CN101695055A - Multi-broadcasting route optimization searching method based on improving the clonal niche algorithm - Google Patents
Multi-broadcasting route optimization searching method based on improving the clonal niche algorithm Download PDFInfo
- Publication number
- CN101695055A CN101695055A CN200910236368A CN200910236368A CN101695055A CN 101695055 A CN101695055 A CN 101695055A CN 200910236368 A CN200910236368 A CN 200910236368A CN 200910236368 A CN200910236368 A CN 200910236368A CN 101695055 A CN101695055 A CN 101695055A
- Authority
- CN
- China
- Prior art keywords
- antibody
- algorithm
- population
- group
- node
- 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.)
- Pending
Links
Images
Abstract
The invention discloses a multi-broadcasting route optimization searching method based on improving the clone niche algorithm in the technical field of communication, which comprises obtaining network information and generating an alternative route base, randomly generating a first antibody aggregation according to the first alternative route base, determining an antibody in a memory pool, calculating the affinity degree of the first antibody aggregation, forming final n groups of antibody clusters according to the affinity degree, carrying about the clone proliferation for n groups of the antibody clusters and variation then, selecting a Pareto solution in each group of the antibody clusters and then putting the Pareto solution in a second antibody aggregation, processing the second antibody aggregation, selecting out the Pareto solution of the second antibody aggregation and then putting the Pareto solution in the memory pool, carrying out the similarity suppression and gradient decision, determining whether the partial crowding distance of the antibody in the memory pool is smaller than or equal to a preset upper-limit threshold value, operating a partial crowding system if the partial crowed distance is larger than the threshold value, determining whether reaching iteration times, and completing the optimization searching if reaching the iteration times. The invention can optimize a plurality of QoS parameters in the optimizing process of multi-broadcasting route.
Description
Technical field
The invention belongs to communication technical field, relate in particular to a kind of based on the multicast route optimization method that improves clone's microhabitat algorithm.
Background technology
Along with the continuous development of wideband multimedia network, various broadband networks are used and are emerged in an endless stream.The legacy network that adopts unicast technique to make up can't satisfy the requirement of Wideband network at bandwidth and network service quality (QoS, Quality of Service) aspect.For addressing this problem the introducing multicasting technology, realize that the important step of multicasting technology is exactly the establishment in multicast transmission path.Different with the unicast transmission path, the transmission topology of multicast packet is a multicast tree, so realize that multicasting technology is equal to the structure multicast tree.Consider that present wideband multimedia network application requirement has QoS to guarantee, so how to make up the huge challenge that a multicast tree that satisfies qos requirement becomes the multicast research field.
By the multicast route technology that satisfies qos requirement is studied as can be known, when the constraints of Route Selection comprises two or more addition types tolerance, as delay, cost etc., when perhaps comprising the combination of addition type tolerance and multiplication type tolerance (as Loss Rate), then this Route Selection problem is a np complete problem.
Proposed multiple algorithm application at present in the multicast routing issue, the overall goal of these algorithms all is the minimum multicast tree of structure cost, utilizes weighting scheme that a plurality of optimization aim function linearities are combined as the single goal function, and then tries to achieve the optimal value of this target function.There are two defectives in these class methods: the result who at first is weighted sum is very responsive to weight vectors, and need the user certain understanding to be arranged to finding the solution problem, as the priority of parameter, the parameter situation etc. that influences to other parameters, this has limited the practicality of this class routing algorithm to a certain extent; Secondly, because the optimization result who obtains is single separating, the nonoptional leeway of user, and in practice, can what the user was concerned about be obtain one group of acceptable optimum noninferior solution, therefrom selects suitable practicality to separate according to form of service again.
Summary of the invention
The objective of the invention is to propose a kind of multicast route optimization method, when satisfying multi-constraint condition, can optimize a plurality of qos parameters based on improvement clone microhabitat algorithm.
Technical scheme of the present invention is that a kind of multicast route optimization method based on improvement clone microhabitat algorithm is characterized in that described method comprises the following steps:
Step 1: obtain network configuration information, generate each network paramter matrix;
Step 2: according to constraints updating network parameters matrix;
Step 3: use the alternative path storehouse that the depth-first algorithm generates network, calculate the affinity of every alternative path;
Step 4: produce the first antibody set at random according to the alternative path storehouse, the antibody number is set at N, calculates the affinity of antibody in the first antibody set;
Step 5: use reverse choice mechanism, according to the region of rejection in memory pond, first antibody set and the antibody in the memory pond are compared, remove the antibody that drops in the described region of rejection, and in the first antibody set, replenish antibody quantity to N, calculate the affinity of the antibody that replenishes; Again the antibody that replenishes is compared with the antibody in the memory pond, drop on region of rejection outside or additional antibody carries out next step when reaching set point number up to the antibody that replenishes;
Step 6: N antibody is sorted respectively according to every kind of affinity, and N antibody after will sorting then resolves into n organizes sub-antibody population, and the antibody quantity of every group of sub-antibody population becomes normal distribution with the antibody affinity; Then, merge the identical sub-antibody population of group number, form final n group antibody population;
Step 7:, described n group antibody population is carried out clonal expansion according to the propagation multiple of setting;
Step 8: according to the different antibodies group, set different variation probability, n group antibody population is made a variation;
Step 9: the Pareto of selecting in every group of antibody population is separated, and puts into the second antibody set;
Step 10: the similitude inhibition is carried out in set to second antibody, removes redundant antibody; Find the solution second antibody set Pareto then and separate, the result is put into the memory pond, carry out similitude and suppress, and carry out the gradient judgement, whether the antibody quantity that judgement memory Chi Zhongsuo comprises is less than or equal to the upper limit threshold of setting, if, execution in step 11; Otherwise, execution in step 12.
Step 11: inoperation local congestion mechanism;
Step 12: operation local congestion mechanism, choose the bigger antibody of local congestion distance in the memory pond, stay in the memory pond;
Step 13: judge whether to reach iterations,, otherwise forward step 4 to, carry out next iteration if then finish.
The described network configuration information of obtaining comprises the parameter value that obtains the cost between each node, time delay and Loss Rate in the network.
Described antibody comprises the combination that is clipped to any paths of each destination node from the start node branch, also comprises overall delay and total cost in described path.
Described overall delay is meant between each node of every paths to be to prolong sum.
Described total cost is meant the cost sum between each node of every paths.
Described additional antibody is chosen from the alternative path storehouse of corresponding destination node.
Describedly respectively N antibody is resolved into n according to affinity and organize sub-antibody population and adopt the microhabitat algorithm.
Described variation is meant that with path stochastic transformation from the start node to the destination node in the antibody be in the alternative path storehouse, another paths from identical start node to identical destination node.
The local congestion distance calculation formula of the antibody in the described memory pond is:
Wherein,
n
aNumber of individuals in the expression memory pond, x
ImAnd x
InWhen expression is only sorted to the individuality in the memory pond by i target function respectively, with immediate two individualities of x; If f
i(x)=min f
i(x
j), j ∈ 1,2 ..., n
a, j ≠ i makes f
i(x
In)=M
i, M
iBe set at an enough big number, if f
i(x)=max f
i(x
j), j ∈ 1,2 ..., n
a, j ≠ i then makes f
i(x
Im)=M
i, M
iBe set at an enough big number.
Effect of the present invention is, adopts method provided by the invention can obtain one group of final noninferior solution and offers the user, and the user just can select suitable separating according to the needs of reality use.So just overcome the single goal optimized Algorithm and can only try to achieve single deficiency of separating.
Description of drawings
Fig. 1 is the multicast route optimization method flow chart based on improvement clone microhabitat algorithm that the embodiment of the invention provides;
Fig. 2 is that the start node that the embodiment of the invention provides is 1, and destination node is 10 alternative path storehouse schematic diagram;
Fig. 3 is that the start node that the embodiment of the invention provides is 1, and destination node is 7 alternative path storehouse schematic diagram;
Fig. 4 is an embodiment of the invention first antibody set schematic diagram;
Fig. 5 is that the set of embodiment of the invention first antibody is by the ascending ordering schematic diagram of total cost;
Fig. 6 is that the set of embodiment of the invention first antibody is by the ascending ordering schematic diagram of overall delay;
Fig. 7 is the first sub-antibody population schematic diagram that the embodiment of the invention is divided according to total cost;
Fig. 8 is the second sub-antibody population schematic diagram that the embodiment of the invention is divided according to total cost;
Fig. 9 is the 3rd sub-antibody population schematic diagram that the embodiment of the invention is divided according to total cost;
Figure 10 is the first sub-antibody population schematic diagram that the embodiment of the invention is divided according to overall delay;
Figure 11 is the second sub-antibody population schematic diagram that the embodiment of the invention is divided according to overall delay;
Figure 12 is the 3rd sub-antibody population schematic diagram that the embodiment of the invention is divided according to overall delay;
Figure 13 is that the embodiment of the invention is respectively according to first group of antibody population schematic diagram after the first sub-antibody population merging of total cost and overall delay division;
Figure 14 is that the embodiment of the invention is respectively according to second group of antibody population schematic diagram after the second sub-antibody population merging of total cost and overall delay division;
Figure 15 is that the embodiment of the invention is respectively according to the 3rd group of antibody population schematic diagram after the 3rd sub-antibody population merging of total cost and overall delay division;
Figure 16 is first group of antibody population clonal expansion schematic diagram after the embodiment of the invention merges;
Figure 17 is second group of antibody population clonal expansion schematic diagram after the embodiment of the invention merges;
Figure 18 is the 3rd a group of antibody population clonal expansion schematic diagram after the embodiment of the invention merges;
Figure 19 is first group of antibody population variation schematic diagram behind the embodiment of the invention clonal expansion;
Figure 20 is second group of antibody population variation schematic diagram behind the embodiment of the invention clonal expansion;
Figure 21 is the 3rd a group of antibody population variation schematic diagram behind the embodiment of the invention clonal expansion;
Figure 22 is that the Pareto that first group of antibody population after the embodiment of the invention variation selected is separated schematic diagram;
Figure 23 is that the Pareto that second group of antibody population after the embodiment of the invention variation selected is separated schematic diagram;
Figure 24 is that the 3rd group of Pareto that antibody population is selected after the embodiment of the invention variation separated schematic diagram;
Figure 25 is the second antibody set schematic diagram of the embodiment of the invention;
Figure 26 is that the result schematic diagram after similitude suppresses is carried out in the second antibody set of the embodiment of the invention;
Figure 27 is that the second antibody of the embodiment of the invention is gathered the result schematic diagram after the Pareto of selecting is separated;
Figure 28 be after iteration of second antibody set of the embodiment of the invention in the memory pond antibody according to the ascending ordering schematic diagram of total cost;
Figure 29 be after iteration of second antibody set of the embodiment of the invention in the memory pond antibody according to the ascending ordering schematic diagram of overall delay.
Figure 30 is the result schematic diagram after the second antibody of the embodiment of the invention is gathered the antibody operation local congestion mechanism of remembering in the pond.
Embodiment
Below in conjunction with accompanying drawing, preferred embodiment is elaborated.Should be emphasized that following explanation only is exemplary, rather than in order to limit the scope of the invention and to use.
The present invention is mainly used in the actual communication network.For convenience of explanation, the present invention adopts the random network of emulation generation as embodiment.
Because the random network that adopts Waxman commonly used to propose carries out emulation, obtain network network of " evenly " not very often, the node number of degrees that promptly have can be very big, and other node is very little, even have the situation of isolated node.So at this situation, the network generating algorithm that adopts Salama and Reeves on the basis of Waxman network, to propose, generate random network by emulation, because characteristics such as it is general, accurate that it possesses that artificial network should possess, the nearly all researcher who for this reason is studied multicast routing algorithm adopts.The random network that emulation generates should satisfy: the average nodal number of degrees of network are 4, change the number of network node from 10 to 100, and the destination node number is made as 20% of the total node number of network, chain-circuit time delay, cost, link bandwidth are respectively [0,100] ms, [2,90], [0,10] are gone up evenly and are distributed.The time delay upper limit of link is between [10,90] ms, and the link minimum bandwidth constrains in [2,8] and goes up evenly distribution.
In the present embodiment, be that 10 nodes (nodal scheme is respectively 1-10) are that example describes with network size.
Fig. 1 is the multicast route optimization method flow chart based on improvement clone microhabitat algorithm that the embodiment of the invention provides.Among Fig. 1, concrete implementation step of the present invention is:
Step 101: obtain network configuration information, generate each network paramter matrix.
Obtain the information such as the number of degrees, destination node, chain-circuit time delay, cost, link bandwidth and Loss Rate that network configuration information comprises each node that obtains number of network node, network.
Multicast Route Selection problem can be described as a undirected weighted graph G=(V, E), V={V wherein
iBe the network node collection, and can be expressed as switch, router, also can be subnet; E={e}, e={＜u, v〉| u, v ∈ V} is a network link.(V, E), wherein V is the network node collection to a undirected weighted graph G=, E={e
IjBe network link, wherein e
IjBe expressed as node i to a link between the node j.For
Four weights functions are all arranged: chain-circuit time delay delay (e): E → R
+, cost cost (e): E → R
+, Loss Rate loss (e): E → R
+With link bandwidth bandwidth (e): E → R
+In the present invention, the weights function claims target function again, and in an embodiment of the present invention, setting bandwidth, time delay and cost is constraints, generates three parameter matrixs after obtaining information.
Step 102: according to constraints updating network parameters matrix.
In the present embodiment, require the time delay upper limit between the adjacent node between [10,90] ms, minimum bandwidth is constrained to [2,8], if not in this scope, then this value in the corresponding matrix is changed to ∞.Bandwidth of operation, cost, three matrixes of time delay are ∞ if having a value in any one matrix subsequently, then value corresponding in cost, time delay, the bandwidth matrices are changed to ∞.As long as the point-to-point transmission bandwidth meets the demands, there is not customer requirements, in the present embodiment so bandwidth is not operated as target function in the back.
Step 103: use the depth-first algorithm and generate the alternative path storehouse that reaches the network destination node, calculate the affinity of every alternative path.
The alternative path storehouse of optimizing network destination node is meant that to the fixedly set in all paths of destination node, the quantity in alternative path storehouse is identical with the quantity of destination node from start node.The specific implementation process of depth-first search is as follows: at first, from start node, select a node associated therewith randomly, 2 are linked to each other; Continue to select randomly the next node related from the node of selecting then, when connecting, will judge whether the loop with it, if any the loop then by recalling the selection of carrying out node again, until all paths that search this destination node of arrival.Generate the alternative path storehouse of other destination node correspondence according to same method.
In the present embodiment, getting target function is cost and time delay, and the value of target function is the affinity of alternative path.Target function can be the additivity parameter herein, also can be the property taken advantage of parameter.The additivity parameter is done addition to time delay and cost between each node exactly, the property taken advantage of parameter is done multiplication to time delay and cost between each node exactly, cost that adopts in the present embodiment and time delay are the additivity parameter, according to parameter character, total cost of antibody and overall delay are respectively in the alternative path, the summation of each internodal cost and time delay.Each internodal cost and time delay obtain by step 101.
In the present embodiment, with 10 nodes (nodal scheme is respectively 1-10), the N=4 of setting is that example describes.In the network of Sheng Chenging, requiring start node is node 1 at random, and destination node is 7 and 10.
The alternative path of destination node 10 is:
Alternative path A1:1,3,5,10;
Alternative path A2:1,6,3,10;
Above-mentioned alternative path A1, A2 have formed the alternative path storehouse of destination node 10, see Fig. 2.
The alternative path of destination node 7 is:
Alternative path B1:1,2,7;
Alternative path B2:1,4,6,7;
Alternative path B3:1,3,6,7.
Above-mentioned alternative path B1, B2, B3 have formed the alternative path storehouse of destination node 7, see Fig. 3.
Step 104: produce the first antibody set at random according to the alternative path storehouse, the antibody number is set at N, calculates the affinity of antibody in the first antibody set.
Producing the first antibody set at random specifically is, from the alternative path storehouse of destination node, chooses a paths at random separately and is combined into an antibody, sets N=4 in the present embodiment, and therefore 4 antibody are formed the first antibody set.
In the present embodiment, specifically be from the alternative path storehouse of destination node 10, to choose a paths at random, simultaneously choose a paths at random from the alternative path storehouse of destination node 7, two paths are combined into an antibody, and wherein the antibody element of delegated path is represented with path number.Because 10 alternative path has 2 from node 1 to destination node, there are 3 to the alternative path of destination node 7, so the alternative path combination that was clipped to destination node 10 and 7 in 1 minute from node has 6 kinds.
The affinity of antibody refers to each self-corresponding affinity sum of contained two paths of antibody.After 4 antibody were chosen, the cost addition of two paths that antibody is contained was as the total cost of antibody; The time delay addition of two paths is as the antibody overall delay.Fig. 4 is an embodiment of the invention first antibody set schematic diagram.Among Fig. 4, each antibody comprises two paths, and one is from start node 1 to destination node 10; Another is from start node 1 to destination node 7, and each antibody back is with total cost and overall delay are arranged.
Step 105: use reverse choice mechanism,, antibody in first antibody set and the memory pond is compared, remove the antibody that drops in the described region of rejection according to the region of rejection in memory pond, and additional antibody; And in the first antibody set, replenish antibody quantity to 4.
The antibody that replenishes needs 104 to calculate affinity set by step, the antibody that replenishes is compared with the antibody in the memory pond again, drops on region of rejection outside or additionally carries out next step when reaching set point number up to the antibody that replenishes.
Oppositely the purpose of choice mechanism is to remove redundant antibody.If the absolute value that is specially in overall delay that antibody is arranged in the first antibody set and total cost and the memory pond difference of the overall delay of antibody and total cost arbitrarily is all less than a set point (being set at 0.002 in the present embodiment), just say that this antibody has dropped within the region of rejection in memory pond, and this set point (0.002) is just remembered the region of rejection in pond.Because the Chi Weikong of iteration memory for the first time, so the first antibody set does not change after moving reverse choice mechanism.
The set point number here be consider if each antibody that adds all in the region of rejection in memory pond, tediously long or algorithm of the time that may cause is not restrained.Therefore, even if also have the antibody that replenishes in region of rejection after arriving set point number, also carry out next step.
Step 106: N antibody is sorted respectively according to every kind of affinity, and N antibody after will sorting then resolves into n organizes sub-antibody population, and the antibody quantity of every group of sub-antibody population becomes normal distribution with the antibody affinity; Then, merge the identical sub-antibody population of group number, form final n group antibody population.
Fig. 5 is that the set of embodiment of the invention first antibody is by the ascending ordering schematic diagram of total cost.Among Fig. 5, each antibody is according to the ascending rank order of total cost.Fig. 6 is that the set of embodiment of the invention first antibody is by the ascending ordering schematic diagram of overall delay.Among Fig. 6, each antibody is according to the ascending rank order of overall delay.
After total cost and overall delay ordering, adopt the microhabitat algorithm, respectively N antibody is resolved into n according to affinity and organize sub-antibody population, and guarantee that the quantity of antibody in every group of sub-antibody population becomes normal distribution with the antibody affinity.The thinking of microhabitat algorithm is, the total cost and the overall delay of the antibody of first group of sub-antibody population of division are all high, and the variation probability of setting behind the clonal expansion of back is also high; The total cost and the overall delay of last group antibody population of dividing are all low, and the variation probability is also low behind the clonal expansion, and middle groups clone back variation probability is medium.This " dividing the variation of gene section " operation, only in having the sub-antibody population of similar gene expression characteristics, carry out, and the variation of each sub-antibody population all is complementary with the feature of this antibody population, be not to make a variation arbitrarily, thereby we can say that the variation of antibody each time all is to carry out towards specific optimization direction.Like this, different variation probability is arranged, given full play to the effect of variation mechanism in the immune system, make algorithm have stronger multiple target optimizing ability at the different genes section.
In the present embodiment, set n=3, promptly be divided into 3 groups, afterwards grouping as follows: the antibody with Fig. 5 is divided into 3 groups of sub-antibody populations according to total cost earlier, and every group of sub-antibody population is made up of two antibody.Fig. 7 is the first sub-antibody population schematic diagram that the embodiment of the invention is divided according to total cost, and among Fig. 7, the first sub-antibody population is two antibody of total cost minimum, comprises antibody 3 and antibody 4.Fig. 8 is the second sub-antibody population schematic diagram that the embodiment of the invention is divided according to total cost, and among Fig. 8, the second sub-antibody population is total cost two antibody placed in the middle, comprises antibody 4 and antibody 1.Fig. 9 is the 3rd sub-antibody population schematic diagram that the embodiment of the invention is divided according to total cost, and among Fig. 9, the 3rd sub-antibody population is two antibody of total cost maximum, comprises antibody 1 and antibody 2.
Antibody with Fig. 6 is divided into 3 groups of antibody populations according to overall delay again, and each antibody population is made up of two antibody.Figure 10 is the first sub-antibody population schematic diagram that the embodiment of the invention is divided according to overall delay, and among Figure 10, the first sub-antibody population is two antibody of overall delay minimum, comprises antibody 1 and antibody 4.Figure 11 is the second sub-antibody population schematic diagram that the embodiment of the invention is divided according to overall delay, and among Figure 11, the second sub-antibody population is overall delay two antibody placed in the middle, comprises antibody 4 and antibody 3.Figure 12 is the 3rd sub-antibody population schematic diagram that the embodiment of the invention is divided according to overall delay, and among Figure 12, the 3rd sub-antibody population is two antibody of overall delay maximum, comprises antibody 3 and antibody 2.
Merge the identical antibody population of group number and be meant, will merge according to the first sub-antibody population, the second sub-antibody population and the 3rd sub-antibody population that total cost and overall delay are divided respectively.Figure 13 is that the embodiment of the invention is respectively according to first group of antibody population schematic diagram after the first sub-antibody population merging of total cost and overall delay division.Among Figure 13, preceding two antibody that antibody is total cost minimum, latter two antibody is the antibody of overall delay minimum.Figure 14 is that the embodiment of the invention is respectively according to second group of antibody population schematic diagram after the second sub-antibody population merging of total cost and overall delay division.Among Figure 14, preceding two antibody are total cost antibody placed in the middle, and latter two antibody is overall delay antibody placed in the middle.Figure 15 is that the embodiment of the invention is respectively according to the 3rd group of antibody population schematic diagram after the 3rd sub-antibody population merging of total cost and overall delay division.Among Figure 15, preceding two antibody that antibody is total cost maximum, latter two antibody is the antibody of overall delay maximum.Through merging, form final 3 groups of new antibody populations.
Step 107: the propagation multiple according to setting, carry out clonal expansion to above-mentioned 3 groups of antibody populations.
The propagation multiple is set according to actual needs, and in the present embodiment, setting the increment multiple is 3.Above-mentioned 3 groups of antibody populations are carried out clonal expansion, be about to above-mentioned 3 groups of antibody populations and duplicate 2 parts respectively.
Figure 16 is first group of antibody population clonal expansion schematic diagram after the embodiment of the invention merges.Among Figure 16, the antibody in first group of antibody population after merging is duplicated 2 times.Figure 17 is second group of antibody population clonal expansion schematic diagram after the embodiment of the invention merges.Among Figure 17, the antibody in second group of antibody population after merging is duplicated 2 times.Figure 18 is the 3rd a group of antibody population clonal expansion schematic diagram after the embodiment of the invention merges.Among Figure 18, the antibody in the 3rd group of antibody population after merging is duplicated 2 times.
Step 108: according to the different antibodies group, set different variation probability, n group antibody population is made a variation.
At first, according to previous step rapid in, different variation probability is set in 3 groupings behind the clonal expansion.First group (antibody population shown in Figure 16) is low variation probability district, and setting between region of variability is [0.1 0.2]; Second group (antibody population shown in Figure 17) is middle variation probability district, and setting between region of variability is [0.5 0.6]; The 3rd group (antibody population shown in Figure 180) is high variation probability district, and setting between region of variability is [0.8 0.9].[x y] wherein is for independently between region of variability, and the variation situation in the contained path of antibody is according to determining between the region of variability in Different Variation probability district.
Concrete mutation process, the district is the first behavior example of first group (antibody population shown in Figure 16) with low variation probability, between the region of variability in known low variation probability district is [0.1 0.2], at an antibody, choose the random number r between 0-1, when r 〉=0.5, the alternative path of destination node 10 does not need variation; And when r＜0.5, the alternative path of destination node 10 needs variation.Variation is that alternative path is transformed in the alternative path storehouse, and any one is starting point with node 1, and node 10 is the path of destination node.When r 〉=0.2, the alternative path of destination node 7 does not need variation; And when r＜0.2, the alternative path of destination node 7 needs variation, and variation is that alternative path is transformed in the alternative path storehouse, and any one is starting point with node 1, and node 7 is the path of destination node.After the variation, recomputate total cost and overall delay.After three groups of antibody populations make a variation in the manner described above, enter next step.
Figure 19-Figure 21 is respectively the antibody population variation schematic diagram behind the embodiment of the invention clonal expansion.Wherein, the part of boldface type represents that antibody makes a variation.Figure 19 is first group of antibody population variation schematic diagram behind the embodiment of the invention clonal expansion, and among Figure 19, the path of the destination node 7 of the 2nd row antibody makes a variation; The destination node 10 of the 6th row antibody and 7 path make a variation; Figure 20 is second group of antibody population variation schematic diagram behind the embodiment of the invention clonal expansion.Among Figure 20, most paths make a variation.Figure 21 is the 3rd a group of antibody population variation schematic diagram behind the embodiment of the invention clonal expansion. among Figure 21, almost all the path makes a variation.
Step 109: the Pareto of selecting in every group of antibody population is separated, and forms the second antibody set.
With more total in twos cost of the antibody of every group of antibody population and overall delay, obtain Pareto and separate and merge and put into second antibody set.
Pareto optimality is meant a kind of state of resource allocation, and Pareto is separated, and to ask method be more total in twos cost of antibody and overall delay in will gathering:
If total cost of a certain antibody and overall delay are all less than another antibody, then this antibody is that Pareto is separated, and another antibody is not;
If total cost of a certain antibody and overall delay one of them greater than another antibody, one less than another antibody, then two antibody are Pareto and separate;
If total cost of a certain antibody and overall delay all greater than or be equal to another antibody, then this antibody is that non-Pareto is separated, another antibody is that Pareto is separated;
If one of them equals another antibody total cost of a certain antibody and overall delay, one then this antibody is that non-Pareto is separated greater than another antibody, and another antibody is that Pareto is separated;
If one of them equals another antibody total cost of a certain antibody and overall delay, one then this antibody is that Pareto is separated less than another antibody, and another antibody is not.
Specific operation process in conjunction with the embodiments is, in Figure 19, antibody 1 and antibody 2 relatively, total cost of antibody 1 is less, the overall delay of antibody 2 is less, keeps antibody 1,2; Antibody 1 compares with antibody 3, and the two is identical, leaves out antibody 3; By that analogy.
Figure 22 is that first group of antibody population Pareto after the embodiment of the invention variation separated schematic diagram, and among Figure 22, Pareto is separated and comprised 3 antibody.Figure 23 is that second group of antibody population Pareto after the embodiment of the invention variation separated schematic diagram, and among Figure 23, Pareto is separated and comprised 4 antibody.Figure 24 is that the 3rd group of antibody population Pareto after the embodiment of the invention variation separated schematic diagram, and among Figure 24, Pareto is separated and comprised 4 antibody.
Pareto in each group separated put into a new antibody set, form the second antibody set.Figure 25 is the second antibody set schematic diagram of the embodiment of the invention.
Step 110: the similitude inhibition is carried out in set to second antibody, removes redundant antibody; Find the solution the Pareto of second antibody set then and separate, the result is put into the memory pond, carry out similitude and suppress, and carry out the gradient judgement, whether the antibody quantity that judgement memory Chi Zhongsuo comprises is less than or equal to the upper limit threshold of setting, if, execution in step 111; Inoperation local congestion mechanism.Otherwise, execution in step 112, operation local congestion mechanism.
It specifically is that antibody compares in twos in the second antibody set, if antibody is arranged in the region of rejection of another one antibody, then leaves out this antibody that second antibody set is carried out that similitude suppresses.Here, the definition of region of rejection is identical with the region of rejection definition in memory pond, if promptly the total cost of antibody X and overall delay respectively with the absolute value of the difference of the total cost of antibody Y and overall delay less than set point 0.002, think that then antibody X is in the region of rejection of antibody Y.Figure 26 is that the result schematic diagram after similitude suppresses is carried out in the second antibody set of the embodiment of the invention.Among Figure 26, second antibody set similitude has suppressed in fact to delete the identical antibody in the antibody set.
Afterwards, select second antibody set Pareto and separate, put into the memory pond.Figure 27 is that the second antibody of the embodiment of the invention is gathered the result schematic diagram after the Pareto of selecting is separated.
This Pareto separated put into memory Chi Zhonghou, the antibody of memory in the pond is carried out similitude suppress.Under the situation of iteration, the memory pond is initially sky for the first time, and it is constant that similitude suppresses the back result; And repeatedly remember the antibody that Chi Zhonghui has last iteration after the iteration, carry out similitude and suppress to remove redundant antibody.
The gradient judgement promptly selects the Pareto in the memory pond to separate operation.Under the situation of iteration, the memory pond is initially sky for the first time, and the back result is constant in the gradient judgement; And repeatedly remember the antibody that Chi Zhonghui has last iteration after the iteration, carry out having guaranteed that the antibody in the memory pond is that Pareto is separated after the gradient judgement.
The number of separating when Pareto hour can not influence operational performance substantially.Yet when the number of resulting Pareto was a lot, the operand of algorithm will become very big when the back iteration was carried out associative operation to the memory pond.And may there be the too big antibody of total cost or overall delay in the memory pond.For this reason, the antibody quantity in the memory pond is set a upper limit threshold, limited the scale in memory pond.
In the present embodiment, the capping threshold value is 3.But this moment, the antibody quantity in the memory pond is 4, operates local congestion mechanism, execution in step 112 this moment.
Step 112: operation local congestion mechanism, choose the bigger antibody of local congestion distance and stay in the memory pond.
The local congestion distance calculation formula of antibody is in the memory pond:
In the formula
n
aAntibody number in the expression memory pond, x
ImAnd x
InWhen representing respectively the antibody of memory in the pond to be sorted, with immediate two individualities of x according to the target function value sum; If f
i(x)=min f
i(x
j), j ∈ 1,2 ..., n
a, j ≠ i makes f
i(x
In)=M
i, M
iBe set at an enough big number, if f
i(x)=max f
i(x
j), j ∈ 1,2 ..., n
a, j ≠ i then makes f
i(x
Im)=M
i, M
iBe set at an enough big number.
In the present embodiment, elder generation sorts according to the total cost and the overall delay of antibody respectively.Set M1=50, M2=40.
Figure 28 be after iteration of second antibody set of the embodiment of the invention in the memory pond antibody according to the ascending ordering schematic diagram of total cost.
(|2.467-50|+|2.467-3.534|+|5.282-40|+|5.282-5.18|)/4＝20.855
(|3.534-2.467|+|3.534-4.872|+|5.18-5.282|+|5.18-3.669|)/4＝1.0045
(|4.872-3.534|+|4.872-5.939|+|3.669-5.18|+|3.669-3.567|)/4＝1.0045
(|5.939-4.872|+|5.939-50|+|3.567-3.669|+|3.567-40|)/4＝20.41575
Figure 29 be after iteration of second antibody set of the embodiment of the invention in the memory pond antibody according to the ascending ordering schematic diagram of overall delay.
(|5.939-50|+|5.939-4.872|+|3.567-40|+|3.567-3.669|)/4+20.41575＝40.7805
(|4.872-5.939|+|4.872-3.534|+|3.669-3.567|+|3.669-5.18|)/4+1.0045＝2.009
(|3.534-4.872|+|3.534-2.467|+|5.18-3.669|+|5.18-5.282|)/4+1.0045＝2.009
(|2.467-3.534|+|2.467-50|+|5.282-5.18|+|5.282-40|)/4+20.855＝41.1765
As seen the local congestion of original antibody 4 and original antibody 2 keeps apart from maximum; Because setting the upper limit threshold of memory antibody pool quantity is 3, the local congestion distance of original antibody 1 and original antibody 3 equates, then fetch bit is put forward original antibody 1 and kept, leave out original antibody 3, so far, residue original antibody 1,2,4 enters next iteration in the memory pond, sees Figure 30.Figure 30 is the result schematic diagram after the second antibody of the embodiment of the invention is gathered the antibody operation local congestion mechanism of remembering in the pond.
Step 113: judge whether to reach iterations, if then finish; Otherwise forward step 104 to, carry out next iteration.
If reach iterations, then remember the optimum noninferior solution that the antibody in the pond is obtained for the present invention, the client can select as required.If do not reach iterations, repeating step 104-step 113 then.Iterations is according to network complexity and the precision set of finding the solution.
The above; only for the preferable embodiment of the present invention, but protection scope of the present invention is not limited thereto, and anyly is familiar with those skilled in the art in the technical scope that the present invention discloses; the variation that can expect easily or replacement all should be encompassed within protection scope of the present invention.Therefore, protection scope of the present invention should be as the criterion with the protection range of claim.
Claims (9)
1. the multicast route optimization method based on improvement clone microhabitat algorithm is characterized in that described method comprises the following steps:
Step 1: obtain network configuration information, generate each network paramter matrix;
Step 2: according to constraints updating network parameters matrix;
Step 3: use the alternative path storehouse that the depth-first algorithm generates network, calculate the affinity of every alternative path;
Step 4: produce the first antibody set at random according to the alternative path storehouse, the antibody number is set at N, calculates the affinity of antibody in the first antibody set;
Step 5: use reverse choice mechanism, according to the region of rejection in memory pond, first antibody set and the antibody in the memory pond are compared, remove the antibody that drops in the described region of rejection, and in the first antibody set, replenish antibody quantity to N, calculate the affinity of the antibody that replenishes; Again the antibody that replenishes is compared with the antibody in the memory pond, drop on region of rejection outside or additional antibody carries out next step when reaching set point number up to the antibody that replenishes;
Step 6: N antibody is sorted respectively according to every kind of affinity, and N antibody after will sorting then resolves into n organizes sub-antibody population, and the antibody quantity of every group of sub-antibody population becomes normal distribution with the antibody affinity; Then, merge the identical sub-antibody population of group number, form final n group antibody population;
Step 7:, described n group antibody population is carried out clonal expansion according to the propagation multiple of setting;
Step 8: according to the different antibodies group, set different variation probability, n group antibody population is made a variation;
Step 9: the Pareto of selecting in every group of antibody population is separated, and puts into the second antibody set;
Step 10: the similitude inhibition is carried out in set to second antibody, removes redundant antibody; Find the solution second antibody set Pareto then and separate, the result is put into the memory pond, carry out similitude and suppress, and carry out the gradient judgement, whether the antibody quantity that judgement memory Chi Zhongsuo comprises is less than or equal to the upper limit threshold of setting, if, execution in step 11; Otherwise, execution in step 12.
Step 11: inoperation local congestion mechanism;
Step 12: operation local congestion mechanism, choose the bigger antibody of local congestion distance in the memory pond, stay in the memory pond;
Step 13: judge whether to reach iterations,, otherwise forward step 4 to, carry out next iteration if then finish.
2. according to claim 1 a kind of based on the multicast route optimization method that improves clone's microhabitat algorithm, it is characterized in that described random network structural information comprises the parameter value that obtains the cost between each node, time delay and Loss Rate in the network.
3. according to claim 1 a kind of based on the multicast route optimization method that improves clone's microhabitat algorithm, it is characterized in that described antibody comprises the combination that is clipped to any paths of each destination node from the start node branch, also comprise overall delay and total cost in described path.
4. according to claim 3 a kind of based on the multicast route optimization method that improves clone's microhabitat algorithm, it is characterized in that described overall delay is meant the time delay sum between each node of every paths.
5. according to claim 3 a kind of based on the multicast route optimization method that improves clone's microhabitat algorithm, it is characterized in that described total cost is meant the cost sum between each node of every paths.
6. according to claim 1 a kind of based on the multicast route optimization method that improves clone's microhabitat algorithm, it is characterized in that described additional antibody chooses from the alternative path storehouse of corresponding destination node.
7. according to claim 1 a kind of based on the multicast route optimization method that improves clone's microhabitat algorithm, it is characterized in that describedly respectively N antibody being resolved into n according to affinity and organizing sub-antibody population employing microhabitat algorithm.
8. according to claim 1 a kind of based on the multicast route optimization method that improves clone's microhabitat algorithm, it is characterized in that described variation is meant that with path stochastic transformation from the start node to the destination node in the antibody be in the alternative path storehouse, another paths from identical start node to identical destination node.
9. according to claim 1 a kind of based on the multicast route optimization method that improves clone's microhabitat algorithm, it is characterized in that the computing formula of the local congestion distance of the antibody in the described memory pond is:
Wherein,
n
aNumber of individuals in the expression memory pond, x
ImAnd x
InWhen expression is only sorted to the individuality in the memory pond by i target function respectively, with immediate two individualities of x; If f
i(x)=minf
i(x
j), j ∈ 1,2 ..., n
a, j ≠ i makes f
i(x
In)=M
i, M
iBe set at an enough big number, if f
i(x)=maxf
i(x
j), j ∈ 1,2 ..., n
a, j ≠ i then makes f
i(x
Im)=M
i, M
iBe set at an enough big number.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN200910236368A CN101695055A (en) | 2009-10-20 | 2009-10-20 | Multi-broadcasting route optimization searching method based on improving the clonal niche algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN200910236368A CN101695055A (en) | 2009-10-20 | 2009-10-20 | Multi-broadcasting route optimization searching method based on improving the clonal niche algorithm |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN101695055A true CN101695055A (en) | 2010-04-14 |
Family
ID=42093997
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN200910236368A Pending CN101695055A (en) | 2009-10-20 | 2009-10-20 | Multi-broadcasting route optimization searching method based on improving the clonal niche algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101695055A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104539535A (en) * | 2015-01-28 | 2015-04-22 | 浪潮电子信息产业股份有限公司 | Data transmission path determination method and data transmission path determination device |
-
2009
- 2009-10-20 CN CN200910236368A patent/CN101695055A/en active Pending
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104539535A (en) * | 2015-01-28 | 2015-04-22 | 浪潮电子信息产业股份有限公司 | Data transmission path determination method and data transmission path determination device |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108521375B (en) | SDN-based network multi-service flow QoS transmission and scheduling method | |
| CN105430707B (en) | A kind of wireless sense network multiple-objection optimization method for routing based on genetic algorithm | |
| US20120317264A1 (en) | Assigning Telecommunications Nodes to Community of Interest Clusters | |
| CN103179052A (en) | Virtual resource allocation method and system based on proximity centrality | |
| CN101616074B (en) | Multicast routing optimization method based on quantum evolution | |
| CN114143264B (en) | Flow scheduling method based on reinforcement learning under SRv network | |
| CN110417652B (en) | Software defined network routing method based on segmented routing strategy | |
| CN106685745B (en) | A kind of constructing network topology method and device | |
| CN113225370B (en) | Block chain multi-objective optimization method based on Internet of things | |
| CN102750286A (en) | Novel decision tree classifier method for processing missing data | |
| CN104009907A (en) | All-to-all message exchange in parallel computing systems | |
| CN115460130A (en) | Multi-path joint scheduling method in time-sensitive network | |
| CN115242295B (en) | Satellite network SDN multi-controller deployment method and system | |
| CN105704025B (en) | Routing optimization method based on Chaos Search and Artificial Immune Algorithm | |
| CN101013955A (en) | Fast simulated annealing for traffic matrix estimation | |
| CN107483079B (en) | Double-population genetic ant colony routing method for low-voltage power line carrier communication | |
| CN103595652B (en) | The stage division of QoS efficiency in a kind of powerline network | |
| Abdel-Kader | An improved discrete PSO with GA operators for efficient QoS-multicast routing | |
| Guo et al. | A novel cluster-head selection algorithm based on hybrid genetic optimization for wireless sensor networks | |
| CN104125146B (en) | A kind of method for processing business and device | |
| CN107135155B (en) | A kind of opportunistic network routing method based on node social relationships | |
| CN101695055A (en) | Multi-broadcasting route optimization searching method based on improving the clonal niche algorithm | |
| CN102768735A (en) | Network community partitioning method based on immune clone multi-objective optimization | |
| CN108111991B (en) | D2D network building method based on scalable video streaming user experience quality | |
| CN101741749A (en) | Method for optimizing multi-object multicast routing based on immune clone |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C02 | Deemed withdrawal of patent application after publication (patent law 2001) | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Open date: 20100414 |