CN114520939A - Optical network resource allocation method based on index perceptual evolution - Google Patents
Optical network resource allocation method based on index perceptual evolution Download PDFInfo
- Publication number
- CN114520939A CN114520939A CN202210246628.4A CN202210246628A CN114520939A CN 114520939 A CN114520939 A CN 114520939A CN 202210246628 A CN202210246628 A CN 202210246628A CN 114520939 A CN114520939 A CN 114520939A
- Authority
- CN
- China
- Prior art keywords
- service
- service request
- scheme
- path
- allocation
- 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.)
- Granted
Links
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04Q—SELECTING
- H04Q11/00—Selecting arrangements for multiplex systems
- H04Q11/0001—Selecting arrangements for multiplex systems using optical switching
- H04Q11/0005—Switch and router aspects
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04Q—SELECTING
- H04Q11/00—Selecting arrangements for multiplex systems
- H04Q11/0001—Selecting arrangements for multiplex systems using optical switching
- H04Q11/0062—Network aspects
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04Q—SELECTING
- H04Q11/00—Selecting arrangements for multiplex systems
- H04Q11/0001—Selecting arrangements for multiplex systems using optical switching
- H04Q11/0062—Network aspects
- H04Q2011/0079—Operation or maintenance aspects
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04Q—SELECTING
- H04Q11/00—Selecting arrangements for multiplex systems
- H04Q11/0001—Selecting arrangements for multiplex systems using optical switching
- H04Q11/0062—Network aspects
- H04Q2011/0086—Network resource allocation, dimensioning or optimisation
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04Q—SELECTING
- H04Q11/00—Selecting arrangements for multiplex systems
- H04Q11/0001—Selecting arrangements for multiplex systems using optical switching
- H04Q11/0062—Network aspects
- H04Q2011/009—Topology aspects
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
The invention discloses an optical network resource allocation method based on index perception evolution, which mainly solves the problems of long working path, large link load difference and low overall network reliability caused by planning a transmission service scheme in an optical network in the prior art. The implementation scheme is as follows: initializing network information and service request information of resource allocation; preprocessing the service request according to the relevance of the service request to generate a pre-distribution scheme of the service request, wherein each scheme in the current pre-distribution scheme represents routing information of all services; iteratively updating the current pre-allocation scheme by adopting a mixed evolution mode based on indexes to obtain an approximately optimal service request allocation scheme; and according to each scheme in the service request allocation schemes, carrying out resource allocation of the service and placement of the optical network equipment in the network. The invention can more fully and reasonably utilize the optical network resources, simultaneously realize the reduction of the total length of the service working path and the guarantee of the overall reliability of the network, and effectively improve the comprehensive performance of the network.
Description
Technical Field
The invention relates to the technical field of communication, in particular to an optical network resource allocation method based on index adaptive perception evolution in the technical field of network communication. The invention can be used for optimizing various mutually restricted resources in the optical network and performing more optimal resource allocation on the optical network.
Background
In an Optical Network, Network resource optimization is a prerequisite condition for reducing Network cost and improving Network fault tolerance, various indexes of a backbone Network such as Network cost, reliability and load balance can be influenced by selection of a service path and placement of an OTN (Optical Transport Network) device, and the indexes can be abstracted into a plurality of mutually-constrained Network resource targets such as total length of the service path, total Network unreliability and link load difference. In addition, since there is correlation between service requests in a real scene, the correlation between services needs to be considered when the associated services carried on the OTN are transmitted.
A network resource allocation method based on routing and spectrum of energy consumption is disclosed in "resource optimization research under IP + light' cooperative architecture" (doctor paper of tokyo post and telecommunications university, 2019, 6 months) published by tanyanxia. The method comprises the following implementation steps: considering that only a single physical line connection exists between directly connected nodes and NSFNET topology without mutual influence among all the links is considered, and meanwhile, service requests have independence, a network energy consumption model comprising three energy consumption components, namely a bandwidth-variable optical transceiver port, an erbium-doped optical fiber amplifier and a bandwidth-variable optical cross connector is designed; secondly, randomly generating the weight of each node in the network, and generating a transition probability matrix for calculating a routing path according to the weight value of each node; thirdly, combining the routing with minimized energy consumption, the modulation format and the spectrum allocation strategy to obtain a path and a modulation mode solution of each service request; the fourth step: and calculating an adaptive value to obtain an energy consumption value required by the current solution. The method has the disadvantages that the relevance of the service request in the application scene of engineering practice is ignored, so that part of the service request cannot be successfully carried, and the part of the service request is blocked, thereby causing the network resource allocation failure of the part of the service request.
The patent document "a multi-service network resource allocation method" (patent application No. 202111184662.5, application publication No. CN 113709885 a) applied by the university of tokyo post and telecommunications discloses a multi-service bandwidth resource allocation method. The method comprises the following implementation steps: on the premise of meeting the requirement of the transmission rate of the service b, constructing a single-target total delay function containing three delays, namely the transmission delay of a sending terminal access network of the service a, the transmission delay of a core network of the service a and the transmission delay of a receiving terminal access network of the service a by taking the minimum total transmission delay of the service a as an optimization target, and constructing a network slice facing to the service a; secondly, on the premise of meeting the transmission rate requirement of the service b and the added limiting conditions, constructing a network slice facing the service b; and thirdly, distributing network resources according to the network slice facing the service a and the network slice facing the service b. The method has the defects that the problem of three-target optical network resource allocation is converted into the problem of single-target optical network resource allocation, and only a single optical network resource allocation scheme is obtained, so that various corresponding optical network resource allocation schemes cannot be obtained at one time when the target requirements are different.
The west ann electronic technology university discloses a multi-target SDN controller placement method based on evolution awareness in the patent document "multi-target controller placement method based on evolution awareness in network" (patent application No. 202110284085.0, application publication No. CN 113037425A). The method comprises the following specific steps: initializing network topology information and algorithm parameter setting information and preprocessing the initialized information; the second step: generating an initial population by adopting a clustering and uniform design method, and representing an SDN controller deployment scheme by each individual in the current population; the third step: iteratively evolving the current population by using the information entropy index; the fourth step: carrying out validity check on the current population and correcting illegal individuals; the fifth step: and calculating to obtain an approximate optimal SDN controller placement scheme set according to the next generation of population. The method has the disadvantages that the information entropy index is only used for measuring the distribution of the result, the convergence of the result is neglected, and the final result obtained by the evolution is not the optimal scheme set.
Disclosure of Invention
The invention aims to provide an optical network resource allocation method based on index perception evolution aiming at the defects of the prior art, which is used for solving the problem that the service blockage is increased due to neglecting the relevance of a service request during service resource allocation, and the problem that the obtained optical network resource allocation scheme is single when the optical network resource allocation is solved in a single-target mode, and the final result obtained by evolution is not the optimal scheme set because the distribution of the result is measured only by using information entropy indexes in the algorithm evolution process.
The idea for realizing the purpose of the invention is that for all the unassociated service requests, the invention directly utilizes the K shortest path algorithm to select the candidate paths for all the unassociated service requests, and for the service requests of all the associated services, the K shortest path algorithm is firstly utilized to respectively select the candidate paths for any one service request in each pair of associated services, and then the K shortest path algorithm is utilized to select the candidate path for the other service request in the pair of associated services, thereby obtaining the service path set of the pair of associated services. Thereby causing the problem of network resource allocation failure of part of service requests. The invention abstracts three objective functions of the total length of a service routing scheme path, the load difference of a resource distribution optical network link and the total unreliability of the service routing scheme in a resource distribution optical network, and then calculates three objective function values of each service routing scheme in a pre-distribution scheme by using the three objective functions. The invention uses the MOEA/D algorithm to adopt single-point cross variation, uses the NSGA-II algorithm to adopt multi-point cross variation, and uses the ultra-volume and information entropy index to measure the performance of the current result so as to select an evolution mode.
The specific steps for realizing the purpose of the invention are as follows:
step 1, generating a network topology matrix and a service source matrix:
step 1.1, constructing a network topology matrix with N rows and M columns, wherein each row of the matrix represents an initial node in a resource allocation optical network, and each column represents a termination node in the resource allocation optical network;
step 1.2, constructing a service source matrix with R rows and 4 columns, wherein each row of the matrix represents a service request, the 1 st column of the matrix represents a source node of the service request, the 2 nd column represents a destination node of the service request, and the 3 rd column represents a service request number, and the 4 th column represents a service request associated service number;
step 2, preprocessing the service request of the non-associated service:
selecting K shortest paths for each unrelated service request by using a K shortest path algorithm to obtain a service path set of the unrelated service;
step 3, preprocessing the service request of the associated service:
step 3.1, selecting K shortest paths for any service request C in each pair of associated services by utilizing a K shortest path algorithm to obtain a candidate path set of the service request C;
step 3.2, selecting each path in the candidate path set of the service request C, updating the network topology matrix, and obtaining a candidate path set of another service request C' in each pair of associated services under each path by using a K shortest path algorithm;
Step 3.3, in each pair of associated services, the candidate path set of the service request C and the candidate path set of the service request C' form a service path set of the associated services after pretreatment;
step 4, generating a service request pre-allocation scheme:
randomly selecting a pre-distribution path for all the unrelated service requests from the service path set of the unrelated service, randomly selecting a pre-distribution path for all the related service pairs from the service path set of the related service, forming a service routing scheme by the pre-distribution path of the unrelated service request and the pre-distribution paths of all the related service pairs, and combining all the service routing schemes to form the pre-distribution scheme of the service request;
and 5, calculating three objective function values of each service routing scheme in the pre-allocation scheme.
Calculating three objective function values of each service routing scheme in the pre-allocation scheme according to the following formula:
wherein, FgAn objective function, R, representing the g-th service routing schemegRepresents the total number of service requests in the g-th service routing scheme, sigma represents the summation operation, Lg,zIndicates the path length of the z-th service request in the g-th service routing scheme, where z is 1 g(ii) a E denotes the total number of links in the resource-allocated optical network,
indicating whether the e link exists in the path selected by the b service request in the g service routing scheme or not, if yes, judging that the e link exists in the path selected by the b service request in the g service routing scheme
Is set to 1, otherwise to 0, E-1, E, b-1, RgX represents a multiplication operation, Pg,eIndicating the packet loss rate of the e link in the g service routing scheme,
indicating whether the p link exists in the path selected by the c service request in the g service routing scheme or not, if yes, the p link is judged to exist in the path selected by the c service request in the g service routing scheme
Is set to 1, otherwise is set to 0, -represents a subtraction operation，
Indicating whether the q link exists in the path selected by the d service request in the g service routing scheme or not, if yes, the q link is judged to exist in the path selected by the d service request in the g service routing scheme
Is set to 1, otherwise 0, q 1, E, d 1, Rg，[·]2Represents a squaring operation,/represents a dividing operation, (.)TRepresenting a transpose operation.
And 6, updating the pre-allocation scheme by using a single-point cross variation method and an MOEA/D algorithm:
step 6.1, selecting two unselected service routing schemes P and Q from the current pre-allocation scheme, randomly selecting a service request from P, replacing the path number of the service request selected in P by the path number of the service request in Q, which has the same type as the service request selected by P, to obtain a crossed service routing scheme P ', and randomly generating an integer to replace the path number of the service associated with the selected service request in P' if the selected service request has associated services;
Step 6.2, randomly selecting a service request from P ', randomly generating an integer to replace the path number of the selected service request in P ', and obtaining a service routing scheme after P ' variation;
step 6.3, updating the service routing scheme in the current pre-allocation scheme by using the varied service routing scheme by using a three-objective function and an MOEA/D algorithm to obtain an updated pre-allocation scheme set;
step 6.4, judging whether all the service routing schemes in the current pre-allocation scheme set are selected, if so, executing step 7, otherwise, executing step 6.1;
step 7, judging whether the updated pre-allocation scheme set is completely the same as the updated pre-allocation scheme set, if so, executing step 12, otherwise, executing step 8;
step 8, calculating the super volume and the super volume change rate of the pre-distribution scheme;
step 9, judging whether the pre-distribution scheme meets the over-volume conversion condition, if so, executing step 10, otherwise, executing step 6;
step 10, updating the pre-allocation scheme by using a multipoint cross mutation method and an NSGA-II algorithm:
step 10.1, selecting two unselected service routing schemes B and A from the current pre-allocation scheme;
step 10.2, selecting a service request to be processed from B, generating a random integer, if the random number is less than or equal to the multipoint cross variation threshold, replacing the path number of the service request selected in B with the path number of the service request in A, which has the same type as the service request selected by B, and if the selected service request has associated services, replacing the path number of the service associated with the selected service request in B with the randomly generated integer;
Step 10.3, judging whether all the services in the service routing scheme B are selected, if so, obtaining a crossed service routing scheme B', otherwise, executing step 10.2;
step 10.4, selecting a service request to be processed from B ', generating a random integer, and if the random integer is less than or equal to the multipoint cross variation threshold, replacing the path number of the selected service request in B' with the randomly generated integer;
step 10.5, judging whether all the services in the service routing scheme B' are selected, if so, generating a variant service routing scheme, otherwise, executing step 10.4;
step 10.6, judging whether all the service routing schemes in the current pre-allocation scheme set are selected, if so, combining all the varied service routing schemes to obtain the varied pre-allocation scheme, otherwise, executing step 8.1;
step 10.7, updating the current pre-allocation scheme by using the mutated pre-allocation scheme by using a three-objective function and an NSGA-II algorithm to obtain an updated pre-allocation scheme;
step 10.8, the information entropy formula is utilized to calculate the information entropy value of the updated pre-distribution scheme
Step 11, judging whether the pre-allocation scheme meets the information entropy conversion condition, if so, executing step 7, otherwise, executing step 10;
And 12, outputting an optimal service request pre-allocation scheme.
Compared with the prior art, the invention has the following advantages:
firstly, the invention respectively carries out route selection pretreatment on the non-associated service request and the associated service request, and carries out route selection for all the service requests in advance from the angle of the relevance of the service requests, thereby overcoming the defects that in the prior art, the relevance of the service requests in the actual scene is ignored, partial services cannot be successfully established, partial service requests are blocked, and further the allocation of optical network resources fails, so that the invention can process the service requests with more complex requirements in the resource allocation optical network, and has wider application range.
Secondly, the invention abstracts three optimization targets of each service routing scheme in the resource allocation optical network into a three-target function, overcomes the problem that the resource allocation problem of the three-target optical network is converted into single-target resource allocation and the obtained optical network resource allocation scheme is single in the prior art, and ensures that the invention can obtain the corresponding optical network resource allocation scheme when the service requirements are different.
Thirdly, the invention selects a proper service routing scheme by combining single-point cross variation and multipoint cross variation with index perception evolution, overcomes the defect that the obtained result can not be completely converged due to the fact that the result is measured by using only information entropy indexes in the prior art, improves the convergence of the searched scheme while considering the distribution of the searched scheme, and can possibly find more and more optimal resource allocation schemes.
Drawings
FIG. 1 is a flow chart of the present invention;
FIG. 2 is a comparison graph of simulation of the present invention.
Detailed Description
The following describes the implementation steps of the present invention in further detail with reference to the accompanying drawings and specific examples, but the embodiments of the present invention are not limited thereto.
The specific steps of the implementation of the present invention are further described with reference to fig. 1 and the examples.
Step 1: generating a network topology matrix and a traffic source matrix
Step 1, constructing a network topology matrix with N rows and M columns, wherein each row of the matrix represents a starting node in the resource allocation optical network, and each column represents a terminating node in the resource allocation optical network. If the starting node and the terminating node are two end points of the same link respectively, setting element values of a row where the starting node is located and a column where the terminating node is located to be 1 in a network topology matrix; otherwise, setting the element value to 0, wherein T represents the total number of nodes in the resource allocation optical network, N is T, and M is T, and taking the network topology matrix as the input of the subsequent routing algorithm.
Step 2, constructing a service source matrix with R rows and 4 columns, wherein each row of the matrix represents a service request, the 1 st column of the matrix represents a source node of the service request, the 2 nd column represents a destination node of the service request, the 3 rd column represents a service request number, and the 1 st, 2 nd and 3 rd columns of data are all generated at random with equal probability in a given interval; column 4 represents the service request associated service number, which means if the R-th service is 1The row represents the number r1Is the R < th > service request2The number represented by a row is r2Is associated with the service request, then the service source matrix is located at (R)1And 4) setting the element value of the position as r3，r3＝r2While in the traffic source matrix is located at (R)2And 4) setting the element value of the position as r4，r4＝r1，1≤R1≤R，1≤R2R is less than R, otherwise, the service source matrix is positioned in (R)1And 4) setting the value of the element at the position to 0 represents the R-th1The number represented by a row is r1The service request of (2) has no associated service. Each service request can only be associated with another service request at most, wherein R is S, and S represents the total number of the service requests.
And 2, preprocessing the service request without the associated service.
Selecting K shortest paths for each unrelated service request by using a K shortest path algorithm, performing ascending arrangement according to the length of each path in the K shortest paths to obtain a candidate path set of the service request, forming the candidate path set of all the unrelated service requests into a service path set of the preprocessed unrelated service, wherein K represents the number of the shortest paths.
And 3, preprocessing the service request of the associated service.
And step 1, selecting K shortest paths for any service request C in each pair of associated services by using a KSP algorithm, and performing ascending arrangement according to the length of each path in the K shortest paths to obtain a candidate path set of the service request C.
And 2, randomly selecting an unprocessed path from the candidate path set of the service request C, and setting element values of a row where a starting node of each link in the selected path is located and a row where a terminating node of the link is located as 0 in the network topology matrix to obtain an updated network topology matrix.
And step 3, obtaining a candidate path set of another service request C' in each pair of associated services under the selected path in the same way as the step 1 in the step.
And step 4, generating a candidate path set of another service request C' in each pair of associated services by adopting the same operation as the steps 2 to 3.
And 5, combining the candidate path set of the service request C and the candidate path set of the service request C' in each pair of associated services into a service path set of the associated services after preprocessing.
And 4, generating a pre-allocation scheme of the service request by utilizing the service request set of the non-associated service and the service request sets of all the associated services.
Randomly selecting a pre-distribution path for all the unrelated service requests from the service path set of the unrelated service, randomly selecting a pre-distribution path for all the related service pairs from the service path set of the related service, forming a service routing scheme by the pre-distribution path of the unrelated service request and the pre-distribution paths of all the related service pairs, generating at least 2 service routing schemes, and combining all the service routing schemes to form the pre-distribution scheme of the service requests.
In the present example, the implementation of generating a set of pre-allocation schemes of 120 service requests is as follows. For each pre-allocation scheme, random numbers of service requests of all non-associated services are randomly generated, the random number of the service request of each non-associated service is represented in a service request set of the non-associated service, and the service request selects a candidate path with the number of the random number. Then a service request r is generated in each pair of associated services5Random number y of1And another service request r6Random number y of2Random number y1Indicating a service request r in a service request set associated with a service5Choose the random number y1Candidate path of (3), random number y2Representing a service request r6At service request r5Choose the random number y1Under the condition of the candidate path, the random number y numbered in the service request set of the selected associated service1The candidate route of (2).
And 5, generating three objective functions of each service routing scheme in the pre-distribution scheme.
The three objective functions for each service routing scheme are generated as follows:
wherein, FgAn objective function, R, representing the g-th service routing schemegRepresents the total number of service requests in the g-th service routing scheme, sigma represents the summation operation, L g,zIndicates the path length of the z-th service request in the g-th service routing scheme, and z is 1g(ii) a E denotes the total number of links in the resource allocation optical network,
indicating to judge whether the e link exists in the g serviceIn the path selected by the b-th service request in the routing scheme, if yes, the path is selected
Is set to 1, otherwise 0, E1, E, b 1, RgX represents a multiplication operation, Pg,eIndicating the packet loss rate of the e link in the g service routing scheme,
indicating whether the p link exists in the path selected by the c service request in the g service routing scheme or not, if yes, the p link is judged to exist in the path selected by the c service request in the g service routing scheme
Is set to 1, otherwise is set to 0, -represents a subtraction operation,
indicating whether the q link exists in the path selected by the d service request in the g service routing scheme or not, if yes, the q link is judged to exist in the path selected by the d service request in the g service routing scheme
Is set to 1, otherwise 0, q 1, E, d 1, Rg，[·]2Represents a squaring operation,/represents a dividing operation, (.)TRepresenting a transpose operation.
And 6, updating the pre-allocation scheme by using a single-point cross variation method and an MOEA/D algorithm.
Step 1, two to-be-processed service routing schemes P and Q are selected from the current pre-allocation scheme.
And step 2, randomly selecting a service request from P, replacing the path number of the service request selected in P by the path number of the service request with the same type as the service request selected by P in Q, and obtaining a service routing scheme P' after intersection.
Step 3, if the selected service request has associated service, using randomly generated integer w1Replacing services associated with selected service requests in P1 is not more than w1And ≦ K, the associated service of the service request refers to the service request in contact with the service request.
Step 4, a service request randomly selected from P' is selected by using randomly generated integer w2Replacing the path number of the selected service request in P', w is more than or equal to 12And (5) obtaining the service routing scheme after P' variation when the number is less than or equal to K.
And 5, updating the service routing scheme in the current pre-allocation scheme by using the varied service routing scheme by using a three-objective function and an MOEA/D algorithm to obtain an updated pre-allocation scheme set.
And 6, judging whether all the service routing schemes in the current pre-distribution scheme set are selected, if so, executing the step 7, otherwise, executing the step 1.
And 7, judging whether the updated pre-allocation scheme set is completely the same as the updated pre-allocation scheme set, if so, executing the step 12, otherwise, executing the step 8.
And 8, calculating the super-volume and the super-volume change rate of the pre-distribution scheme.
The hyper-volume and rate of change of the hyper-volume of the pre-allocation scheme are calculated as follows:
vh＝1-HVh/HVh′
wherein, HVtIndicating the ultra volume value of the tth pre-distribution scheme, omega indicating the total number of Pareto solution sets in the tth pre-distribution scheme, U indicating the operation of solution sets, i indicating the serial number of the Pareto solution sets in the tth pre-distribution scheme, Vt,iRepresents the hyper-volume, v, enclosed by the i-th solution in the Pareto solution set of the t-th pre-allocation scheme and the maximum target value in the solution sethRepresenting the rate of change of the over-volume, HV, of the h-th pre-allocation schemehThe method comprises the steps of utilizing a three-objective function and an MOEA/D algorithm, and updating a service routing scheme in a current pre-allocation scheme by using a mutated service routing scheme to obtain an updated service routing schemeThe hyper-volume value, HV, of the last h pre-allocation schemeh' denotes the over-volume value before the h-th pre-allocation scheme update.
And 9, judging whether the pre-distribution scheme meets the over-volume conversion condition, if so, executing the step 10, and otherwise, executing the step 7.
The hyper-volume conversion condition refers to a situation that the following 2 conditions are satisfied simultaneously:
condition 1: whether the super-volume of the pre-allocation scheme is greater than a super-volume transform threshold;
Condition 2: whether the super-volume change rate of the pre-distribution scheme before updating and the pre-distribution scheme after updating is smaller than a super-volume change rate conversion threshold value or not is judged, and the super-volume threshold value and the super-volume change rate threshold value are set according to historical multi-objective optimization index empirical data.
And step 10, updating the pre-allocation scheme by using a multipoint cross mutation method and an NSGA-II algorithm.
Step 1, two to-be-processed service routing schemes B and A are selected from a current pre-allocation scheme set.
Step 2, selecting a service request to be processed from B, and generating a random integer u1，1≤u1U is judged to be less than or equal to 1001Whether or not it is less than or equal to the multipoint cross variation threshold PcoIf yes, executing the step 2 of the step, otherwise, executing the step 3 of the step.
And step 3, replacing the path number of the service request selected in the B by the path number of the service request with the same type as the service request selected in the B in the A.
Step 4, if the selected service request has associated service, using randomly generated integer u2Replacing the path number of the service associated with the selected service request in B, u being more than or equal to 12And ≦ K, the associated service of the service request refers to the service request in contact with the service request.
And 5, judging whether all the services in the service routing scheme B are selected, if so, obtaining a crossed service routing scheme B', and otherwise, executing the step 2 of the step.
Step 7, using randomly generated integer u4Replacing the path number of the selected service request in B', u is more than or equal to 14≤K。
And 8, judging whether all the services in the service routing scheme B' are selected, if so, generating a varied service routing scheme, and otherwise, executing the step 6 of the step.
And 9, judging whether all the service routing schemes in the current pre-allocation scheme set are selected, if so, combining all the varied service routing schemes to obtain the varied pre-allocation scheme, and otherwise, executing the step 1 of the step.
And step 10, updating the current pre-allocation scheme by using the mutated pre-allocation scheme by using a three-objective function and an NSGA-II algorithm to obtain an updated pre-allocation scheme.
And 11, calculating the information entropy value of the updated pre-distribution scheme by using an information entropy formula.
And 11, judging whether the pre-allocation scheme meets the information entropy transformation condition, if so, executing the step 6, otherwise, executing the step 10.
The information entropy change condition refers to a situation that satisfies the following condition:
conditions are as follows: and whether the information entropy index of the current pre-allocation scheme is larger than an information entropy threshold value, wherein the value of the threshold value is 0.4.
And 12, outputting the optimal service request distribution scheme.
The effects of the present invention can be further explained by the following simulation experiments.
The effect of the present invention can be further demonstrated by the following simulation.
1. And (5) simulating experimental conditions.
The software platform of the simulation experiment of the invention is as follows: windows 10 operating system and Matlab R2020 a.
The data of the simulation experiment of the invention is Internet2 IP with 9 nodes and 17 links as the resource distribution optical network topology, the total number of the pre-distribution schemes of the service request is 120, the total number of the service request is 300, the multi-point cross variation threshold is Pco300, the shortest path number K is 3, the super volume threshold is set to 0.5, the super volume change rate threshold is set to 0.1, the information entropy threshold is set to 0.4, and each target value is composed of the total length of the path of the traffic routing scheme, the total unreliability of the traffic routing scheme, and the load difference of the resource allocation optical network link.
2. And (5) simulating content and result analysis.
The simulation experiment 1 of the present invention is to adopt the method of the present invention and two existing technologies to respectively simulate all service requests in the resource allocation optical network topology to obtain three service request allocation schemes, then respectively calculate three objective function values of the three service request allocation schemes, and draw the obtained three objective function values into three curves as shown in fig. 2.
The prior art 1 refers to an open-source decomposition-based multi-objective evolutionary algorithm MOEA/D in Matlab PlatEMO.
The prior art 2 refers to an open-source multi-objective evolutionary algorithm NSGA-II based on non-dominated sorting and elite selection strategies in PlatEMO of Matlab.
The effect of the present invention is further described below with reference to the simulation diagram of fig. 2.
The abscissa in fig. 2 represents the total length of the path of the traffic routing scheme in the three-objective function, the ordinate represents the total unreliability of the traffic routing scheme in the three-objective function, and the ordinate represents the load difference of the resource allocation optical network link in the three-objective function. The curve marked by a triangle represents a three-objective function value curve of the service request allocation scheme obtained by simulation by the method, the curve marked by a square represents the three-objective function value curve of the service request allocation scheme obtained by the prior art 1, and the curve marked by a five-pointed star represents the three-objective function value curve of the service request allocation scheme obtained by the prior art 2. As can be seen from fig. 2, compared with the existing method, the present invention can obtain smaller overall unreliability of the service routing scheme under the condition of the total length of the paths of the service routing scheme and the load difference of the links of the resource allocation optical network, and when the total length of the paths of the service routing scheme is similar, the load difference and the overall unreliability of the links of the resource allocation optical network obtained by the present invention are lower. The resource allocation designed according to the invention can improve the operation cost of the network, reduce the overall risk degree of the network, promote the reasonable utilization of the resources and optimize the performance of the network.
Claims (6)
1. An optical network resource allocation method based on index sensing evolution is characterized in that routing pretreatment is respectively carried out on non-associated service requests and associated service requests to generate three objective functions of each service routing scheme, and a proper service routing scheme is selected by utilizing a mode of combining single-point cross variation and multipoint cross variation with index sensing evolution, wherein the method comprises the following steps:
step 1, generating a network topology matrix and a service source matrix:
step 1.1, constructing a network topology matrix with N rows and M columns, wherein each row of the matrix represents an initial node in a resource allocation optical network, and each column represents a termination node in the resource allocation optical network;
step 1.2, a service source matrix with R rows and 4 columns is constructed, each row of the matrix represents a service request, the 1 st column of the matrix represents a source node of the service request, the 2 nd column represents a sink node of the service request, the 3 rd column represents a service request number, and the 4 th column represents a service request correlation service number;
step 2, preprocessing the service request of the non-associated service:
selecting K shortest paths for each unrelated service request by using a K shortest path algorithm to obtain a service path set of the unrelated service;
Step 3, preprocessing the service request of the associated service:
step 3.1, selecting K shortest paths for any service request C in each pair of associated services by utilizing a K shortest path algorithm to obtain a candidate path set of the service request C;
step 3.2, selecting each path in the candidate path set of the service request C, updating the network topology matrix, and obtaining a candidate path set of another service request C' in each pair of associated services under each path by using a K shortest path algorithm;
step 3.3, in each pair of associated services, the candidate path set of the service request C and the candidate path set of the service request C' form a service path set of the associated services after pretreatment;
step 4, generating a service request pre-allocation scheme:
randomly selecting a pre-distribution path for all the uncorrelated service requests from the service path set of the uncorrelated service, randomly selecting a pre-distribution path for all the correlated service pairs from the service path set of the correlated service, forming the pre-distribution paths of the uncorrelated service requests and the pre-distribution paths of all the correlated service pairs into a service routing scheme, and combining all the service routing schemes into the pre-distribution scheme of the service requests;
Step 5, calculating three objective function values of each service routing scheme in the pre-allocation scheme according to the following formula:
wherein, FgAn objective function, R, representing the g-th traffic routing schemegRepresenting the total number of service requests in the g-th service routing scheme, sigma, the summation operation, Lg,zIndicates the path length of the z-th service request in the g-th service routing scheme, where z is 1g(ii) a E denotes the total number of links in the resource allocation optical network,
indicating whether the e link exists in the path selected by the b service request in the g service routing scheme or not, if yes, the e link is judged to exist in the path selected by the b service request in the g service routing scheme
Is set to 1, otherwise 0, E1, E, b 1, RgX represents a multiplication operation, Pg,eIndicating the packet loss rate of the e link in the g service routing scheme,
indicating whether the p link exists in the path selected by the c service request in the g service routing scheme or not, if yes, the p link is judged to exist in the path selected by the c service request in the g service routing scheme
Is set to 1, otherwise is set to 0, -represents a subtraction operation,
indicating whether the q link exists in the path selected by the d service request in the g service routing scheme or not, if yes, the q link is judged to exist in the path selected by the d service request in the g service routing scheme
Is set to 1, otherwise 0, q 1, E, d 1, R g，[·]2Represents a squaring operation,/represents a dividing operation, (. cndot.)TRepresenting a transpose operation;
and 6, updating a pre-allocation scheme by using a single-point cross variation method and an MOEA/D algorithm:
step 6.1, selecting two unselected service routing schemes P and Q from the current pre-allocation scheme, randomly selecting a service request from P, replacing the path number of the service request selected in P by the path number of the service request in Q, which has the same type as the service request selected by P, to obtain a crossed service routing scheme P ', and randomly generating an integer to replace the path number of the service associated with the selected service request in P' if the selected service request has associated services;
step 6.2, randomly selecting a service request from the P ', randomly generating an integer to replace the path number of the selected service request in the P ', and obtaining a service routing scheme after P ' variation;
step 6.3, updating the service routing scheme in the current pre-allocation scheme by using the varied service routing scheme by using a three-objective function and an MOEA/D algorithm to obtain an updated pre-allocation scheme set;
step 6.4, judging whether all the service routing schemes in the current pre-allocation scheme set are selected, if so, executing step 7, otherwise, executing step 6.1;
Step 7, judging whether the updated pre-allocation scheme set is completely the same as the updated pre-allocation scheme set, if so, executing step 12, otherwise, executing step 8;
step 8, calculating the super volume and the super volume change rate of the pre-distribution scheme;
step 9, judging whether the pre-distribution scheme meets the over-volume conversion condition, if so, executing step 10, otherwise, executing step 6;
step 10, updating the pre-allocation scheme by using a multipoint cross mutation method and an NSGA-II algorithm:
step 10.1, selecting two unselected service routing schemes B and A from the current pre-allocation scheme;
step 10.2, selecting a service request to be processed from B, generating a random integer, if the random number is less than or equal to the multipoint cross variation threshold, replacing the path number of the service request selected in B with the path number of the service request in A which has the same type as the service request selected in B, and if the selected service request has associated services, replacing the path number of the service associated with the selected service request in B with the randomly generated integer;
step 10.3, judging whether all the services in the service routing scheme B are selected, if so, obtaining a crossed service routing scheme B', otherwise, executing step 10.2;
Step 10.4, selecting a service request to be processed from B ', generating a random integer, and replacing the path number of the selected service request in B' with the randomly generated integer if the random integer is less than or equal to the multipoint cross variation threshold;
step 10.5, judging whether all the services in the service routing scheme B' are selected, if so, generating a variant service routing scheme, otherwise, executing step 10.4;
step 10.6, judging whether all the service routing schemes in the current pre-allocation scheme set are selected, if so, combining all the varied service routing schemes to obtain the varied pre-allocation scheme, otherwise, executing step 10.1;
step 10.7, updating the current pre-allocation scheme by using the mutated pre-allocation scheme by using a three-objective function and an NSGA-II algorithm to obtain an updated pre-allocation scheme;
step 10.8, the information entropy formula is utilized to calculate the information entropy value of the updated pre-distribution scheme
Step 11, judging whether the pre-allocation scheme meets the information entropy transformation condition, if so, executing step 7, otherwise, executing step 10;
and 12, outputting the optimal service request pre-allocation scheme.
2. The method for allocating optical network resources based on index-aware evolution of claim 1, wherein: the network topology matrix of N rows and M columns described in step 1.1 is generated according to the following formula: if the starting node and the terminating node are the same link, setting the element values of the row where the starting node is located and the column where the terminating node is located to be 1 in the network topology matrix; otherwise, the element value is set to 0, T represents the total number of nodes in the resource allocation optical network, N equals T, and M equals T.
3. The method for allocating optical network resources based on index-aware evolution of claim 1, wherein: the R row and 4 column traffic source matrix in step 1.2 is generated according to the following formula: the 1 st, 2 nd and 3 rd data are all randomly generated with equal probability in a given interval; column 4 represents the service request associated service number, which means if the R-th service is1The number represented by a row is r1Is the R < th > service request2The number represented by a row is r2Associated with the service request, then bits in the service source matrixIn (R)1And 4) setting the element value of the position as r3，r3＝r2While in the traffic source matrix is located at (R)2And 4) setting the element value of the position as r4，r4＝r1，1≤R1≤R，1≤R2R is less than R, otherwise, the service source matrix is positioned in (R)1And 4) setting the value of the element at the position to 0 represents the R-th1The number represented by a row is r1The service request of (2) has no associated service, each service request can only be associated with another service request at most, wherein R is S, and S represents the total number of the service requests.
4. The method for allocating optical network resources based on index-aware evolution of claim 1, wherein: the calculation of the super-volume and super-volume rate of change for each pre-dispense protocol described in step 8 is given by:
vh＝1-HVh/HVh′
Wherein HVtIndicating the hyper-volume value of the tth pre-allocation scheme, omega indicates the total number of Pareto solutions in the tth pre-allocation scheme, U indicates a union operation, i indicates the sequence number of the solutions in the Pareto solutions in the tth pre-allocation scheme, Vt,iRepresents the hyper-volume, v, enclosed by the ith solution in the Pareto solution set of the tth pre-allocation scheme and the maximum target value in the solution sethRepresents the hyper-volume rate of change, HV, of the h-th pre-dispense protocolhThe representation utilizes a three-objective function and an MOEA/D algorithm, a service routing scheme in the current pre-allocation scheme is updated by a varied service routing scheme to obtain an ultra-volume value of the updated h pre-allocation scheme, and HVh' represents the over-volume value before the h-th pre-allocation scheme update.
5. The method for allocating optical network resources based on index-aware evolution of claim 1, wherein: the condition for the ultra-volume change in step 9 is a case where the following 2 conditions are satisfied simultaneously:
condition 1: whether the super-volume of the pre-allocation scheme is greater than a super-volume transform threshold;
condition 2: whether the super-volume change rate of the pre-update pre-allocation scheme and the post-update pre-allocation scheme is less than a super-volume change rate conversion threshold;
the ultra-volume threshold and the ultra-volume change rate threshold are set according to empirical data of historical multi-objective optimization indexes.
6. The method for allocating optical network resources based on index aware evolution of claim 1, wherein: the information entropy change condition in step 11 refers to: the information entropy index of the current pre-allocation scheme is greater than an information entropy threshold, and the value of the threshold is 0.4.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210246628.4A CN114520939B (en) | 2022-03-14 | 2022-03-14 | Optical network resource allocation method based on index perceptual evolution |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210246628.4A CN114520939B (en) | 2022-03-14 | 2022-03-14 | Optical network resource allocation method based on index perceptual evolution |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114520939A true CN114520939A (en) | 2022-05-20 |
| CN114520939B CN114520939B (en) | 2023-02-10 |
Family
ID=81598357
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210246628.4A Active CN114520939B (en) | 2022-03-14 | 2022-03-14 | Optical network resource allocation method based on index perceptual evolution |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114520939B (en) |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110011922A (en) * | 2019-04-07 | 2019-07-12 | 西安电子科技大学 | Using the RSA method of traffic forecast and frequency spectrum conversion in elastic optical network |
| US20200266894A1 (en) * | 2017-12-13 | 2020-08-20 | Soochow University | Optical network planning method for asymmetric traffic transmission over multi-core fiber optical network and network using the same |
| CN112367572A (en) * | 2020-10-27 | 2021-02-12 | 西安电子科技大学 | Multi-target RSCA method based on fiber core variability in elastic optical network |
| CN112911429A (en) * | 2021-02-04 | 2021-06-04 | 烽火通信科技股份有限公司 | Routing frequency spectrum allocation method and device based on dynamic elastic optical network |
| EP3883258A1 (en) * | 2020-03-20 | 2021-09-22 | Nokia Solutions and Networks Oy | An optical network unit and an optical line terminal |
| WO2022042262A1 (en) * | 2020-08-31 | 2022-03-03 | 中兴通讯股份有限公司 | Otn network resource optimization method and apparatus, computer device, and storage medium |
-
2022
- 2022-03-14 CN CN202210246628.4A patent/CN114520939B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20200266894A1 (en) * | 2017-12-13 | 2020-08-20 | Soochow University | Optical network planning method for asymmetric traffic transmission over multi-core fiber optical network and network using the same |
| CN110011922A (en) * | 2019-04-07 | 2019-07-12 | 西安电子科技大学 | Using the RSA method of traffic forecast and frequency spectrum conversion in elastic optical network |
| EP3883258A1 (en) * | 2020-03-20 | 2021-09-22 | Nokia Solutions and Networks Oy | An optical network unit and an optical line terminal |
| WO2022042262A1 (en) * | 2020-08-31 | 2022-03-03 | 中兴通讯股份有限公司 | Otn network resource optimization method and apparatus, computer device, and storage medium |
| CN112367572A (en) * | 2020-10-27 | 2021-02-12 | 西安电子科技大学 | Multi-target RSCA method based on fiber core variability in elastic optical network |
| CN112911429A (en) * | 2021-02-04 | 2021-06-04 | 烽火通信科技股份有限公司 | Routing frequency spectrum allocation method and device based on dynamic elastic optical network |
Non-Patent Citations (2)
| Title |
|---|
| THOMAS D. LAGKAS: "《Joint Spatial and Spectral Resource Optimization over Both Wireless and Optical Fronthaul Domains of 5G Architectures》", 《2020 22ND INTERNATIONAL CONFERENCE ON TRANSPARENT OPTICAL NETWORKS (ICTON)》 * |
| 宣贺君等: "弹性光网络中考虑节点安全性的频谱分配算法", 《中国激光》 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114520939B (en) | 2023-02-10 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101777990B (en) | Method for selecting multi-objective immune optimization multicast router path | |
| CN103428704B (en) | A kind of frequency spectrum sensing method and device | |
| CN104994021A (en) | Method and device for determining optimal path | |
| CN111245722B (en) | SDN data center network flow forwarding method based on genetic algorithm | |
| CN109951391B (en) | Network path rapid calculation method based on multi-QoS constraint | |
| CN101013955B (en) | Fast simulated annealing for traffic matrix estimation | |
| KR20090103188A (en) | Routing Tree Searching Method Using Multi-Objective Genetic Algorithm and Corresponding Sensor Network | |
| CN109861925A (en) | Data transmission method, relevant apparatus and network | |
| CN113472671A (en) | Method and device for determining multicast route and computer readable storage medium | |
| Zhang et al. | Fine-grained service offloading in B5G/6G collaborative edge computing based on graph neural networks | |
| CN114520939B (en) | Optical network resource allocation method based on index perceptual evolution | |
| CN104009943B (en) | LLDP message transmitting methods and DCB equipment | |
| CN115499306B (en) | Method and device for constructing flow scheduling model, electronic equipment and storage medium | |
| CN103795641A (en) | Optical network resource management method based on multi-dimensional frame mapping | |
| CN116132403A (en) | Route distribution method and device of computing power network, electronic equipment and storage medium | |
| CN113285832B (en) | NSGA-II-based power multi-mode network resource optimization allocation method | |
| CN112600753B (en) | Equipment node communication path selection method and device according to equipment access amount | |
| CN115021399A (en) | Topology identification method and device adaptive to park multi-energy power supply network | |
| CN113542115B (en) | SDN power communication network-based data path determination method, device and system | |
| Vallejos et al. | Fast blocking probability evaluation of end-to-end optical burst switching networks with non-uniform ON–OFF input traffic model | |
| CN116032363B (en) | Optical network routing and wavelength distribution method, system, equipment and medium | |
| CN113037425B (en) | Multi-target controller placement method based on evolution perception in network | |
| CN114650257B (en) | SDN network congestion control system and method based on RTT | |
| CN113422709B (en) | Data transmission method suitable for blockchain transaction verification | |
| CN114978947B (en) | Method, device and equipment for calculating simulation communication time of power system |
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 |