source node to the intermediate node in step 3 are respectively:
Figure BDA0001608682830000041
wherein, Wsv,q(q∈[1,|Psv|]) Cost index vector P of q route from source node to neighbor node of intermediate nodesvIs a set of routes, | P, from the source node to the neighbor nodes of the intermediate nodesvI is the number of the routes from the source node to the neighbor nodes of the intermediate node, namely the number of the cost index vectors, and PsvAny route in the network does not contain intermediate nodes, namely network nodes u (u belongs to [1, N ]])；
Comparing any two cost index vectors in the step 3, and randomly selecting two cost index vectors Wsv,g(g∈[1,|Psv|]) And Wsv,h(h∈[1,|Psv|]) And g ≠ h, if Wsv,g(g∈[1,|Psv|]) Each element in (1) is less than or equal to Wsv,h(h∈[1,|Psv|]) For each corresponding element in (i.e., for any K (K e [1, K))]) Satisfies Wsv,g(k)≤Wsv,h(k) Then W will besv,hThe route set P of the h route from the source node to the neighbor node of the intermediate nodesvDeleting and cutting;
traversing all cost index vectors from the source node to the neighbor nodes of the intermediate node, wherein the residual routing quantity from the source node to the neighbor nodes of the intermediate node, namely the quantity of the residual cost index vectors is as follows according to the comparison method
Figure BDA0001608682830000042
Figure BDA0001608682830000043
The residual route set from the source node to the neighbor node of the intermediate node is obtained;
preferably, the number of remaining routes from the source node to the neighbor nodes of the intermediate node in step 4, that is, the number of remaining cost indicator vectors, is
Figure BDA0001608682830000044
The residual route set from the source node to the neighbor node of the intermediate node is selected if
Figure BDA0001608682830000045
Finishing the secondary cutting;
if it is
Figure BDA0001608682830000046
The residual cost index vector from the source node to the neighbor node of the intermediate node in the step 4 is
Figure BDA0001608682830000047
In a multi-dimensional space HKCutting;
step length changing segmentation is carried out on each dimension axis of the multidimensional space according to the step sequence established in the step 1, and the multidimensional space H is divided into a plurality of step lengthsKDividing the data into a plurality of subspaces, wherein the residual cost index vector from the source node to the neighbor node of the intermediate node in the step 4 is
Figure BDA0001608682830000051
Respectively belong to S subspaces, each continuous subspace respectively contains M1,M2,...,MSA residual cost index vector;
for respectively containing M1,M2,...,MSEach subspace of cost index vectors, according to any one subspace, the number of the remaining cost index vectors contained in the subspace is MfMf∈[M1,MS]The subspace includes a cost index vector of
Figure BDA0001608682830000052
Respectively calculating the weighted value of each cost index vector as follows:
Figure BDA0001608682830000053
wherein the content of the first and second substances,
Figure BDA0001608682830000054
is the minimum value, the source node corresponding to the minimum value is connected to the neighbor node of the intermediate node
Figure BDA0001608682830000055
F in (1)zOne route is reserved, M remainsf-1 route all from
Figure BDA0001608682830000056
Deleting, namely cutting;
preferably, in step 5, network node s (s e [1, N ]) is selected from the network nodes as a source node, network node t (t e [1, N ]) is selected as a target node, and N is the number of nodes in the network;
in step 5, the cost index vectors from the source node to the target node are respectively
Figure BDA0001608682830000057
The number of routes from the source node to the target node, i.e. the number of cost indicator vectors, is
Figure BDA0001608682830000058
Respectively calculating the weighted value of each cost index vector as follows for the route set from the source node to the target node:
Figure BDA0001608682830000059
wherein the content of the first and second substances,
Figure BDA00016086828300000510
is the minimum value, corresponding to the source node to the target node
Figure BDA00016086828300000511
N inminOne route is reserved, the rest
Figure BDA00016086828300000512
All routes are from
Figure BDA00016086828300000513
Deleting, namely cutting;
preferably, the number of times that the iteration in step 6 repeatedly performs step 5 is MoptN-1, N is the number of nodes in the network.
Compared with the prior art, the method cuts off other paths, thereby greatly reducing the search space; when the iteration is finished and the operation is carried out to the target node, the rest paths on the target node are approximate multi-cost optimal paths; the method is simple and easy to implement, and can process high-dimensional cost indexes in polynomial time.
Drawings
FIG. 1: is a flow chart of the method of the present invention;
FIG. 2: is a network topology diagram of an embodiment of the invention.
Detailed Description
In order to facilitate the understanding and implementation of the present invention for those of ordinary skill in the art, the present invention is further described in detail with reference to the accompanying drawings and examples, it is to be understood that the embodiments described herein are merely illustrative and explanatory of the present invention and are not restrictive thereof.
Fig. 1 is a flowchart of a method of the present invention, and fig. 2 is a network topology diagram of an embodiment of the present invention. The embodiment of the invention carries out python language programming on a Mininet simulation platform, Floodlight is used as a controller in the Mininet, 6 switches are selected as network nodes, namely the number of the network nodes is 6, two-way links are arranged between every two network nodes, each link in the graph comprises a two-dimensional cost index vector, the links are marked by a pair of brackets, the first element of the cost index vector represents the delay cost of the link, the second element of the cost index vector represents the bandwidth cost of the link, the network node 1 is selected as a source node, and the network node 6 is selected as a target node.
The following describes specific steps of an embodiment of the present invention with reference to fig. 1 and fig. 2, and the present invention provides a route cutting optimization method based on multiple cost indexes, which specifically includes the following steps:
step 1: constructing a multi-dimensional space through the cost index vector, and performing variable length segmentation on each dimensional axis of the multi-dimensional space according to a stepping sequence established by precision;
the cost index vector in step 1 is:
Wxy,i＝(wxy,i,1,wxy,i,2,...,wxy,i,K)x∈[1,N],y∈[1,N],i∈[1,|Pxy|],x≠y
wherein, Wxy,iCost index vector, P, for the ith route from network node x to network node yxyFor the set of routes from network node x to network node y, | PxyI is the number of the cost index vectors, i.e. the number of routes from the network node x to the network node y, N is 6, K is 2, w is the number of the nodes in the networkxy,i,kk∈[1,K]A kth cost index of a cost index vector of an ith route from the network node x to the network node y;
constructing a multidimensional space H according to the number K of indexes in the network being 2KThe dimension of the compound is K;
let j (1 ≤ j ≤ K) dimension generation in all pathsLower bound of valence is LBj1, UB as the upper boundj10, precision e, and there is one MjE.z + (positive integer) such that
Figure BDA0001608682830000072
Then
The step sequence established according to the precision e in the step 1 is as follows:
Figure BDA0001608682830000073
wherein, the mth element in the stepping sequence is (1+ e)m-1m∈[1,Mj]；
The pair of multidimensional spaces H in step 1KIs segmented into:
Figure BDA0001608682830000071
wherein, the nth segment on the j dimension axis is [ (1+ e)n-1,(1+e)n]n∈[1,Mj]Multidimensional space HKThe K axes are subjected to variable length segmentation according to the stepping sequence;
step 2: selecting a source node and a target node from network nodes, selecting an intermediate node from the nodes routed from the source node to the target node, and calculating cost index vectors of neighbor nodes from the source node to the intermediate node according to the cost index vectors from the source node to the intermediate node;
in the step 2, a network node s (s belongs to [1, N ]) is selected from the network nodes as a source node, a network node t (t belongs to [1, N ]) is selected as a target node, and the number of the nodes in the network is N-6;
in step 2, the network node u (u belongs to [1, N ]) is selected as the intermediate node, and the cost index vector from the source node to the intermediate node in step 2
Wsu,l＝(wsu,l,1,wsu,l,2,...,wsu,l,K)s∈[1,N],u∈[1,N],l∈[1,|Psu|],s≠u
Wherein, Wsu,lCost index vector, P, for the ith route from source node to intermediate nodesuIs a set of routes, | P, from the source node to the intermediate nodexyI is the number of routes from the source node to the intermediate node, namely the number of cost index vectors, N is 6 is the number of nodes in the network, K is 2 is the number of indexes in the network, w issu,l,kk∈[1,K]The kth cost index of the cost index vector of the ith route from the source node to the intermediate node in the network;
in step 2, the cost index vector from the intermediate node to the neighbor node of the intermediate node is as follows:
Wuv,o＝(wuv,o,1,wuv,o,2,...,wuv,o,K)u∈[1,N],v∈[1,N],o∈[1,|Puv|],u≠v
wherein, Wuv,oCost index vector P of the o-th route from the intermediate node to the neighbor node of the intermediate nodesuIs a set of routes, | P, from intermediate node to intermediate node's neighbor nodexyI is the number of the routes from the intermediate node to the neighbor nodes of the intermediate node, namely the number of the cost index vectors, N is 6 is the number of the nodes in the network, K is 2 is the number of the indexes in the network, and w isuv,o,kk∈[1,K]The kth cost index of the cost index vector of the o-th route from the intermediate node to the neighbor node of the intermediate node;
step 2, the cost index vector from the source node to the neighbor node of the intermediate node is as follows:
Wsv,q＝Wsu,l+Wuv,o
wherein, Wsv,q(q∈[1,|Psv|]) Cost index vector P of q route from source node to neighbor node of intermediate nodesvIs a set of routes, | P, from the source node to the neighbor nodes of the intermediate nodesvI is the number of the routes from the source node to the neighbor nodes of the intermediate node, namely the number of the cost index vectors, and PsvAny route in the network does not contain intermediate nodes, namely network nodes u (u belongs to [1, N ]])；
And step 3: all cost index vectors from a source node to a neighbor node of an intermediate node are traversed, and any two cost index vectors are compared to achieve one-time routing cutting to obtain residual cost index vectors and residual routing sets from the node to the neighbor node of the intermediate node;
in step 3, the cost index vectors from the source node to the neighbor nodes of the intermediate node are respectively:
Figure BDA0001608682830000081
wherein, Wsv,q(q∈[1,|Psv|]) Cost index vector P of q route from source node to neighbor node of intermediate nodesvIs a set of routes, | P, from the source node to the neighbor nodes of the intermediate nodesvI is the number of the routes from the source node to the neighbor nodes of the intermediate node, namely the number of the cost index vectors, and PsvAny route in the network does not contain intermediate nodes, namely network nodes u (u belongs to [1, N ]])；
Comparing any two cost index vectors in the step 3, and randomly selecting two cost index vectors Wsv,g(g∈[1,|Psv|]) And Wsv,h(h∈[1,|Psv|]) And g ≠ h, if Wsv,g(g∈[1,|Psv|]) Each element in (1) is less than or equal to Wsv,h(h∈[1,|Psv|]) For each corresponding element in (i.e., for any K (K e [1, K))]) Satisfies Wsv,g(k)≤Wsv,h(k) Then W will besv,hThe route set P of the h route from the source node to the neighbor node of the intermediate nodesvDeleting and cutting;
traversing all cost index vectors from the source node to the neighbor nodes of the intermediate node, wherein the residual routing quantity from the source node to the neighbor nodes of the intermediate node, namely the quantity of the residual cost index vectors is as follows according to the comparison method
Figure BDA0001608682830000091
Figure BDA0001608682830000092
The residual route set from the source node to the neighbor node of the intermediate node is obtained;
and 4, step 4: performing secondary cutting on the remaining routes from the source node to the neighbor nodes of the intermediate node in the multi-dimensional space according to the remaining cost index vectors from the source node to the neighbor nodes of the intermediate node;
the number of the remaining routes from the source node to the neighbor node of the intermediate node in the step 4, that is, the number of the remaining cost index vectors is
Figure BDA0001608682830000093
The residual route set from the source node to the neighbor node of the intermediate node is selected if
Figure BDA0001608682830000094
Finishing the secondary cutting;
if it is
Figure BDA0001608682830000095
The residual cost index vector from the source node to the neighbor node of the intermediate node in the step 4 is
Figure BDA0001608682830000096
In a multi-dimensional space HKCutting;
step length changing segmentation is carried out on each dimension axis of the multidimensional space according to the step sequence established in the step 1, and the multidimensional space H is divided into a plurality of step lengthsKDividing the data into a plurality of subspaces, wherein the residual cost index vector from the source node to the neighbor node of the intermediate node in the step 4 is
Figure BDA0001608682830000097
Respectively belong to S subspaces, each continuous subspace respectively contains M1,M2,...,MSA residual cost index vector;
for respectively containing M1,M2,...,MSEach subspace of cost index vectors, according to any one subspace, the number of the remaining cost index vectors contained in the subspace is Mf Mf∈[M1,MS]The subspace includes a cost index vector of
Figure BDA0001608682830000098
Respectively calculating the weighted value of each cost index vector as follows:
Figure BDA0001608682830000099
wherein the content of the first and second substances,
Figure BDA00016086828300000910
is the minimum value, the source node corresponding to the minimum value is connected to the neighbor node of the intermediate node
Figure BDA00016086828300000911
F in (1)zOne route is reserved, M remainsf-1 route all from
Figure BDA00016086828300000912
Deleting, namely cutting;
and 5: repeating the steps 3 to 4 until the route from the source node to the target node is searched and carrying out optimization selection;
in step 5, selecting network nodes s (s belongs to [1, N ]) as source nodes from the network nodes, selecting network nodes t (t belongs to [1, N ]) as target nodes, and taking the number of the nodes in the network as N ═ 6;
in step 5, the cost index vectors from the source node to the target node are respectively
Figure BDA0001608682830000101
The number of routes from the source node to the target node, i.e. the number of cost indicator vectors, is
Figure BDA0001608682830000102
Respectively calculating the weighted value of each cost index vector as follows for the route set from the source node to the target node:
Figure BDA0001608682830000103
wherein the content of the first and second substances,
Figure BDA0001608682830000104
is the minimum value, corresponding to the source node to the target node
Figure BDA0001608682830000105
N inminOne route is reserved, the rest
Figure BDA0001608682830000106
All routes are from
Figure BDA0001608682830000107
Deleting, namely cutting;
step 6: repeatedly executing the step 5 for multiple iterations;
the number of times of iterative repeated execution of the step 5 is MoptThe optimal path from the source node, network node 1, to the target node, network node 6, is 1-2-4-6, 5.
It should be understood that the above description of the preferred embodiments is given for clarity and not for any purpose of limitation, and that various changes, substitutions and alterations can be made herein without departing from the spirit and scope of the invention as defined by the appended claims.

Claims (7)

1. A route cutting optimization method based on multi-cost indexes is characterized by comprising the following steps:
step 1: constructing a multi-dimensional space through the cost index vector, and performing variable length segmentation on each dimensional axis of the multi-dimensional space according to a stepping sequence established by precision;
step 2: for a given network source node and a given network target node, performing breadth traversal from the source node to the target node to obtain an intermediate node, and calculating cost index vectors of neighbor nodes from the source node to the intermediate node according to the cost index vectors from the source node to the intermediate node;
and step 3: all cost index vectors of neighbor nodes from a source node to an intermediate node are traversed, and any two cost index vectors are compared to achieve one-time routing cutting to obtain residual cost index vectors and residual routing sets of the neighbor nodes from the source node to the intermediate node;
and 4, step 4: performing secondary cutting on the remaining routes from the source node to the neighbor nodes of the intermediate node in the multi-dimensional space according to the remaining cost index vectors from the source node to the neighbor nodes of the intermediate node;
and 5: repeating the steps 3 to 4 until the route from the source node s to the target node t is searched and optimized;
step 6: step 5 is repeatedly performed for a plurality of iterations.
2. The multi-cost-index-based route pruning optimization method according to claim 1, wherein the cost index vector in step 1 is:
Wxy,i＝(wxy,i,1,wxy,i,2,...,wxy,i,K)x∈[1,N],y∈[1,N],i∈[1,|Pxy|],x≠y
wherein, Wxy,iCost index vector, P, for the ith route from network node x to network node yxyFor the set of routes from network node x to network node y, | PxyI is the number of the routes from the network node x to the network node y, namely the number of the cost index vectors, N is the number of the nodes in the network, K is the number of the indexes in the network, and w isxy,i,k k∈[1,K]A kth cost index of a cost index vector of an ith route from the network node x to the network node y;
constructing a multidimensional space H according to the number K of indexes in the networkKThe dimension of the compound is K;
let the lower bound of the jth dimension cost in all paths be LBjThe upper bound is UBjThe precision is e, j is more than or equal to 1 and less than or equal to K, and one M existsj∈Z+(positive integer) such that
Figure RE-FDA0003097243530000011
Then
The step sequence established according to the precision e in the step 1 is as follows:
Figure RE-FDA0003097243530000012
wherein, the mth element in the stepping sequence is LBj*(1+e)m-1，m∈[1,Mj+1]；
The pair of multidimensional spaces H in step 1KIs segmented into:
Figure RE-FDA0003097243530000021
wherein, the nth segment on the j dimension axis is [ LB ]j*(1+e)n-1,LBj*(1+e)n]n∈[1,Mj]Multidimensional space HKThe K axes of (a) are all length-variable segmented according to the stepping sequence.
3. The multi-cost-index-based route cutting optimization method according to claim 1, wherein the network source node in the step 2 is set to be s, s belongs to [1, N ], the target node is set to be t, t belongs to [1, N ], and the number of nodes in the network is set to be N;
in step 2, the breadth traversal is carried out from the source node s to obtain a network intermediate node u, u belongs to [1, N ], and the cost index vector of the ith route from the source node s to the intermediate node u is
Wsu,l＝(wsu,l,1,wsu,l,2,...,wsu,l,K),s∈[1,N],u∈[1,N],l∈[1,|Psu|]
Wherein, Wsu,lCost index vector, P, for the ith route from source node to intermediate nodesuIs a set of routes, | P, from the source node to the intermediate nodesuI is the number of routes from the source node to the intermediate node, i.e. the number of cost index vectors, N is the number of nodes in the network, K is the number of indexes in the network, wsu,l,kCost index direction of the ith route from source node s to node u in networkThe kth dimension of the quantity is a cost index, K belongs to [1, K ∈]；
In step 2, the cost index vector of the o-th route from the intermediate node u to the neighbor node v of the intermediate node is as follows:
Wuv,o＝(wuv,o,1,wuv,o,2,...,wuv,o,K),u∈[1,N],v∈[1,N],o∈[1,|Puv|],u≠v
wherein, Wuv,oCost index vector P of the o-th route from the intermediate node to the neighbor node of the intermediate nodeuvIs a set of routes, | P, from intermediate node to intermediate node's neighbor nodeuvI is the number of the routes from the intermediate node to the neighbor node of the intermediate node, namely the number of the cost index vectors, N is the number of the nodes in the network, K is the number of the indexes in the network, and w is the number of the indexes in the networkuv,o,kFor the kth dimension cost index of the cost index vector of the route from the node u to the node v, K belongs to [1, K ]]；
Step 2, the cost index vector from the source node s to the neighbor node v of the intermediate node is as follows:
Wsv,q＝Wsu,l+Wuv,o
wherein, Wsv,qA cost index vector of a q-th route from a source node to a neighbor node of an intermediate node, wherein the q-th route passes through the ith route from s to u and the ith route from u to v respectively, and q belongs to [1, | Psv|]，PsvIs a set of routes, | P, from the source node to the neighbor nodes of the intermediate nodesvI is the number of the routes from the source node to the neighbor nodes of the intermediate node, namely the number of the cost index vectors, and PsvThere is no loop in any of the routes, i.e. there are no two identical nodes.
4. The multi-cost-index-based route cutting optimization method according to claim 1, wherein cost index vectors from the source node to the neighbor nodes of the intermediate node in step 3 are respectively:
Figure RE-FDA0003097243530000034
wherein, Wsv,qA cost index vector of a q route from a source node to a neighbor node of an intermediate node, wherein q belongs to [1, | P |)sv|]，PsvIs a set of routes, | P, from the source node to the neighbor nodes of the intermediate nodesvI is the number of the routes from the source node to the neighbor nodes of the intermediate node, namely the number of the cost index vectors, and PsvAny route in the routing does not contain a certain node twice, namely, no loop exists;
comparing any two cost index vectors in the step 3, and randomly selecting two cost index vectors Wsv,gAnd Wsv,hAnd g ≠ h, if Wsv,gEach element in (1) is less than or equal to Wsv,hThat is, w is satisfied for any ksv,g,k≤wsv,h,kThen W will besv,hThe route set P of the h route from the source node to the neighbor node of the intermediate nodesvDeleting, i.e. clipping, g belongs to [1, | P |)sv|],h∈[1,|Psv|],g∈[1,|Psv|],
k∈[1,K]；
Traversing all cost index vectors from the source node to the neighbor nodes of the intermediate node, wherein the residual routing quantity from the source node to the neighbor nodes of the intermediate node, namely the quantity of the residual cost index vectors is as follows according to the comparison method
Figure RE-FDA0003097243530000031
And collecting the residual routes from the source node to the neighbor nodes of the intermediate node.
5. The multi-cost-index-based route cutting optimization method according to claim 1, wherein the number of remaining routes from the source node to the neighbor nodes of the intermediate node in step 4, that is, the number of remaining cost index vectors, is
Figure RE-FDA0003097243530000032
Figure RE-FDA0003097243530000033
The residual route set from the source node to the neighbor node of the intermediate node is selected if
Figure RE-FDA0003097243530000041
Finishing the secondary cutting;
if it is
Figure RE-FDA0003097243530000042
4, the residual cost index vector from the source node to the neighbor node of the intermediate node in the step 4 is obtained
Figure RE-FDA0003097243530000044
In a multi-dimensional space HKCutting;
step length changing segmentation is carried out on each dimension axis of the multidimensional space according to the step sequence established in the step 1, and the multidimensional space H is divided into a plurality of step lengthsKDividing the data into a plurality of subspaces, and obtaining the residual cost index vector from the source node to the neighbor node of the intermediate node in the step 4
Figure RE-FDA00030972435300000414
Respectively belong to S subspaces, each subspace respectively contains M1,M2,...,M|S|A vector of residual cost indicators, where a relationship exists
Figure RE-FDA0003097243530000046
For respectively containing M1,M2,...,M|S|Each subspace of cost index vectors, if any subspace F contains the residual cost index vector set as F,
Figure RE-FDA0003097243530000047
the number of the residual cost index vectors in the set is MfThe subspace includes a cost index vector of
Figure RE-FDA0003097243530000048
Setting the weight of each dimension cost in the K dimension cost in the system as { alpha [ ]1,...,αK},(α1+...+αK1), respectively calculating the weighted value of each cost index vector as:
Figure RE-FDA0003097243530000049
wherein the device is provided with a plurality of sensors,
Figure RE-FDA00030972435300000410
is the minimum value, the source node corresponding to the minimum value is connected to the neighbor node of the intermediate node
Figure RE-FDA00030972435300000411
F in (1)zOne route is reserved, M remainsf-1 route all from
Figure RE-FDA00030972435300000412
In deletion, i.e. clipping, Fz∈F。
6. The multi-cost-index-based route cutting optimization method according to claim 1, wherein in step 5, a network node s is selected from the network nodes as a source node, a network node t is selected as a target node, N is the number of nodes in the network, s belongs to [1, N ], t belongs to [1, N ];
in step 5, the cost index vectors of the routes from the source node s to the target node t are respectively
Figure RE-FDA00030972435300000413
Wherein the route set from the source node to the target node is PstThe number of routes, i.e. the number of cost indicator vectors, is
Figure RE-FDA0003097243530000051
Figure RE-FDA0003097243530000052
Setting the weight of each dimension cost in K dimension costs in the system as { alpha ] respectively for a route set from a source node to a target node1,...,αK},α1+...+αKReferring to step 3 and step 4, respectively calculating the weighted value of each cost index vector in each sub-partition as:
Figure RE-FDA0003097243530000053
is provided therein
Figure RE-FDA0003097243530000054
Is the minimum value, corresponding to the source node to the target node
Figure RE-FDA0003097243530000055
Fz middle route is reserved, and the rest is
Figure RE-FDA0003097243530000056
All routes are from
Figure RE-FDA0003097243530000057
In deletion, i.e. clipping, FzAnd E is F, and F is a set of residual cost index vectors.
7. The multi-cost-index-based route pruning optimization method according to claim 1, wherein the number of times that step 5 is iteratively performed in step 6 is N-1, where N is the number of nodes in the network.
CN201810254608.5A 2018-03-26 2018-03-26 Route cutting optimization method based on multi-cost index Active CN110365585B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201810254608.5A CN110365585B (en) 2018-03-26 2018-03-26 Route cutting optimization method based on multi-cost index

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201810254608.5A CN110365585B (en) 2018-03-26 2018-03-26 Route cutting optimization method based on multi-cost index

Publications (2)

Publication Number Publication Date
CN110365585A CN110365585A (en) 2019-10-22
CN110365585B true CN110365585B (en) 2021-08-03

Family

ID=68212779

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201810254608.5A Active CN110365585B (en) 2018-03-26 2018-03-26 Route cutting optimization method based on multi-cost index

Country Status (1)

Country Link
CN (1) CN110365585B (en)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113328950B (en) * 2021-05-25 2022-06-17 桂林电子科技大学 SDN routing system construction method based on tree structure

Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2003061194A2 (en) * 2001-12-26 2003-07-24 Tropic Networks Inc. Multi-constraint routine system and method
CN101778041A (en) * 2009-12-31 2010-07-14 福建星网锐捷网络有限公司 Method, device and network equipment for path selection
CN102523360A (en) * 2012-01-12 2012-06-27 江苏电力信息技术有限公司 Call routing algorithm based on multi-dimensional queue model
CN103036787A (en) * 2011-10-09 2013-04-10 华为技术有限公司 Network route convergence processing method and network route convergence processing device
CN104038418A (en) * 2014-05-19 2014-09-10 暨南大学 Routing method for hybrid topologic structure data center, path detection mechanism and message processing mechanism
EP2940946A1 (en) * 2014-04-30 2015-11-04 Alcatel Lucent Method for operating a communication network
CN105591915A (en) * 2014-10-22 2016-05-18 中兴通讯股份有限公司 Maintenance method and apparatus of routing table
CN105847151A (en) * 2016-05-25 2016-08-10 安徽大学 Multi-constrained QoS (Quality of Service) routing strategy designing method for software defined network

Patent Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2003061194A2 (en) * 2001-12-26 2003-07-24 Tropic Networks Inc. Multi-constraint routine system and method
CN101778041A (en) * 2009-12-31 2010-07-14 福建星网锐捷网络有限公司 Method, device and network equipment for path selection
CN103036787A (en) * 2011-10-09 2013-04-10 华为技术有限公司 Network route convergence processing method and network route convergence processing device
CN102523360A (en) * 2012-01-12 2012-06-27 江苏电力信息技术有限公司 Call routing algorithm based on multi-dimensional queue model
EP2940946A1 (en) * 2014-04-30 2015-11-04 Alcatel Lucent Method for operating a communication network
CN104038418A (en) * 2014-05-19 2014-09-10 暨南大学 Routing method for hybrid topologic structure data center, path detection mechanism and message processing mechanism
CN105591915A (en) * 2014-10-22 2016-05-18 中兴通讯股份有限公司 Maintenance method and apparatus of routing table
CN105847151A (en) * 2016-05-25 2016-08-10 安徽大学 Multi-constrained QoS (Quality of Service) routing strategy designing method for software defined network

Non-Patent Citations (4)

* Cited by examiner, † Cited by third party
Title
A tursted routing protocol for wireless mobile ad hoc networks;Huang Chuanhe et al.;《CCWMSN》;20071214;全文 *
Minimal Road-side unit placement for delay-bounded applicaitons in bus ad-hoc networks;HaizhouBao et al.;《IPCCC》;20171212;全文 *
基于Dijistra算法的多约束多播路由算法的研究;汪胡青;《计算机技术与发展》;20111231;全文 *
基于多路广播树的SDN多路径路由算法;覃匡宇;《计算机科学》;20180131;全文 *

Also Published As

Publication number Publication date
CN110365585A (en) 2019-10-22

Similar Documents

Publication Publication Date Title
Xu et al. Link-state routing with hop-by-hop forwarding can achieve optimal traffic engineering
CN111770019B (en) Q-learning optical network-on-chip self-adaptive route planning method based on Dijkstra algorithm
US20210203557A1 (en) System and method for generating and using physical roadmaps in network synthesis
WO2017045578A1 (en) Topological graph optimal path algorithm with constraint conditions
US10404576B2 (en) Constrained shortest path determination in a network
CN111181792A (en) SDN controller deployment method and device based on network topology and electronic equipment
CN110365585B (en) Route cutting optimization method based on multi-cost index
Becker et al. Near-optimal approximate shortest paths and transshipment in distributed and streaming models
CN102210128A (en) Path calculation order deciding method, program and calculating apparatus
CN110661704B (en) Calculation method of forwarding path and SDN controller
CN112015518B (en) Method and system for realizing real-time migration of multiple virtual machines in incremental deployment SDN environment
CN108683593A (en) A kind of computational methods of K short paths
CN113328950B (en) SDN routing system construction method based on tree structure
CN105515984A (en) Multipath multi-communication means route planning method
EP3961472A2 (en) System and method for generating connectivity models in network synthesis
Foerster et al. Input-dynamic distributed algorithms for communication networks
JP5595342B2 (en) Multiple path search method and apparatus
CN104462829B (en) The processing method of complicated multi-region grid in space propultion solution
JP2013062628A (en) Path rearrangement method and device
Chan et al. An experiment on the performance of shortest path algorithm
JP5898112B2 (en) Network design apparatus and network design program
Rashedi et al. Evolutionary algorithms for solving routing and wavelength assignment problem in optical networks: A comparative study
Xiao et al. Approximation and heuristic algorithms for delay constrained path selection under inaccurate state information
Mikoshi et al. High-speed calculation method for large-scale multi-layer network design problem
Sun et al. A novel multi-objective genetic algorithm for the qos based multicast routing and wavelength allocation problem in wdm network

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant