CN104202188A - Method for carrying out AFDX network path optimization by genetic algorithm - Google Patents

Method for carrying out AFDX network path optimization by genetic algorithm Download PDF

Info

Publication number
CN104202188A
CN104202188A CN201410441205.3A CN201410441205A CN104202188A CN 104202188 A CN104202188 A CN 104202188A CN 201410441205 A CN201410441205 A CN 201410441205A CN 104202188 A CN104202188 A CN 104202188A
Authority
CN
China
Prior art keywords
switch
path
chromosome
population
virtual link
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
Application number
CN201410441205.3A
Other languages
Chinese (zh)
Other versions
CN104202188B (en
Inventor
何锋
代真
张宇静
熊华钢
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Beihang University
Original Assignee
Beihang University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Beihang University filed Critical Beihang University
Priority to CN201410441205.3A priority Critical patent/CN104202188B/en
Publication of CN104202188A publication Critical patent/CN104202188A/en
Application granted granted Critical
Publication of CN104202188B publication Critical patent/CN104202188B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Abstract

The invention discloses a method for carrying out AFDX network path optimization by a genetic algorithm. The method is virtual link path optimization carried out on an AFDX network configured with a VL path. The method disclosed by the invention comprises the following steps of: establishing a connection matrix and a virtual link path group of a switch at first; and then carrying out chromosome coding on a virtual link path; finally carrying out genetic operations of crossing, variation and selection on a chromosome group, obtaining the optimal chromosome in case of meeting an end condition, and extracting out a valid virtual link path, wherein the path is used as the optimized virtual link path. According to the method disclosed by the invention, path optimization for the virtual link VL in the configured AFDX network is solved, and real-time path optimization for the virtual link VL in the configured AFDX network is carried out by applying the genetic algorithm, thus improving the message transmission real-time performance of the AFDX network.

Description

A kind of method that adopts genetic algorithm to carry out the optimization of AFDX network path
Technical field
The invention belongs to avionics system communication network optimize design field, more particularly, refer to a kind of method that adopts genetic algorithm to carry out the optimization of AFDX network path.
Background technology
Avionic full-duplex switched-type Ethernet (Avionics Full Duplex Switched Ethernet, AFDX) is to transform through applicability on the basis of industrial standard Ethernet, and is applicable to the interconnected network technology of avionics system.AFDX interconnection technique has obtained successful Application in passenger vehicle A380 and Boeing 787 aloft.
AFDX network carries out message data flow transmission by virtual link (Virtual Link, VL), and AFDX network provides definite bandwidth and definite route for VL.The path effects of VL the transmission delay of virtual link bearer messages, also affects the flow equalization of AFDX network.
The 39th the 6th phase of volume of " BJ University of Aeronautics & Astronautics's journal " June in 2013 discloses " time is triggered design and the real time analysis of AFDX network ", author Liu Cheng.Fig. 8 of the document discloses a comparatively traditional AFDX network topology structure (being cited as shown in Fig. 1), have 8 switches and formed network backbone, on each switch, connect 8 end systems, one has 64 end systems, and link bandwidth is configured to 10Mbit/s.
Summary of the invention
In order to reduce in avionic full-duplex switched-type Ethernet AFDX, the transmission delay of virtual link bearer messages, the present invention proposes a kind of method that adopts genetic algorithm to carry out the optimization of AFDX network path.The method has solved the path optimization of virtual link VL in the AFDX network having configured, be that application genetic algorithm is carried out real-time route optimization to the VL path of AFDX network, thereby improved the transmission of messages real-time (being that the propagation delay time is short) of AFDX network.
The present invention is a kind of method that adopts genetic algorithm to carry out the optimization of AFDX network path, and described AFDX network refers to the AFDX network that disposes VL path; It is characterized in that including the following step:
Step 1: build switch connection matrix;
Carry out the connection initialization between all switches to disposing the AFDX network in VL path, obtain n × n switch connection matrix n is the number of the switch in AFDX network;
Step 2: build virtual link path population;
Virtual link path population scale U={u is set 1, u 2..., u x; Switch connection matrix SC to step 1 chooses and meets U={u according to path exclusion condition 1, u 2..., u xsource-object-switch path R - u x = ( S a A → u x S b B ) ;
Step 3: build chromosome;
Source-object-switch path that step 2 is obtained in the identification number of switch as gene, the encoded chromosome population CH={h that obtains 1, h 2..., h x;
Step 4: calculate chromosomal fitness function value;
Described fitness function Fitness=AVL max+ time postpone, wherein, maximum link load factor AVL max = max n Σ VL ∈ C nn r max t BAG 100 × 10 6 × 100 % , Message transmission delay time rate
Step 5: genetic manipulation;
Current chromosome population is designated as CH i, previous generation chromosome population is designated as CH i-1, chromosome population of future generation is designated as CH i+1;
Step 501, chromosomal the first cross modal:
The mode that uses single-point to intersect, crosspoint is selected in public branch exchange place, if there are multiple public branch exchanges, select at random one as crosspoint, then the part after exchange pairing crosspoint, two new chromosomes that generate, and described new chromosome is added in the population of virtual link of future generation path; If duplicate gene in new chromosome after intersecting, delete the switch of any one repetition, and the new chromosome of having deleted after repeated exchanged machine is added in the population of virtual link of future generation path.
Step 502, chromosomal the second cross modal:
The mode that uses single-point to intersect, if there is not public branch exchange, in the middle of selecting, chromosome intersects, the then part after exchange pairing crosspoint, two new chromosomes that generate, and described new chromosome is added in the population of virtual link of future generation path; If duplicate gene in new chromosome after intersecting, delete the switch of any one repetition, and add in the population of virtual link of future generation path and add in the population of virtual link of future generation path having deleted new chromosome after repeated exchanged machine.
Step 503, chromosomal the first variant form:
The mode that uses single-point variation, change point is selected in except source switch with object switch any switch place, select with described switch exist link-attached another switch of physics as variation result, obtain new chromosome, and described new chromosome added in the population of virtual link of future generation path; If duplicate gene in new chromosome after variation, delete the switch of any one repetition, and the new chromosome of having deleted after repeated exchanged machine is added in the population of virtual link of future generation path.
Step 504, chromosomal the second variant form:
The mode that uses single-point variation, change point is selected in except source switch with object switch any switch place, directly described switch is deleted, obtain new chromosome, and described new chromosome added in the population of virtual link of future generation path.
Step 505, chromosomal selection mode:
Chromosomal selection mode is by the chromosome population CH when former generation i={ h 1, h 2..., h xand by the new chromosome population of intersect-variation generation mix, obtain mixed population then pass through Fitness=AVL max+ time postponedescribed in calculating in each chromosomal fitness function value and choose x minimum chromosome as population CH of future generation i+1.
Step 6: judge whether to meet hereditary end condition;
Judge the fitness function value Fh of described minimum minimumafter continuous 3 generation genetic manipulations, described Fh minimumwhile variation, stop genetic manipulation; If do not meet, to CH={h 1, h 2..., h xin every item chromosome carry out genetic manipulation.
The advantage that the present invention adopts genetic algorithm to carry out the optimization of AFDX network path is:
1. genetic algorithm is incorporated in the AFDX network that disposes VL path, is more conducive to find in real time the optimal path of virtual link.Can improve the transmission of messages real-time (being that the propagation delay time is short) of AFDX network.
2. apply chromosome in genetic algorithm carries out switch identification number coding to virtual link path, this is conducive to carry out intersection, the mutation operation of genetic algorithm.
3. adopt binary system to carry out encode to switch connection matrix, effectively provide basis for estimation for virtual link path code.
Brief description of the drawings
Fig. 1 is the AFDX schematic network structure having built.
Fig. 2 is AFDX network topology structure schematic diagram of the present invention.
Fig. 3 is that the present invention adopts genetic algorithm to carry out the flow chart of AFDX network path optimization.
Fig. 4 is the AFDX network topology structure schematic diagram of embodiment 1.
Embodiment
Below in conjunction with drawings and Examples, the present invention is described in further detail.
For the AFDX network topology structure shown in Fig. 1, AFDX network packet contains n switch, adopts set form to be expressed as SWITH={S 1, S 2..., S a..., S b..., S n; Described SWITH={S 1, S 2..., S a..., S b..., S nmiddle S 1represent first switch, S 2represent second switch, S arepresent a switch, S brepresent b switch, S nrepresenting n switch, is also last switch; For general knowledge explanation, S aalso referred to as one of them any switch, S balso referred to as another any switch wherein, S nalso referred to as any one switch, a, b, the identification number that n is switch, n is also the sum of switch in AFDX network.In Fig. 1, n=8, has S 1, S 2, S 3, S 4, S 5, S 6, S 7, S 8.
For the AFDX network topology structure shown in Fig. 1, AFDX network packet contains m end system, adopts set form to be expressed as ES={E 1, E 2..., E c..., E d..., E m; Described ES={E 1, E 2..., E c..., E d..., E mmiddle E 1represent first end system, E 2represent second end system, E crepresent c end system, E drepresent d end system, E mrepresenting m end system, is also last end system; For general knowledge explanation, E calso referred to as one of them arbitrary end system, E dalso referred to as another arbitrary end system wherein, E malso referred to as any one end system, c, d, the identification number that m is end system.
For the AFDX network topology structure shown in Fig. 1, AFDX network packet contains q bar virtual link, adopts set form to be expressed as VL={L 1, L 2..., L e..., L f..., L q; Described VL={L 1, L 2..., L e..., L f..., L qmiddle L 1represent Article 1 virtual link, L 2represent Article 2 virtual link, L erepresent e article of virtual link, L frepresent f article of virtual link, L qrepresenting q article of virtual link, is also the last item virtual link; For general knowledge explanation, L ealso referred to as any virtual link wherein, L falso referred to as another any virtual link wherein, L qalso referred to as any virtual link, e, f, the identification number that q is virtual link.
The employing genetic algorithm that the present invention proposes is carried out the method for path optimization, and described AFDX network refers to the AFDX network that disposes VL path, and concrete path optimization includes step:
Step 1: build switch connection matrix
Carry out the physical connection initialization between all switches to disposing the AFDX network in VL path, obtain the n × n switch connection matrix that adopts matrix form to express
C nnrepresent n switch S nwith n switch S nconnection, represent the connection between switch self; In like manner, C 11represent the 1st switch S 1with the 1st switch S 1connection, C 22represent the 2nd switch S 2with the 2nd switch S 2connection, C aarepresent a switch S awith a switch S aconnection, C bbrepresent b switch S bwith b switch S bconnection, C abrepresent a switch S awith b switch S bconnection.
C 12represent the 1st switch S 1with the 2nd switch S 2connection, C 1nrepresent the 1st switch S 1with n switch S nconnection;
C 21represent the 2nd switch S 2with the 1st switch S 1connection, C 2nrepresent the 2nd switch S 2with n switch S nconnection;
C n1represent n switch S nwith the 1st switch S 1connection, C n2represent n switch S nwith the 2nd switch S 2connection.
In the present invention, the assignment condition of switch connection matrix SC has three kinds, the first is that the connection between switch self is designated as 0, and the second is between switch, not have the connection of physical link to be designated as 0, and the third has the connection of physical link to be designated as 1 between switch.Switch connection matrix as shown in Figure 1
SC = C 11 C 12 C 13 C 14 C 15 C 16 C 17 C 18 C 21 C 22 C 23 C 24 C 25 C 26 C 27 C 28 C 31 C 32 C 33 C 34 C 35 C 36 C 37 C 38 C 41 C 42 C 43 C 44 C 45 C 46 C 47 C 48 C 51 C 52 C 53 C 54 C 55 C 56 C 57 C 58 C 61 C 62 C 63 C 64 C 65 C 66 C 67 C 68 C 71 C 72 C 73 C 74 C 75 C 76 C 77 C 78 C 81 C 82 C 83 C 84 C 85 C 86 C 87 C 88 = 0 1 1 0 0 0 0 1 1 0 0 1 0 0 0 1 1 0 0 1 1 0 1 1 0 1 1 0 0 1 1 1 0 0 1 0 0 1 1 0 0 0 0 1 1 0 1 0 0 0 1 1 1 1 0 0 1 1 1 1 0 0 0 0 .
In the present invention, application binary " 0,1 " encode represents the physical connection of switch, is in order comprehensively to represent the connection state between switch in AFDX network.
Step 2: build virtual link path population
(A) virtual link path population scale U={u is set 1, u 2..., u x; According in AFDX network for the setting of virtual link path population scale U, the number in the virtual link path from source end system A to destination system B.Usually, population number x=3~20.
U 1represent the mark in the effective virtual link of Article 1 path;
U 2represent the mark in the effective virtual link of Article 2 path;
U xrepresent the mark in the effective virtual link of the last item path, for general knowledge explanation, also referred to as any effective virtual link path.
(B) switch matrix to step 1 choose and meet U={u according to path exclusion condition 1, u 2..., u xsource-object-switch path
represent the switch S of carrying source end system A a;
represent the switch S of carrying destination system B b;
For example, in Fig. 2: in the time of value population number x=6, source end system A is connected to switch S 1upper, destination system B is connected to switch S mupper, there is the effective virtual link of the Article 1 selecting according to path exclusion condition path R - u 1 = ( S a A → u 1 S b B ) = S 1 - S 3 - S 5 - S a - S b - S m ;
The effective virtual link of Article 2 path R - u 2 = ( S a A → u 2 S b B ) = S 1 - S 2 - S 4 - S 6 - S m ;
The effective virtual link of Article 3 path R - u 3 = ( S a A → u 3 S b B ) = S 1 - S 8 - S 4 - S 7 - S 6 - S m ;
The effective virtual link of Article 4 path R - u 4 = ( S a A → u 4 S b B ) = S 1 - S 3 - S 5 - S 6 - S b - S m ;
The effective virtual link of Article 5 path R - u 5 = ( S a A → u 5 S b B ) = S 1 - S 8 - S 4 - S 6 - S m ;
The effective virtual link of Article 6 path R - u 6 = ( S a A → u 6 S b B ) = S 1 - S 2 - S 4 - S 8 - S 3 - S 5 - S 6 - S m .
(C) in the present invention, described path exclusion condition includes three kinds of situations:
Eliminating situation in the first path refers to that source end system A and destination system B are connected on same switch, do not choose this effective virtual link path, referred to as single switch path.
The second path eliminating situation refer to source end system A to destination system B in the virtual link path of process, have the same switch of repetition, do not choose this effective virtual link path, there is loop referred to as path.
The third path eliminating situation refer to source end system A to destination system B in the virtual link path of process, between switch, there is no physical link, do not choose this effective virtual link path, referred to as without corresponding physical link.
For example, situation is got rid of in the second path, and source end system A is connected to switch S 1upper, destination system B is connected to switch S mupper, the path of the switch that passes through is S 1-S 3-S 5-S 3-S 7-S 6-S m, due at switch S 3on there is loop situation, get rid of described S 1-S 3-S 5-S 3-S 7-S 6-S mpath.
For example, situation is got rid of in the third path, and source end system A is connected to switch S 1upper, destination system B is connected to switch S mupper, the path of the switch that passes through is S 1-S 3-S 6-S b-S m, get rid of described S 1-S 3-S 6-S b-S mpath.Because switch S 3with switch S 6between there is no physical link, i.e. C in switch matrix SC 36connection code be " 0 ".
Step 3: build chromosome
Source-object-switch path that step 2 is obtained in the identification number of switch as gene, the encoded chromosome population CH={h that obtains 1, h 2..., h x; Then perform step four;
H 1represent first chromosomal mark;
H 2represent second chromosomal mark;
H xrepresent last chromosomal mark, for general knowledge explanation, also referred to as any one chromosome;
R - u 1 = ( S a A → u 1 S b B ) = S 1 - S 3 - S 5 - S a - S b - S m The encoded chromosome obtaining is h 1=1-3-5-a-b-m;
R - u 2 = ( S a A → u 2 S b B ) = S 1 - S 2 - S 4 - S 6 - S m The encoded chromosome obtaining is h 2=1-2-4-6-m;
R - u 3 = ( S a A → u 3 S b B ) = S 1 - S 8 - S 4 - S 7 - S 6 - S m The encoded chromosome obtaining is h 3=1-8-4-7-6-m;
R - u 4 = ( S a A → u 4 S b B ) = S 1 - S 3 - S 5 - S 6 - S b - S m The encoded chromosome obtaining is h 4=1-3-5-6-b-m;
R - u 5 = ( S a A → u 5 S b B ) = S 1 - S 8 - S 4 - S 6 - S m The encoded chromosome obtaining is h 5=1-8-4-6-m.
R - u 6 = ( S a A → u 6 S b B ) = S 1 - S 2 - S 4 - S 8 - S 3 - S 5 - S 6 - S m The encoded chromosome obtaining is h 6=1-2-4-8-3-5-6-m.
In the present invention, the chromosome in application genetic algorithm carries out the identification number coding of switch to virtual link path, and this is conducive to carry out intersection, the mutation operation of genetic algorithm.
Step 4: calculate chromosomal fitness function value
For equilibrium disposed VL path AFDX network flow and reduce propagation delay time of message, the present invention adopts maximum link load factor AVL maxpropagation delay time rate time with message postponesum is as fitness function Fitness=AVL max+ time postpone; Then perform step five.
In the present invention, AVL max = max n Σ VL ∈ C nn r max t BAG 100 × 10 6 × 100 % , R maxrepresent the virtual link VL={L in AFDX network 1, L 2..., L e..., L f..., L qmaximum bag long, t bAGrepresent the virtual link VL={L in AFDX network 1, L 2..., L e..., L f..., L qthe bandwidth allocation interval of transmission of messages, n represents the sum of switch in AFDX network, C nnrepresent n switch S nwith n switch S nconnection.
In the present invention, delay qrepresent any virtual link L that Adoption Network calculation method obtains qthe message transmission delay time, M qrepresent any virtual link L qmessage transmission delay boundary value.
What network calculus method adopted is " electric light and control " the 15th the 9th phase of volume, in September, 2008, and author Yang Yun, bear Hua Gang, the content of Section 3 in " calculating the network calculus method that AFDX postpones ", " utilizes the network delay of the theoretical AFDX of calculating of network calculus ".
In the present invention, according to fitness function Fitness=AVL max+ time postponecalculate CH={h 1, h 2..., h xin each chromosomal fitness function value be designated as FCH={Fh 1, Fh 2..., Fh x, and from described FCH={Fh 1, Fh 2..., Fh xin select minimum fitness function value Fh minimum.
In the present invention, calculate all chromosomal fitness function values, the fitness function value of different virtual link paths differs greatly, if the method that directly uses roulette to select will cause some chromosomes almost can not be copied in the next generation and go, has had a strong impact on the diversity of population.The less chromosome representing of fitness function value is more excellent.
Step 5: genetic manipulation
In the present invention, when the chromosome population of former generation is designated as CH i, previous generation chromosome population is designated as CH i-1, chromosome population of future generation is designated as CH i+1.
Step 501, chromosomal the first cross modal:
In the present invention, the mode that uses single-point to intersect: (A) first crosspoint is selected in public branch exchange place, then to the chromosome population CH when former generation iin chromosome carry out exchange pairing, the part after exchange pairing crosspoint fills in chromosome separately, thus two new chromosomes that generate; Finally described new chromosome is added to virtual link of future generation path population CH i+1in.
In the present invention, the mode that uses single-point to intersect: (B) first crosspoint is selected in public branch exchange place, then to the chromosome population CH when former generation iin chromosome carry out exchange pairing, the part after chiasma just enters in chromosome separately, thus two new chromosomes that generate; If duplicate gene (being that loop appears in path) in two described new chromosomes, delete the switch of any one repetition, and the new chromosome of having deleted after repeated exchanged machine is added to virtual link of future generation path population CH i+1in.
In the present invention, described public branch exchange place refers to except source switch with object switch switch in addition.If there are multiple public branch exchanges, select at random the crosspoint of a public branch exchange as genetic manipulation.
For example: chromosome h 2=1-2-4-6-m and chromosome h 3=1-8-4-7-6-m selects switch S 4as crosspoint, after intersecting, obtain two new chromosomes,
For example: chromosome h 3=1-8-4-7-6-m and chromosome h 6=1-2-4-8-3-5-6-m selects switch S 4as crosspoint, after intersecting, obtain two new chromosomes, ? the middle switch S repeating that occurred 8, delete the switch S of a rear repetition 8, and by the new chromosome of having deleted after repeated exchanged machine add in the population of virtual link of future generation path.
Step 502, chromosomal the second cross modal:
In the present invention, the mode that uses single-point to intersect: (A) first crosspoint is not selected to be in public branch exchange place, then to the chromosome population CH when former generation iin chromosome select the chromosome in the middle of being positioned to intersect, the part after exchange pairing crosspoint fills in chromosome separately, thus two new chromosomes that generate; Finally described new chromosome is added to virtual link of future generation path population CH i+1in.
In the present invention, the mode that uses single-point to intersect: (B) first crosspoint is not selected to be in public branch exchange place, then to the chromosome population CH when former generation iin chromosome select the chromosome in the middle of being positioned to intersect, the part after exchange pairing crosspoint fills in chromosome separately, thus two new chromosomes that generate; If duplicate gene (being that loop appears in path) in two described new chromosomes, delete the switch of any one repetition, and add and in the population of virtual link of future generation path, add virtual link of future generation path population CH having deleted new chromosome after repeated exchanged machine i+1in.
For example: chromosome h 1=1-3-5-a-b-m and chromosome h 5=1-8-4-6-m selects that middle switch of described chromosome (at h 1in=1-3-5-a-b-m, be switch S 5, at h 5in=1-8-4-6-m, be switch S 4) as crosspoint, after intersecting, obtain two new chromosomes,
Step 503, chromosomal the first variant form:
In the present invention, use the mode of single-point variation: (A) to the chromosome population CH when former generation ifirst change point is selected in except source switch with object switch any switch place in addition, selects to exist link-attached another switch of physics as variation result with described switch, obtains new chromosome after variation, then by described variation afterwards new chromosome add virtual link of future generation path population CH i+1.
In the present invention, use the mode of single-point variation: (B) to the chromosome population CH when former generation ifirst change point is selected in except source switch with object switch any switch place in addition, selects to exist link-attached another switch of physics as variation result with described switch new chromosome after acquisition variation; If duplicate gene (being that loop appears in path) in new chromosome after variation, delete the switch of any one repetition, then the new chromosome of having deleted after repeated exchanged machine is added to virtual link of future generation path population CH i+1in.
For example: chromosome h 4=1-3-5-6-b-m selects switch S 5make a variation, switch S 5variation is switch S 7, obtain new chromosome and be
For example: chromosome h 6=1-2-4-8-3-5-6-m selects switch S 8make a variation, switch S 8variation is switch S 2, obtain new chromosome and be delete the switch S of a rear repetition 2, the new chromosome that has obtained deleting after repeated exchanged machine is
Step 504, chromosomal the second variant form:
In the present invention, use the mode of single-point variation, to the chromosome population CH when former generation ifirst change point is selected in except source switch with object switch any switch place in addition, directly deletes described switch, obtains new chromosome after variation, then by described variation afterwards new chromosome add virtual link of future generation path population CH i+1in.
For example: chromosome h 1=1-3-5-a-b-m selects switch S amake a variation, directly delete switch S aobtaining new chromosome is
In the example of enumerating in the present invention, new chromosome to add virtual link of future generation path population CH i+1in select operation.
For the variation in the present invention, there is a kind of chromosome only containing two genes, be source switch and object switch, if in initial population, not containing this chromosome, above crossover and mutation operation can not produce this chromosome yet, and this chromosome is probably optimal solution, therefore here to mutation operation correct, this chromosome may be occurred after variation.After random selection change point, not necessarily to change gene, also can directly delete gene.
Step 505, chromosomal selection mode:
In the present invention, chromosomal selection mode is by the chromosome population CH when former generation i={ h 1, h 2..., h xand by the new chromosome population of intersect-variation generation mix, obtain mixed population then pass through Fitness=AVL max+ time postponedescribed in calculating in each chromosomal fitness function value and choose x minimum chromosome as population CH of future generation i+1.
In the example of enumerating in the present invention, chromosomal selection mode is by the chromosome population CH when former generation i={ h 1, h 2, h 3, h 4, h 5, h 6and by the new chromosome population of intersect-variation generation mix, obtain mixed population then pass through Fitness=AVL max+ time postponedescribed in calculating in each chromosomal fitness function value and choose six minimum chromosomes as population CH of future generation i+1.
Step 6: judge whether to meet hereditary end condition
Judge the fitness function value Fh of described minimum minimumafter continuous 3 generation genetic manipulations, described Fh minimumwhile variation, stop genetic manipulation; If do not meet, to CH={h 1, h 2..., h xin every item chromosome re-start the genetic manipulation of step 5.
Described termination genetic manipulation refers to that the chromosome of selecting fitness function value minimum in generation population is in the end as optimum results, from chromosome, extract the switch of source end system A to destination system B process, thereby obtain effective virtual link path.
embodiment 1
The present embodiment is to carry out emulation based on matlab (version number 7.13) platform; Matlab is a kind of advanced techniques computational language and interactive environment for algorithm development, data visualization, data analysis and numerical computations.
Shown in Figure 4, to having configured in the AFDX network in VL path, there are 7 virtual links: i.e. L 1configuration path be S 1-S 2-S 4-S 6, source end system is E 1, destination system is E 6; L 2configuration path be S 2-S 4-S 6, source is system E 2, destination system is E 6; L 3configuration path be S 3-S 7-S 6, source end system is E 3, destination system is E 6; L 4configuration path be S 4-S 6, source end system is E 4, destination system is E 6; L 5configuration path be S 5-S 6, source end system is E 5, destination system is E 6; L 6configuration path be S 7-S 6, source end system is E 7, destination system is E 6; L 7configuration path be S 8-S 3-S 5-S 6, source end system is E 8, destination system is E 6.Configuration r max=4000bits, t bAG=4ms, M q=1ms.
Choose virtual link L 1be optimized its source end system E 1(it is connected to switch S 1on), destination system E 6(it is connected to switch S 6on), from source end system E 1to destination system E 6path be S 1-S 2-S 4-S 6, L before Adoption Network calculation is optimized 1the propagation delay time of bearer messages is 629.6 microseconds.
The switch matrix building according to Fig. 4 SC = 0 1 1 0 0 0 0 1 1 0 0 1 0 0 0 1 1 0 0 1 1 0 1 1 0 1 1 0 0 1 1 1 0 0 1 0 0 1 1 0 0 0 0 1 1 0 1 0 0 0 1 1 1 1 0 0 1 1 1 1 0 0 0 0 ;
While making population number x=3, the effective virtual link of the Article 1 path selecting according to path exclusion condition R - u 1 = ( S 1 E 1 → u 1 S 6 E 6 ) = S 1 - S 2 - S 4 - S 6 , The encoded chromosome that obtains is h 1=1-2-4-6;
The effective virtual link of Article 2 path R - u 2 = ( S 1 E 1 → u 2 S 6 E 6 ) = S 1 - S 8 - S 4 - S 6 , The encoded chromosome that obtains is h 2=1-8-4-6;
The effective virtual link of Article 3 path R - u 3 = ( S 1 E 1 → u 3 S 6 E 6 ) = S 1 - S 8 - S 2 - S 4 - S 7 - S 6 , The encoded chromosome that obtains is h 3=1-8-2-4-7-6.
According to fitness function Fitness=AVL max+ time postponecalculate CH 1={ h 1, h 2, h 3in each chromosomal fitness function value be designated as FCH={Fh 1, Fh 2, Fh 3}={ 0.9096,0.868,1.022}, selects minimum fitness function value Fh minimum=0.868.
The operation of first generation genetic manipulation, to CH 1={ h 1, h 2, h 3carry out genetic manipulation, by chromosome h 1=1-2-4-6 and chromosome h 3=1-8-2-4-7-6 is with switch S 2for crosspoint intersects, obtain chromosome h 2=1-8-4-6 selects switch S 8for change point makes a variation, switch S 8variation is switch S 3, obtain new chromosome and be by the chromosome population CH when former generation 1={ h 1, h 2, h 3and by the new chromosome population of intersect-variation generation mix, obtain mixed population then pass through Fitness=AVL max+ time postponecalculate each chromosomal fitness function value is respectively choose the wherein chromosome h of 3 fitness function value minimums 1, h 2, as population CH of future generation 2chromosome, now, CH 2={ h 1, h 2, h 3the chromosome h of population 1=1-2-4-6, chromosome h 2=1-8-4-6, chromosome h 3=1-3-4-6.Fh in this generation population minimum=0.868, now do not meet hereditary end condition.
The operation of second generation genetic manipulation, to CH 2={ h 1, h 2, h 3carry out genetic manipulation, by chromosome h 1=1-2-4-6 and chromosome h 2=1-8-4-6 is with switch S 4for crosspoint intersects, obtain chromosome h 3=1-3-4-6 selects switch S 4for change point makes a variation, switch S 4variation is switch S 3, obtain new chromosome and be by the chromosome population CH when former generation 2={ h 1, h 2, h 3and by the new chromosome population of intersect-variation generation mix, obtain mixed population then pass through Fitness=AVL max+ time postponecalculate each chromosomal fitness function value is respectively choose the wherein chromosome h of 3 fitness function value minimums 2, h 3, as population CH of future generation 3chromosome, now, CH 3={ h 1, h 2, h 3the chromosome h of population 1=1-8-4-6, chromosome h 2=1-3-4-6, chromosome h 3=1-8-4-6.Fh in this generation population minimum=0.868, now do not meet hereditary end condition.
The operation of third generation genetic manipulation, to CH 3={ h 1, h 2, h 3carry out genetic manipulation, by chromosome h 1=1-8-4-6 and chromosome h 3=1-8-4-6 is with switch S 4for crosspoint intersects, obtain chromosome h 2=1-3-4-6 selects switch S 3for change point makes a variation, switch S 3variation is switch S 2, obtain new chromosome and be by the chromosome population CH when former generation 3={ h 1, h 2, h 3and by the new chromosome population of intersect-variation generation mix, obtain mixed population then pass through Fitness=AVL max+ time postponecalculate each chromosomal fitness function value is respectively choose the wherein chromosome h of 3 fitness function value minimums 1, h 2, h 3as population CH of future generation 4chromosome, now, CH 4={ h 1, h 2, h 3the chromosome h of population 1=1-8-4-6, chromosome h 2=1-3-4-6, chromosome h 3=1-8-4-6.Fh in this generation population minimum=0.868, now after continuous 3 generation genetic manipulations, Fh minimumdo not change, stop genetic manipulation, be i.e. selected population CH 4={ h 1, h 2, h 3in the chromosome of fitness function value minimum as optimum results, be h here 1=1-8-4-6, has source end system E 1to destination system E 6the switch of process is S 1-S 8-S 4-S 6, by path S 1-S 8-S 4-S 6as virtual link L 1path as optimum results.
Have and optimize front L 1path be S 1-S 2-S 4-S 6path as shown in triangle as hollow in dotted line in Fig. 4, the L after genetic manipulation is optimized 1path be S 1-S 8-S 4-S 6path as shown in triangle as solid in dotted line in Fig. 4.L after optimizing 1the propagation delay time of bearer messages is 588 microseconds, with optimize before 629.6 microseconds compared with, its propagation delay time has obtained reducing, the real-time of transmission of messages has obtained enhancing.

Claims (5)

1. adopt genetic algorithm to carry out a method for AFDX network path optimization, described AFDX network refers to the AFDX network that disposes VL path; It is characterized in that including the following step:
Step 1: build switch connection matrix;
Carry out the connection initialization between all switches to disposing the AFDX network in VL path, obtain n × n switch connection matrix n is the number of the switch in AFDX network;
Step 2: build virtual link path population;
Virtual link path population scale U={u is set 1, u 2..., u x, the switch matrix SC of step 1 is chosen and meets U={u according to path exclusion condition 1, u 2..., u xsource-object-switch path
Step 3: build chromosome;
Source-object-switch path that step 2 is obtained in the identification number of switch as gene, the encoded chromosome population CH={h that obtains 1, h 2..., h x;
Step 4: calculate chromosomal fitness function value;
Described fitness function Fitness=AVL max+ time postpone, wherein, maximum link load factor AVL max = max n Σ VL ∈ C nn r max t BAG 100 × 10 6 × 100 % , Message transmission delay time rate
Step 5: genetic manipulation;
Current chromosome population is designated as CH i, previous generation chromosome population is designated as CH i-1, chromosome population of future generation is designated as CH i+1;
Step 501, chromosomal the first cross modal:
The mode that uses single-point to intersect, crosspoint is selected in public branch exchange place, if there are multiple public branch exchanges, select at random one as crosspoint, then the part after exchange pairing crosspoint, two new chromosomes that generate, and described new chromosome is added in the population of virtual link of future generation path; If duplicate gene in new chromosome after intersecting, delete the switch of any one repetition, and the new chromosome of having deleted after repeated exchanged machine is added in the population of virtual link of future generation path;
Step 502, chromosomal the second cross modal:
The mode that uses single-point to intersect, crosspoint is not selected to be in public branch exchange place, selects the chromosome in the middle of being positioned to intersect, the then part after exchange pairing crosspoint, two new chromosomes that generate, and described new chromosome is added in the population of virtual link of future generation path; If duplicate gene in new chromosome after intersecting, delete the switch of any one repetition, and add in the population of virtual link of future generation path and add in the population of virtual link of future generation path having deleted new chromosome after repeated exchanged machine.
Step 503, chromosomal the first variant form:
The mode that uses single-point variation, change point is selected in except source switch with object switch any switch place in addition, selects to exist link-attached another switch of physics as variation result with described switch, obtains new chromosome, and described new chromosome is added in the population of virtual link of future generation path; If duplicate gene in new chromosome after variation, delete the switch of any one repetition, and the new chromosome of having deleted after repeated exchanged machine is added in the population of virtual link of future generation path.
Step 504, chromosomal the second variant form:
The mode that uses single-point variation, change point is selected in except source switch with object switch any switch place in addition, directly deletes described switch, obtains new chromosome, and described new chromosome is added in the population of virtual link of future generation path.
Step 505, chromosomal selection mode:
Chromosomal selection mode is by the chromosome population CH when former generation i={ h 1, h 2..., h xand by the new chromosome population of intersect-variation generation mix, obtain mixed population then pass through Fitness=AVL max+ time postponedescribed in calculating in each chromosomal fitness function value and choose x minimum chromosome as population CH of future generation i+1.
Step 6: judge whether to meet hereditary end condition;
Judge the fitness function value Fh of described minimum minimumafter continuous 3 generation genetic manipulations, described Fh minimumwhile variation, stop genetic manipulation; If do not meet, to CH={h 1, h 2..., h xin every item chromosome carry out genetic manipulation.
2. the method that employing genetic algorithm according to claim 1 is carried out the optimization of AFDX network path, it is characterized in that: the assignment condition of described switch connection matrix SC has three kinds, the first is that the connection between switch self is designated as 0, the second is between switch, not have the connection of physical link to be designated as 0, and the third has the connection of physical link to be designated as 1 between switch.
3. the method that employing genetic algorithm according to claim 1 is carried out the optimization of AFDX network path, is characterized in that: the described path exclusion condition in step 2 includes three kinds of situations;
Eliminating situation in the first path refers to that source end system A and destination system B are connected on same switch, do not choose this virtual link path;
The second path eliminating situation refer to source end system A to destination system B in the virtual link path of process, have the same switch of repetition, do not choose this virtual link path;
The third path eliminating situation refer to source end system A to destination system B in the virtual link path of process, between switch, there is no physical link, do not choose this virtual link path.
4. the method that employing genetic algorithm according to claim 1 is carried out the optimization of AFDX network path, is characterized in that: in virtual link path construction process, the value of population number is 3~20 virtual link paths.
5. the method that employing genetic algorithm according to claim 1 is carried out the optimization of AFDX network path, is characterized in that: can reduce the propagation delay time of the virtual link message of AFDX network, reduce to reach 5~15%.
CN201410441205.3A 2014-09-01 2014-09-01 Method for carrying out AFDX network path optimization by genetic algorithm Active CN104202188B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201410441205.3A CN104202188B (en) 2014-09-01 2014-09-01 Method for carrying out AFDX network path optimization by genetic algorithm

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201410441205.3A CN104202188B (en) 2014-09-01 2014-09-01 Method for carrying out AFDX network path optimization by genetic algorithm

Publications (2)

Publication Number Publication Date
CN104202188A true CN104202188A (en) 2014-12-10
CN104202188B CN104202188B (en) 2017-04-26

Family

ID=52087418

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201410441205.3A Active CN104202188B (en) 2014-09-01 2014-09-01 Method for carrying out AFDX network path optimization by genetic algorithm

Country Status (1)

Country Link
CN (1) CN104202188B (en)

Cited By (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104504477A (en) * 2014-12-30 2015-04-08 中山大学 Route optimizing method based on evolutionary mechanism of birds
CN105071947A (en) * 2015-07-06 2015-11-18 山东超越数控电子有限公司 Link optimization method for SDN network
CN106656797A (en) * 2016-11-30 2017-05-10 中国民航大学 Method for optimizing AFDX (Avionics Full Duplex Switched Ethernet) network path under consideration of risk balance
WO2017107127A1 (en) * 2015-12-24 2017-06-29 Orange Method and system for optimisation of virtual network functions under network function virtualisation environment
CN107396421A (en) * 2017-08-11 2017-11-24 深圳市益鑫智能科技有限公司 A kind of efficient wireless sensor network path optimizing system
CN107404437A (en) * 2017-07-25 2017-11-28 广东工业大学 The determining method of path and device of a kind of message
CN108268710A (en) * 2018-01-03 2018-07-10 北京航空航天大学 A kind of IMA system dynamic restructuring policy optimization methods based on genetic algorithm
CN109598341A (en) * 2018-11-21 2019-04-09 济南浪潮高新科技投资发展有限公司 A kind of detection of convolutional neural networks training result and method for visualizing based on genetic algorithm
CN113379228A (en) * 2021-06-08 2021-09-10 天津大学 Dispatching optimization method for garbage collection and transportation vehicle

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101677286A (en) * 2008-09-19 2010-03-24 中国电信股份有限公司 Optimization method of carrier network
CN102238686A (en) * 2011-07-04 2011-11-09 南京邮电大学 Wireless sensor network routing method for modeling quantum genetic algorithm
US20120084525A1 (en) * 2009-06-05 2012-04-05 AIRBUS OPERATIONS(inc as a Societe par Act Simpl) Method and device for loading and executing instructions with deterministic cycles in a multicore avionic system having a bus of which the access time is not predictable

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101677286A (en) * 2008-09-19 2010-03-24 中国电信股份有限公司 Optimization method of carrier network
US20120084525A1 (en) * 2009-06-05 2012-04-05 AIRBUS OPERATIONS(inc as a Societe par Act Simpl) Method and device for loading and executing instructions with deterministic cycles in a multicore avionic system having a bus of which the access time is not predictable
CN102238686A (en) * 2011-07-04 2011-11-09 南京邮电大学 Wireless sensor network routing method for modeling quantum genetic algorithm

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
桂超等: ""基于遗传算法的最短路径路由优化算法"", 《微计算机信息》 *

Cited By (13)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104504477B (en) * 2014-12-30 2018-01-09 中山大学 A kind of method for optimizing route based on birds spore mechanism
CN104504477A (en) * 2014-12-30 2015-04-08 中山大学 Route optimizing method based on evolutionary mechanism of birds
CN105071947B (en) * 2015-07-06 2019-03-08 山东超越数控电子有限公司 A kind of link optimizing method of SDN network
CN105071947A (en) * 2015-07-06 2015-11-18 山东超越数控电子有限公司 Link optimization method for SDN network
WO2017107127A1 (en) * 2015-12-24 2017-06-29 Orange Method and system for optimisation of virtual network functions under network function virtualisation environment
CN106656797A (en) * 2016-11-30 2017-05-10 中国民航大学 Method for optimizing AFDX (Avionics Full Duplex Switched Ethernet) network path under consideration of risk balance
CN106656797B (en) * 2016-11-30 2019-05-31 中国民航大学 A method of considering the AFDX network path optimization of risk balance
CN107404437B (en) * 2017-07-25 2018-05-18 广东工业大学 The determining method of path and device of a kind of message
CN107404437A (en) * 2017-07-25 2017-11-28 广东工业大学 The determining method of path and device of a kind of message
CN107396421A (en) * 2017-08-11 2017-11-24 深圳市益鑫智能科技有限公司 A kind of efficient wireless sensor network path optimizing system
CN108268710A (en) * 2018-01-03 2018-07-10 北京航空航天大学 A kind of IMA system dynamic restructuring policy optimization methods based on genetic algorithm
CN109598341A (en) * 2018-11-21 2019-04-09 济南浪潮高新科技投资发展有限公司 A kind of detection of convolutional neural networks training result and method for visualizing based on genetic algorithm
CN113379228A (en) * 2021-06-08 2021-09-10 天津大学 Dispatching optimization method for garbage collection and transportation vehicle

Also Published As

Publication number Publication date
CN104202188B (en) 2017-04-26

Similar Documents

Publication Publication Date Title
CN104202188A (en) Method for carrying out AFDX network path optimization by genetic algorithm
CN103298009B (en) The modeling of complicated wireless network control system and fault propagation method
CN104299053A (en) Method for selecting optimal path of power distribution and utilization communication network based on genetic algorithm
CN102711125B (en) Method for improving transmission capability of wireless mesh network
CN103595623B (en) Opportunistic routing behavioral modeling method based on mobile social network node social characteristics
CN103957032A (en) Load redistribution method for electric power coupling network to resist cascade failure
CN101777990B (en) Method for selecting multi-objective immune optimization multicast router path
CN105553869A (en) Method and system for risk balancing for power communication network
CN105722206A (en) Energy efficiency optimization algorithm of multi-user OFDM relay system
CN105467839A (en) Multi-agent system security consensus control method in malicious environment
CN103428805B (en) The virtual mapping method of a kind of wireless network based on link anti-interference
CN105719016A (en) Immune algorithm-based siting planning method of electric power communication network access points
CN107509229B (en) A kind of car networking chance method for routing calculated based on vehicle node centrality
Duell et al. System optimal dynamic lane reversal for autonomous vehicles
CN104811479B (en) A kind of creation method based on forwarding with the virtual network node of control separated protocol
CN103368658A (en) Four-photon entangled W state-based super-dense coding method for quantum signaling
CN105573306A (en) Formation method and device for multiple intelligent agents with blind areas
CN104144134B (en) A kind of array dispatching method and device
CN103825845A (en) Matrix decomposition-based packet scheduling algorithm of reconfigurable VOQ (virtual output queuing) structure switch
CN105453494B (en) Virtual shortest path tree is established and the method and path-calculating element of processing
CN103051537B (en) IP Circuit Service transmission dispatching method and system
CN111193971B (en) Machine learning-oriented distributed computing interconnection network system and communication method
CN107360480A (en) A kind of ONU packet planing methods based on load balancing
CN109039766B (en) Power CPS network risk propagation threshold determination method based on seepage probability
CN107294658B (en) Coding node selection method in network control coding

Legal Events

Date Code Title Description
PB01 Publication
C06 Publication
SE01 Entry into force of request for substantive examination
C10 Entry into substantive examination
GR01 Patent grant
GR01 Patent grant