| CPC H04L 45/70 (2013.01) [H04L 47/24 (2013.01); H04L 49/101 (2013.01); H04W 40/02 (2013.01)] | 5 Claims |
|
1. A service demand-driven routing scheduling and dynamic selection method for an active distribution network, wherein the method comprises:
evaluating a physical importance index and an information importance index of each service traffic according to a physical influence and an information demand of each issued service traffic;
comprehensively taking into account the physical importance index and the information importance index to obtain a comprehensive service traffic importance index;
ranking the comprehensive service traffic importance index to obtain a ranked service traffic set;
quantizing the real-time performance and the reliability of each route link and generating a route distance matrix;
inputting the ranked service traffic set and the route distance matrix into a pre-constructed route scheduling optimization model, and generating an optimal route scheduling scheme of the overall service traffic based on a principle of service traffic importance and route path performance matching; and
scheduling, by a lower router, the traffic according to the optimal route scheduling scheme;
wherein evaluating a physical importance index and an information importance index of each service traffic according to a physical influence and an information demand of each issued service traffic comprises:
using an improved gravity centrality method to evaluate a physical importance, wherein an original gravity centrality formula is:
 where G(i) denotes an obtained node influence index, ks(i) denotes a k-shell value of a physical node i, ks(j) denotes a k-shell value of a physical node j, the k-shell value denotes the influence in the network topology, Ni denotes a neighboring node set of a physical node i, and dij denotes a topological distance between the physical nodes i and j;
improving the formula (1), and obtaining an improved gravity centrality formula:
 where Gp(i) denotes an obtained physical node importance value, pi,norm denotes a power influence of the normalized traffic on the physical node i, pi,normks(i) denotes a comprehensive index of the structural influence and the physical influence of the node i in a network topology, pj,norm denotes a power influence of the normalized traffic on the physical node j, and pj.normks(j) denotes a comprehensive index of the structural influence and the physical influence of the node j in the network topology;
referring to formula (3) for the method of normalizing all variables to be normalized:
 where cnorm denotes a normalized variable value, c denotes the variable to be normalized, cmin denotes a minimum value of the variable c, cmax denotes a maximum value of the variable c, and X denotes the corrected parameter;
using the improved gravity centrality formula (2) to denote the physical importance Ip(k) of the service traffic fk, as shown in formula (4):
Ip(k)=Gp(i) (4)
taking into account the specific real-time demand and reliability demand of different service traffics, constructing a traffic information importance formula (5):
 where k denotes the number of the service traffic, Ic(k) denotes the information importance index of the traffic, τk,norm denotes a normalized deadline demand of the traffic, ek,norm denotes a normalized bit error rate demand of the traffic, α1 and α2 denote weight coefficients which satisfy α1+α2=1;
comprehensively taking into account the physical importance index and the information importance index to obtain a comprehensive service traffic importance index comprises:
based on a traffic physical importance Ip(k) (4) and a traffic information importance (5), giving an information physical comprehensive importance index (6) of the service traffic fk:
Ik=β1Ip(k)+β2Ic(k) (6)
where Ik denotes a comprehensive importance of each traffic taking into account both the physical importance and the information importance, and β1 and β2 denote weight coefficients which satisfy β1+β2=1;
arranging the traffics in a traffic set to be scheduled F in a descending order according to the comprehensive importance Ik of each traffic to obtain a ranked traffic set;
wherein quantizing the real-time performance and the reliability of each route link and generating a route distance matrix comprises:
abstracting an actual router as a communication node, and abstracting a link between two communication nodes as a connecting edge, wherein the communication topology is denoted as G={V,E}, V={u1, u2, . . . , un} denotes a communication node set, un denotes an n-th communication node, E={lij|i≠j} denotes a communication link set, lij denotes a communication link between nodes i and j, an adjacency matrix A=[aij]n×n of the communication network is obtained according to the communication network topology, and when the communication nodes i and j are communicated with each other directly, aij=1, otherwise aij=0; aij denotes the connectivity between nodes i and j, and n denotes a matrix dimension;
each communication link lij has a link state attribute: an available link bandwidth bij, a link transmission delay τij and a link transmission bit error rate eij, and each route link is evaluated taking into account the transmission delay and the bit error rate in the link state, and the evaluation method is shown in formula (7):
 where distanceij denotes an evaluation value of each link, τij,norm denotes a link transmission delay normalized by formula (3), eij,norm denotes a link bit error rate normalized by formula (3), w1 and w2 denote weight coefficients which satisfy w1+w2=1, inf denotes infinity, and the smaller the value distanceij, the better the performance of the link;
calculating the route distance matrix D=[distanceij]n×n according to formula (7) and the adjacency matrix A of the communication network;
wherein the method further comprises:
when a traffic is scheduled, a communication node where the traffic is located having a link failure, so that an original route path is unavailable, wherein at this time, a current communication node acquires real-time link failure information, regards a failed link lij as unreachable, and modifies a corresponding element aij=0 in the adjacency matrix A and a corresponding element distanceij=inf in a route distance matrix;
re-selecting an optimal path based on service traffic importance and route path performance matching, so as to achieve rapid reconstruction of a failed path;
wherein taking into account that an original route scheduling result is not completely unavailable, dynamic selection of a path is only carried out at the failed link, and a complete route path is incapable of being formed, an optimal scheduling result is combined with a reconstruction result to re-form a complete route path from a source node to a destination node of the service traffic, the complete route path is fed back to a control center, and an optimal scheduling route table S is updated;
wherein the route scheduling optimization model has an objective function of:
 where |F| denotes the number of traffics in the traffic set to be scheduled, τk,total denotes a total transmission delay of the service traffic fk on a complete route path, ek,total denotes a total bit error rate of the service traffic fk on a complete route path, and o1 and o2 denote weight coefficients which satisfy o1+o2=1;
the total delay and the total bit error rate of the service traffic on a complete route path are (9) and (10), respectively:
 where lij denotes a link, pathk denotes a complete route path of the service traffic fk, fk denotes a k-th traffic to be scheduled, and F denotes a traffic set to be scheduled;
a transmission delay constraint of the service traffic is (11): each traffic arrives within the deadline;
τk,total≤τk,∀fk∈F (11)
where τk denotes a deadline of the service traffic;
a bit error rate constraint of the service traffic is (12): each traffic ensures that the total bit error rate does not exceed an allowable maximum bit error rate;
ek,total≤ek,∀fk∈F (12)
where ek denotes an allowable maximum bit error rate of the service traffic;
a communication link bandwidth constraint is (13): the total traffic accommodated by each link does not exceed the available link bandwidth;
 where xkij denotes an indicator indicating whether the service traffic fk passes through the communication link lij, when the service traffic fk passes through the communication link lij, xkij=1, otherwise, xkjj=0, bk denotes a bandwidth required for traffic transmission, and bij denotes a total bandwidth of the lij;
a communication node queue constraint is (14): a length of a traffic queue in each communication node does not exceed the allowable maximum queue value;
 where uki denotes an indicator indicating whether the service traffic fk passes through the communication node i, when the service traffic fk passes through the communication node ui, uki=1, otherwise, uki=0, qi,max denotes a maximum traffic storage queue of the communication node i.
|