CN103647603A - Method for optimizing blind ultraviolet non line-of-sight Ad-hoc communication network shared channel - Google Patents

Abstract

The invention discloses a method for optimizing a blind ultraviolet non line-of-sight Ad-hoc communication network shared channel. The method comprises the following steps: firstly, drawing up conflict graphs under three communication modes according to the characteristics of a blind ultraviolet non line-of-sight Ad-hoc network, and proposing a conflict avoidance model; then constructing a link conflict graph based on the conflict avoidance model, and proposing a dyeing theory to be applied on a time slot allocation control method of the blind ultraviolet non line-of-sight Ad-hoc communication network shared channel; further proposing a UVAd-TDMA protocol suitable for the blind ultraviolet non line-of-sight Ad-hoc network. The local link conflict graphs obtained by the method are small in color point quantity, namely the time slots divided within unit time are few, so that the target that different complementary interference channels are distributed to conflicted links by using the least time slot quantity within the unit time by the blind ultraviolet non line-of-sight Ad-hoc communication network is achieved.

Description

Day blind ultraviolet non line of sight Ad-hoc communication network shared channel optimization method

Technical field

The present invention relates to a day blind ultraviolet non line of sight Ad-hoc communication network shared channel optimization method.

Background technology

Day territory, blind area ultraviolet communication is by atmospheric scattering, to carry out a kind of communication mode of communication, has that non line of sight NLOS (Non-Line-of-Sight) transmission, good confidentiality, antijamming capability are strong, construction and an advantage such as maintenance cost is low.Above-mentioned advantage makes day territory, a blind area ultraviolet communication can be used as a kind of good alternative of radio communication and is widely used in the environment of rural area, city outdoor environment and military communication.But the strong absorption to solar blind UV due to near-earth atmospheric molecule and particulate, is restricted its transmission range.

Wireless Ad-hoc Network is a kind of interim acentric distributed network, and its node can move arbitrarily, and the status of each node is equality.Source node can arrive destination node by multi-hop.And day blind ultraviolet non line of sight Ad-hoc communication network can be strong in conjunction with the transmission of ultraviolet communication non line of sight, antijamming capability, the advantage of good confidentiality and the multi-hop of Ad-hoc network, realize transmission at a distance reliably, made up the short defect of solar blind UV non line of sight communication system communication distance.Visible, a day blind ultraviolet non line of sight Ad-hoc communication network is the new effective that meets all kinds of mobile context communicating requirements.

The communication of solar blind UV non line of sight has three kinds of modes: the debit of omnidirectional formula is sent out by omnidirectional, is characterized in: the bandwidth of the scattering volume that positioning requirements is minimum, unlimited, minimum; The directed debit of the omnidirectional formula of sending out, its feature is: in scattering volume, have maximum ultraviolet light energy, limited scattering volume, larger bandwidth; Directed send out directed debit's formula, its feature is: the bandwidth of limited scattering volume, maximum.Accordingly, set up ultraviolet light network,---the neighbours of omnidirectional, omnidirectional---directed neighbours, orientation---situations such as directed neighbours that its internodal neighbouring relations can be divided into omnidirectional by communication mode.Therefore, be similar to the problem that conventional radio frequency communication network is used directional antenna to produce, a day blind ultraviolet non line of sight Ad-hoc communication network adopts directional beam will produce deaf node during contention channel resource in practical communication, the problems such as asymmetric, exposed terminal and new concealed terminal gain.

The problem existing during day blind ultraviolet non line of sight Ad-hoc communication network node shared channel is in fact all to cause owing to existing a plurality of ultraviolet light node transport services to produce conflict in the directivity Ji omnidirectional of directional beam or directional beam coverage.Therefore, one have n ultraviolet light node day blind ultraviolet non line of sight Ad-hoc network system in, due to the directivity of directional beam and the space-time Competition Characteristics of data flow wherein, node in deaf region cannot be communicated by letter, and when mutually the source node of interference region or destination node send at the same time or receive data by the conflict on generation time and space, thereby cause the source node that can successfully transmit data simultaneously---destination node is to being limited.In order to make successfully to transmit within the unit interval, the source node of data---destination node is maximum to number, can concurrent link maximum in blind ultraviolet non line of sight Ad-hoc communication network of this day, just need to carry out transmitting and scheduling to limited ultraviolet light channel resource.Here it is shares the space and time optimization assignment problem of ultraviolet light channel.

Summary of the invention

The problems referred to above that exist for prior art, the technical problem to be solved in the present invention is: how by day blind ultraviolet non line of sight Ad-hoc communication network node shared channel space and time optimization, to distribute, make a day blind ultraviolet Ad-hoc communication network reach good communication efficiency.

For achieving the above object, the present invention adopts following technical scheme: day blind ultraviolet non line of sight Ad-hoc communication network

Shared channel optimization method, comprises the steps:

Step 1: model is avoided in the conflict of having set up under day three kinds of channelling modes of blind ultraviolet non line of sight Ad-hoc communication network:

The conflict of S11: omnidirectional---directional communication mode avoids model suc as formula (1):

|N1-N2|≥(1+Δ)(r ₁+r ₂) （1）；

Wherein, Δ >0, represents boundary belt region, and N1 and N2 all represent the node of day blind ultraviolet non line of sight Ad-hoc communication network, r ₁and r ₂represent respectively node N1, effective propagation path during N2 omnidirectional transfer of data;

S12: omnidirectional---conflict under beam communication mode avoids model suc as formula (2):

|N1-N2|≥(1+Δ)(r ₁'+r ₂) （2）；

Wherein, Δ >0, represents boundary belt region, and N1 and N2 all represent the node of day blind ultraviolet non line of sight Ad-hoc communication network, r ₁' the effective propagation path of expression node N1 when directional data transmits, r ₂effective propagation path while representing node N2 omnidirectional transfer of data;

S13: orientation---conflict under beam communication mode avoids model suc as formula (3):

|N1-N2|≥(1+Δ)(r ₁'+r ₂') （3）；

Wherein, Δ >0, represents boundary belt region, and N1 and N2 all represent the node of day blind ultraviolet non line of sight Ad-hoc communication network, r ₁' the effective propagation path of expression node N1 when directional data transmits, r ₂' the effective propagation path of expression node N2 when directional data transmits;

Step 2: build the local link conflict graph of day blind ultraviolet non line of sight Ad-hoc communication network, concrete steps are as follows:

S21: structure link interference approximate model:

Every link neighbours' link in only jumping with self ξ exists and disturbs,

r=min{r wherein _i, r _i' represent the transmission radius of node, R'=max{r _i, r _i' represent the interference radius of node;

S22: utilize the computing capability of node in day blind ultraviolet non line of sight Ad-hoc communication network to calculate, each node produces the broadcast data packet that includes own business transmission and suffered interfere information, with omnidirectional's light beam, this broadcast data packet is sent to the neighbours' link nodes in its ξ jumping;

S23: the ξ that each node is preserved it jumps the broadcast packet that interior nodes sends, and generate the local link conflict graph of day blind ultraviolet non line of sight Ad-hoc communication network;

Step 3: the local link conflict graph of day blind ultraviolet non line of sight Ad-hoc communication network of setting up according to step 2 calculates optimum dyeing sequence, specific as follows:

S31: the local link conflict graph of establishing day blind ultraviolet non line of sight Ad-hoc communication network of step 2 foundation is M _{λ, c}=(Λ, l _{λ, c}), the μ-Vertex Coloring of this local link conflict graph is from local link conflict graph M _{λ, c}=(Λ, l _{λ, c}) vertex set Λ to a mappings of chromatic number collection PC (μ); And if only if λ _i, λ _j∈ Λ, and λ _iλ _j∈ l _{λ, c}time, σ (λ _i) ≠ σ (λ _j), local link conflict graph M _{λ, c}=(Λ, l _{λ, c}) the set of μ-Vertex Coloring be designated as Ξ _μ(M _{λ, c}), if | Ξ _μ(M _{λ, c}) | ≠ 0, claim local link conflict graph M _{λ, c}=(Λ, l _{λ, c}) be μ-stainable, wherein, λ _i, λ _jrepresent respectively link conflict graph M _{λ, c}=(Λ, l _{λ, c}) vertex set Λ in i and j summit, l _{λ, c}represent conflict link set, σ (λ _i) expression λ _idye number, σ (λ _j) expression λ _jdye number, chromatic number collection PC (μ) represents the set that dye number forms;

S32: adopt genetic algorithm to combine and carry out the dyeing of link conflict graph with greedy algorithm, concrete steps are as follows:

S321: build chromosome space:

With A (M _{λ, c})=(ai _j) _{k * k}represent local link conflict graph M _{λ, c}=(Λ, l _{λ, c}) adjacency matrix, specifically suc as formula (4):

$a_{\mathrm{ij}}\left\{\begin{array}{ccc}1& \mathrm{if}& {\mathrm{\&lambda;}}_i{\mathrm{\&lambda;}}_j\&Element;l_{\mathrm{\&lambda;},\mathrm{<figure-callout\; id="0"\; label="c"\; filenames="HDA0000450489840000052.png"\; state="\{\{state\}\}">c</figure-callout>}}\\ 0& \mathrm{if}& {\mathrm{\&lambda;}}_i{\mathrm{\&lambda;}}_j\&NotElement;l_{\mathrm{\&lambda;},c}\end{array}\right.---\left(4\right);$

Wherein k represents link conflict graph M _{λ, c}=(Λ, l _{λ, c}) exponent number, a _ijrepresent adjacency matrix A (M _{λ, c})=(ai _j) _{k * k}in element, to local link conflict graph M _{λ, c}=(Λ, l _{λ, c}) all dyeing forecast scheme configuration chromosomes space on summit;

Following coding is made in chromosome space:

If M _{λ, c}=(Λ, l _{λ, c}) a kind of order on k summit be n ₁, n ₂... n _k, n wherein ₁, n ₂... n _kbe natural 1,2 ..., a kind of arrangement of k}, the sequence x that is k corresponding to a length ₁x ₂... x _i... x _k, x wherein _iexpression is to M _{λ, c}=(Λ, l _{λ, c}) summit λ _icolor, x _i∈ PC (μ);

After being encoded in chromosome space, the initial population that the N bar chromosome that utilizes greedy algorithm generation to be comprised of integer forms, the sequence x that is k corresponding to a length ₁x ₂... x _i... x _kbecome item chromosome x, x _ibecome this chromosomal i gene;

S322: set fitness function:

Definition penalty function is suc as formula (5):

$p\left(x\right)=\underset{i=1.j=i+1}{\overset{k}{\mathrm{\&Sigma;}}}\mathrm{\&eta;}({\mathrm{\&lambda;}}_i,{\mathrm{\&lambda;}}_j)---\left(5\right);$

And

η (x) represents that in chromosome space, chromosome x violates the number on the limit of constraint;

Definition fitness function is suc as formula (7):

$\mathrm{\&rho;}(x_1{x}_2...x_k)=\frac{1}{p\left(x\right)+|\mathrm{\&chi;}\left(x\right)-\mathrm{\&epsiv;}-1|}---\left(7\right);$

Wherein χ (x) is chromosome x=(x ₁x ₂... x _k) in number of colours used, ε is color minimum value used in the μ that tried to achieve-can dyeing scheme;

S333: probability γ is selected in heredity _idetermine:

For fitness value, be ρ _ii bar chromosome x, probability γ is selected in its heredity _iavailable formula (8) calculates:

${\mathrm{\&gamma;}}_i=\frac{{\mathrm{\&rho;}}_i}{\underset{j=1}{\overset{\mathrm{pop}\_\mathrm{size}}{\mathrm{\&Sigma;}}}{\mathrm{\&rho;}}_j}---\left(8\right);$

The value of the upper limit of wherein suing for peace pop_size is population scale N;

S324: the determining of crossover algorithm:

If x=is (x ₁x ₂..x _i..x _k) and z=(z ₁z ₂... z _i... z _k) be two parent chromosomes participating in intersection, successively to gene x _i, (i=1,2 ..., k) and z _i, (i=1,2 ..., k) divide following two class situations to intersect:

1) as gene x _i< ε and z _iduring < ε, if

make (λ _i, λ _j) ∈ l _{λ, c}, x _i=x _j, make x _i=z _i;

2) otherwise, make x _i=min (x _i, z _i);

S325: the determining of mutation algorithm:

At chromosome x=(x ₁x ₂... x _k) in random selection two positions, the genic value between these two positions is resequenced, the gene of other position keeps original value constant;

S326: reverse determining of algorithm:

At chromosome x=(x ₁x ₂... x _k) in choose at random two positions, then genic value between these two positions is reversed;

S327: the end condition of genetic algorithm:

The average fitness value of every generation population

can calculate by (9):

$\stackrel{\&OverBar;}{\mathrm{\&rho;}}=\frac{\underset{j=1}{\overset{\mathrm{pop}\_\mathrm{size}}{\mathrm{\&Sigma;}}}{\mathrm{\&rho;}}_j}{N}---\left(9\right);$

The end condition of genetic algorithm is

0< θ <1;

Step 4: solve the dyeing sequence that represents the optimum dyeing of local link conflict graph, concrete steps are as follows:

Input: local link conflict graph M _{λ, c}=(Λ, l _{λ, c}), adjacency matrix A (M _{λ, c})=(ai _j) _{k * k};

Output: the dyeing sequence that represents the optimum dyeing of local link conflict graph;

The initialization of Step1 colony, arranges control parameter, and population size is N, and maximum evolutionary generation is gen, and crossover probability is γ _c, variation probability is _{γ m};

Step2 greedy algorithm:

Use integer string encoding mode, adopt greedy algorithm to obtain N local suboptimal solution, make colony's initialization, concrete greedy algorithm is as follows:

If A1 local link conflict graph M _{λ, c}=(Λ, l _{λ, c}) in the summit of being unstained in addition, execution step A2, otherwise greedy algorithm finishes, and carries out genetic algorithm;

A2 is arbitrary summit in arbitrary Maximum Clique Q, finds the summit with the shortest another Maximum Clique of its topology distance P, dyes identical color T, with form can be concurrent link set;

A3 continue to find with steps A 2 in Maximum Clique Q topology distance the shortest and and this Maximum Clique Q in disjunct summit all, all summits, this summit is dyed and color T identical in steps A 2, and be added to can be concurrent link set;

A4 repeating step A3, if cannot find again qualified summit be added to can be concurrent link set, execution step A5;

A5 is from local link conflict graph M _{λ, c}=(Λ, l _{λ, c}) in leave out the aforementioned limit of having dyeed summit and being connected with aforementioned summit, then perform step A1;

Step3 genetic algorithm:

B1gen=0, generates initial population Pop (0), has N bar chromosome in initial population Pop (0);

B2 utilizes fitness function (7) to calculate gen for the chromosomal fitness value of N bar in population;

If B3 meets the end condition of algorithm genetic algorithm, output meets the chromosome of algorithm end condition, and the gene order on this chromosome is the dyeing sequence that represents the optimum dyeing of local link conflict graph; Otherwise carry out B4;

B4 adopts the definite heredity of formula (8) to select probability γ _ithe N bar chromosome copying in initial population obtains new N bar chromosome;

B5 to new N bar chromosome according to crossover probability γ _cselect in pairs N _cbar chromosome, the definite crossover algorithm of applying step S234 intersects between two paired dyeing; Again according to variation probability γ _mselect N _rbar chromosome, the definite mutation algorithm of applying step S325 makes a variation; And then the random N selecting _nbar chromosome, the reverse algorithm in applying step S326 reverses, and obtains population of future generation, establishes gen=gen+1, returns to B3;

Step 5: based on step 4, obtain representing the non-competing MAC agreement of the optimum dyeing sequence foundation day blind ultraviolet non line of sight Ad-hoc communication network dyeing of local link conflict graph, specific as follows:

The node that dyes same color in day blind ultraviolet non line of sight Ad-hoc communication network is distributed in to same time slot, and the node that dyes different colours is distributed in different time-gap, and the number of time slot equals with dyeing number;

In described non-competing MAC agreement, several time slots form a frame, each frame is divided into controls subframe and data burst, starting of each frame is to control subframe, utilize control subframe that each node transport service is assigned to each data slot in data burst, all nodes sharing within the scope of collision domain are controlled scheduling and the data slot assignment information of subframes.

With respect to prior art, tool of the present invention has the following advantages:

1, the local link conflict graph for real-time structure dyes to it, is applicable to the dynamic distributed feature of day blind ultraviolet non line of sight Ad-hoc communication network, the requirement becoming in the time of can meeting network topology.

2, greedy algorithm, in conjunction with genetic algorithm to the dyeing of local link conflict graph, can obtain near-optimization coloration result fast, thereby is applicable to the sharing wireless channel distribution of day blind ultraviolet non line of sight Ad-hoc communication network of topological dynamic change.

The point chromatic number of the local link conflict graph 3, obtaining based on method provided by the invention is less, that is less to the timeslot number of unit interval division, also met a day blind ultraviolet non line of sight Ad-hoc communication network is simultaneously the target that does not interfere with each other channel that conflicting link assignment is different with minimum timeslot number within the unit interval.

4, there is no at present for day blind ultraviolet non line of sight Ad-hoc communication network and propose the MAC agreement that conflict is avoided, the present invention proposes to be first applicable to the conflict of day blind ultraviolet non line of sight Ad-hoc communication network and avoids MAC agreement---UVAd-TDMA agreement.Be showed no relevant report both at home and abroad.

Accompanying drawing explanation

Tu1Wei omnidirectional---the conflict model under directional communication mode.

Tu2Wei omnidirectional---the conflict model under beam communication mode.

Fig. 3 is directed---the conflict model of beam communication mode.

Fig. 4 is a day blind ultraviolet non line of sight Ad-hoc communication network topology figure.

Fig. 5 a is the link conflict graph that Fig. 4 is corresponding, and Fig. 5 b is the Maximum Clique that Fig. 5 a is corresponding.

Fig. 6 a is the optimum dyeing of link conflict graph, and Fig. 6 b is that the time slot based on Fig. 6 a link conflict graph is divided.

Fig. 7 a is the link conflict graph of 10 nodes, and Fig. 7 b is the Color figure of Fig. 7 a link conflict graph.

Fig. 8 a is the link conflict graph of 16 nodes, and Fig. 8 b is the Color figure of Fig. 8 a link conflict graph.

Fig. 9 a is the link conflict graph of 21 nodes, and Fig. 9 b is the Color figure of Fig. 9 a link conflict graph.

Figure 10 is the frame structure under UVAd-TDMA agreement.

Figure 11 is each node time slot allocation and structure of time slot figure.

Figure 12 is UVAd-TDMA agreement and the comparison of IEEE802.11 throughput of system.

Figure 13 is UVAd-TDMA agreement and the comparison of IEEE802.11 end-to-end delay.

Embodiment

Below in conjunction with drawings and Examples, technology of the present invention is described in further details.

Day blind ultraviolet non line of sight Ad-hoc communication network shared channel optimization method, comprises the steps:

Step 1: model is avoided in the conflict of having set up under day three kinds of channelling modes of blind ultraviolet non line of sight Ad-hoc communication network:

The conflict of S11: omnidirectional---directional communication mode avoids model suc as formula (1):

|N1-N2|≥(1+Δ)(r ₁+r ₂) （1）；

Wherein, Δ >0, represents that boundary belt region, boundary belt region can be determined by real network, and N1 and N2 all represent the node of day blind ultraviolet non line of sight Ad-hoc communication network, r ₁and r ₂represent respectively node N1, effective propagation path during N2 omnidirectional transfer of data.

Referring to Fig. 1, node N1 communicates by letter with node A, and node N2 communicates by letter with Node B, and conflict area is the effective propagation path r during with node N1 omnidirectional transfer of data ₁effective propagation path r during for the drawn circle of radius with node N2 omnidirectional transfer of data ₂public part for the drawn circle of radius.For fear of conflict, the distance between node N1 and N2 should be at least (1+ Δ) (r ₁+ r ₂), wherein Δ >0 is boundary belt region, with transmission when avoiding adjacent node on identical ultraviolet light channel; Nature, conflict area can be defined as the region that sender can cover, therefore | N1-N2|>=(1+ Δ) (r ₁+ r ₂) be the beeline that can avoid conflict.Being omnidirectional---the conflict of directional communication mode avoids model suc as formula (1):

|N1-N2|≥(1+Δ)(r ₁+r ₂) （1）。

S12: omnidirectional---conflict under beam communication mode avoids model suc as formula (2):

|N1-N2|≥(1+Δ)(r ₁'+r ₂) （2）；

Wherein, Δ >0, represents boundary belt region, and N1 and N2 all represent the node of day blind ultraviolet non line of sight Ad-hoc communication network, r ₁' the effective propagation path of expression node N1 when directional data transmits, r ₂effective propagation path while representing node N2 omnidirectional transfer of data.

Referring to Fig. 2, node N1 communicates by letter with node A, and node N2 communicates by letter with Node B, and conflict area can use formula (a) to represent:

S _cDEF≈ S _{fan N1CD}-S _{Δ N1EF}(a);

| N1-N2|>=(1+ Δ) (r ₁'+r ₂) be the beeline that can avoid conflict.Omnidirectional---conflict under beam communication mode avoids model suc as formula (2):

|N1-N2|≥(1+Δ)(r ₁'+r ₂) （2）。

S13: orientation---conflict under beam communication mode avoids model suc as formula (3):

|N1-N2|≥(1+Δ)(r ₁'+r ₂') （3）；

Wherein, Δ >0, represents boundary belt region, and N1 and N2 all represent the node of day blind ultraviolet non line of sight Ad-hoc communication network, r ₁' the effective propagation path of expression node N1 when directional data transmits, r ₂' the effective propagation path of expression node N2 when directional data transmits.

Referring to Fig. 3, node N1 communicates by letter with node A, and node N2 communicates by letter with Node B, and interference region is irregular figure CDEF, | N1-N2|>=(1+ Δ) (r ₁'+r ₂') be the beeline that can avoid conflict.Directed---conflict under beam communication mode avoids model suc as formula (3):

|N1-N2|≥(1+Δ)(r ₁'+r ₂') （3）。

Step 2: build the local link conflict graph of day blind ultraviolet non line of sight Ad-hoc communication network, concrete steps are as follows:

S21: structure link interference approximate model:

Every link neighbours' link in only jumping with self ξ exists and disturbs,

r=min{r wherein _i, r _i' represent the transmission radius of node, R'=max{r _i, r _i' represent the interference radius of node;

S22: utilize the computing capability of node in day blind ultraviolet non line of sight Ad-hoc communication network to calculate, each node produces the broadcast data packet that includes own business transmission and suffered interfere information, with omnidirectional's light beam, this broadcast data packet is sent to the neighbours' link nodes in its ξ jumping;

S23: the ξ that each node is preserved it jumps the broadcast packet that interior nodes sends, and generate the local link conflict graph of day blind ultraviolet non line of sight Ad-hoc communication network.Owing to there is not interference in the link of ξ outside jumping and its, so generate this local link conflict graph with enough to its channel allocation.Local link conflict graph has reflected the conflict relationship of all data flow in day blind ultraviolet non line of sight Ad-hoc network.

If one on day blind ultraviolet non line of sight Ad-hoc communication network link the transmitting terminal of data flow or receiving terminal on an other link in the transmitting terminal of flow or the transmission range of receiving terminal, and two transmitting terminals or receiving terminal send simultaneously or receive data, and this two day blind ultraviolet non line of sight Ad-hoc communication network link is called conflict link.Data flow on conflict link can not be simultaneously concurrent.

Conflict graph (CI, collision image), refers to the conflict graph of avoiding link under the constraint of model three kinds of conflicts, by l _ibar link and be subject to l _iits some other time set of blind ultraviolet non line of sight Ad-hoc communication network link of bar link transmission impact.Centering local link conflict graph M of the present invention _{λ, c}=(Λ, l _{λ, c}), Λ={ λ wherein ₁, λ ₂..., λ _kthe set of data flows that in day blind ultraviolet non line of sight Ad-hoc communication network, institute's active node sends, limit λ _iλ _j∈ l _{λ, c}and if only if data flow λ _iand λ _jconflict.

Obviously, link conflict graph is undirected simple graph.Link in link conflict graph in vertex representation day blind ultraviolet non line of sight Ad-hoc communication network if exist and disturb between link, connects a limit between the respective vertices of link conflict graph.Link conflict graph has represented the conflict situations that may exist between all links in day blind ultraviolet non line of sight Ad-hoc communication network.

In classic map opinion, the vertex coloring rule of figure is: adjacent vertex coloring is different; The link scheduling rule of link conflict graph model is: adjacent vertex can not be simultaneously concurrent.As can be seen here, the scheduling rule of link conflict graph is identical with figure dyeing rule, painted identical point represents that corresponding day blind ultraviolet non line of sight Ad-hoc communication link can be simultaneously in running order, and the corresponding day blind ultraviolet non line of sight Ad-hoc communication link of painted different vertex representation can not be simultaneously in running order.In day blind ultraviolet non line of sight Ad-hoc communication network, the time is divided into isometric working time slot, corresponding day blind ultraviolet non line of sight Ad-hoc communication link of painted identical point can be operated in a time slot, i.e. a time slot in the corresponding ultraviolet communication TDMA scheduling of a kind of color.For maximization network throughput, guaranteeing under the prerequisite of required communication task, should minimize the communication scheduling cycle of whole network.Therefore, require a day blind ultraviolet non line of sight Ad-hoc communication TDMA dispatching algorithm should complete with minimum timeslot number the transmission of business in its network.It is that link conflict graph dyes that a minimum timeslot number of therefore, trying to achieve day blind ultraviolet non line of sight Ad-hoc communication network resource scheduling is equivalent to by minimum chromatic number.The dispatch channel resource problem of day blind ultraviolet non line of sight Ad-hoc communications network media access control is converted to the vertex coloring problem of link conflict graph dexterously like this.

If Fig. 4 is a day blind ultraviolet non line of sight Ad-hoc communication network topology figure with 7 nodes, establishes source node 1 and send data traffic λ ₁send data traffic λ with source node 3 ₃time there is conflict, source node 4 sends data traffic λ ₄send data traffic λ with source node 5 ₅time there is conflict, the corresponding link conflict graph of Fig. 4 is as shown in Figure 5.

In classic map opinion, if any two summits are all adjacent in the subgraph of a certain simple non-directed graph, claim that this subgraph is the group of former figure.If but increase arbitrarily behind new summit in this subgraph, it just no longer becomes group, claims that this subgraph is the Maximum Clique of former figure.In a figure, the Maximum Clique that summit number is maximum is called Clique.Obviously, each summit in the same group of link conflict graph (link in corresponding day blind ultraviolet non line of sight Ad-hoc communication network topology) can not transmit data simultaneously, and when different, belonging to can ensuring escapement from confliction concurrent data between the summit of any one group.

As shown in Figure 5, Fig. 5 a is a link conflict graph, and Fig. 5 b shows in this link conflict graph and comprised 2 Maximum Cliques, and each link in same Maximum Clique can not send simultaneously or receive data due to the mutual conflict on space.

First the present invention carries out vertex coloring to day blind ultraviolet non line of sight Ad-hoc communication network local link conflict graph, seeks its chromatic number; Then according to a chromatic number value, a certain unit interval length is divided into isometric time slot, the corresponding ultraviolet light link in summit of same color can be operated in a time slot, be a time slot in the blind ultraviolet non line of sight Ad-hoc communication network working node scheduling in corresponding day of a kind of color, as shown in Figure 6.In Fig. 6 a, use 2 kinds of colors to realize the optimum dyeing of link conflict graph, be equivalent in Fig. 6 b and use 2 time slots to complete the transmission of business in network.

Because day blind ultraviolet non line of sight Ad-hoc communication network is a distributed network system, if adopt centralized approach structure link conflict graph in such distributed system, first topology information and the business need information of whole day blind ultraviolet non line of sight Ad-hoc communication network all must be converged to a node that computing capability is powerful, to produce complete link conflict graph, then You Yige computer center processor calculates, finally result (scheduling scheme) is distributed to each day blind ultraviolet non line of sight Ad-hoc communication network node, obviously this is impossible realize in day blind ultraviolet non line of sight Ad-hoc communication network.In real day blind ultraviolet non line of sight Ad-hoc communication network scene, practical link conflict graph building method should be walk abreast, distributed, only rely on local message, can in the short period of time, obtain one and approach optimum link conflict graph.

Step 3: the local link conflict graph of day blind ultraviolet non line of sight Ad-hoc communication network of setting up according to step 2 calculates optimum dyeing sequence, specific as follows:

S31: the local link conflict graph of establishing day blind ultraviolet non line of sight Ad-hoc communication network of step 2 foundation is M _{λ, c}=(Λ, l _{λ, c}), the μ-Vertex Coloring of this local link conflict graph is from local link conflict graph M _{λ, c}=(Λ, l _{λ, c}) vertex set Λ to a mappings of chromatic number collection PC (μ); And if only if λ _i, λ _j∈ Λ, and λ _iλ _j∈ l _{λ, c}time, σ (λ _i) ≠ σ (λ _j), local link conflict graph M _{λ, c}=(Λ, l _{λ, c}) the set of μ-Vertex Coloring be designated as Ξ _μ(M _{λ, c}), if | Ξ _μ(M _{λ, c}) | ≠ 0, claim local link conflict graph M _{λ, c}=(Λ, l _{λ, c}) be μ-stainable, wherein, λ _i, λ _jrepresent respectively link conflict graph M _{λ, c}=(Λ, l _{λ, c}) vertex set Λ in i and j summit, l _{λ, c}represent conflict link set, σ (λ _i) expression λ _idye number, σ (λ _j) expression λ _jdye number, chromatic number collection PC (μ) represents the set that dye number forms.

Without loss of generality, the operating time of day blind ultraviolet non line of sight Ad-hoc communication network is set as to the unit interval, the unit interval can also can be 1 hour for one day 24 hours, can set according to actual needs.

S32: adopt genetic algorithm to combine and carry out the dyeing of link conflict graph with greedy algorithm, concrete steps are as follows:

S321: build chromosome space:

With A (M _{λ, c})=(ai _j) _{k * k}represent local link conflict graph M _{λ, c}=(Λ, l _{λ, c}) adjacency matrix, specifically suc as formula (4):

$a_{\mathrm{<figure-callout\; id="1"\; label="ij"\; filenames="HDA0000450489840000012.png,HDA0000450489840000021.png"\; state="\{\{state\}\}">ij</figure-callout>}}\left\{\begin{array}{ccc}1& \mathrm{if}& {\mathrm{\&lambda;}}_i{\mathrm{\&lambda;}}_j\&Element;l_{\mathrm{\&lambda;},\mathrm{<figure-callout\; id="0"\; label="c"\; filenames="HDA0000450489840000052.png"\; state="\{\{state\}\}">c</figure-callout>}}\\ 0& \mathrm{if}& {\mathrm{\&lambda;}}_i{\mathrm{\&lambda;}}_j\&NotElement;l_{\mathrm{\&lambda;},c}\end{array}\right.---\left(4\right);$

Wherein k represents link conflict graph M _{λ, c}=(Λ, l _{λ, c}) exponent number, a _ijrepresent adjacency matrix A (M _{λ, c})=(ai _j) _{k * k}in element, to local link conflict graph M _{λ, c}=(Λ, l _{λ, c}) all dyeing forecast scheme configuration chromosomes space on summit;

Following coding is made in chromosome space:

If M _{λ, c}=(Λ, l _{λ, c}) a kind of order on k summit be n ₁, n ₂... n _k, n wherein ₁, n ₂... n _kbe natural 1,2 ..., a kind of arrangement of k}, the sequence x that is k corresponding to a length ₁x ₂... x _i... x _k, x wherein _iexpression is to M _{λ, c}=(Λ, l _{λ, c}) summit λ _icolor, x _i∈ PC (μ); As coded sequence 124... μ represents sequence number n ₁the corresponding diagram M of institute _{λ, c}=(Λ, l _{λ, c}) summit λ ₁color 1, sequence number n ₂the corresponding diagram M of institute _{λ, c}=(Λ, l _{λ, c}) summit λ ₂color 2, sequence number n ₃the corresponding diagram M of institute _{λ, c}=(Λ, l _{λ, c}) summit λ ₃color 4 ..., sequence number n _kthe corresponding diagram M of institute _{λ, c}=(Λ, l _{λ, c}) summit λ _kcolor μ.

After being encoded in chromosome space, the initial population that the N bar chromosome that utilizes greedy algorithm generation to be comprised of integer forms, the sequence x that is k corresponding to a length ₁x ₂... x _i... x _kbecome item chromosome x, xi becomes this chromosomal i gene;

S322: set fitness function:

Definition penalty function is suc as formula (5):

$p\left(x\right)=\underset{i=1.j=i+1}{\overset{k}{\mathrm{\&Sigma;}}}\mathrm{\&eta;}({\mathrm{\&lambda;}}_i,{\mathrm{\&lambda;}}_j)---\left(5\right);$

And

η (x) represents that in chromosome space, chromosome x violates the number on the limit (limit that has conflict) of constraint;

Definition fitness function is suc as formula (7):

$\mathrm{\&rho;}(x_1{x}_2...x_k)=\frac{1}{p\left(x\right)+|\mathrm{\&chi;}\left(x\right)-\mathrm{\&epsiv;}-1|}---\left(7\right);$

Wherein χ (x) is chromosome x=(x ₁x ₂... x _k) in number of colours used, ε is color minimum value used in the μ that tried to achieve-can dyeing scheme;

S333: probability γ is selected in heredity _ithe selection of genetic operator determine:

Probability γ is selected in heredity _iemploying roulette is selected, and the selecteed chance of chromosome that fitness is high is large.For fitness value, be ρ _ii bar chromosome x, probability γ is selected in its heredity _iavailable formula (8) calculates:

${\mathrm{\&gamma;}}_i=\frac{{\mathrm{\&rho;}}_i}{\underset{j=1}{\overset{\mathrm{pop}\_\mathrm{size}}{\mathrm{\&Sigma;}}}{\mathrm{\&rho;}}_j}---\left(8\right);$

The value of the upper limit of wherein suing for peace pop_size is population scale N;

S324: the determining of crossover algorithm:

If x=is (x ₁x ₂..x _i..x _k) and z=(z ₁z ₂... z _i... z _k) be two parent individualities participating in intersection, successively to gene x _i, (i=1,2 ..., k) and z _i, (i=1,2 ..., k) divide following two class situations to intersect:

1) as gene x _i< ε and z _iduring < ε, if

make (λ _i, λ _j) ∈ l _{λ, c}, x _i=x _j, make x _i=z _i;

2) otherwise, make x _i=min (x _i, z _i);

Above-mentioned crossover algorithm, 1) interleaved mode can reduce the blind search of genetic algorithm, 2) interleaved mode, individual number of colours used is reduced gradually, the search volume of problem is dwindled.Under the effect of above-mentioned crossover algorithm, effectively summit carve information is retained, and can produce again some new individualities.Crossover algorithm can reduce intersection offspring color μ used, and (μ >=ε), and fitness function can guarantee that color used is that the chromosome that plant (ε-1) is survived in population of future generation with greater probability.

S325: the determining of mutation algorithm:

At chromosome x=(x ₁x ₂... x _k) in random selection two positions, the genic value between these two positions is resequenced, the gene of other position keeps original value constant; For example:

Father: 323|123|465

Son: 323|123|645

S326: reverse determining of algorithm:

At chromosome x=(x ₁x ₂... x _k) in choose at random two positions, then genic value between these two positions is reversed; For example:

Father: 323|123|465

Son: 323|321|465

Above-mentioned mutation algorithm and reverse algorithm are conducive to keep the diversity of population, can effectively make the unlikely local optimum that is absorbed in of algorithm.

S327: the end condition of genetic algorithm:

The average fitness value of every generation population

can calculate by (9):

$\stackrel{\&OverBar;}{\mathrm{\&rho;}}=\frac{\underset{j=1}{\overset{\mathrm{pop}\_\mathrm{size}}{\mathrm{\&Sigma;}}}{\mathrm{\&rho;}}_j}{N}---\left(9\right);$

The end condition of genetic algorithm is

0< θ <1;

Step 4: solve the dyeing sequence that represents the optimum dyeing of local link conflict graph, concrete steps are as follows:

Input: local link conflict graph M _{λ, c}=(Λ, l _{λ, c}), adjacency matrix A (M _{λ, c})=(ai _j) _{k * k};

Output: the dyeing sequence that represents the optimum dyeing of local link conflict graph;

The initialization of Step1 colony, arranges control parameter, and population size is N, and maximum evolutionary generation is gen, and crossover probability is γ _c, variation probability is γ _m;

Step2 greedy algorithm:

Use integer string encoding mode, adopt greedy algorithm to obtain N local suboptimal solution, make colony's initialization, concrete greedy algorithm is as follows:

If A1 local link conflict graph M _{λ, c}=(Λ, l _{λ, c}) in the summit of being unstained in addition, execution step A2, otherwise greedy algorithm finishes, and carries out genetic algorithm;

A2 is arbitrary summit in arbitrary Maximum Clique Q, finds the summit with the shortest another Maximum Clique of its topology distance P, dyes identical color T, with form can be concurrent link set;

A3 continue to find with steps A 2 in Maximum Clique Q topology distance the shortest and and this Maximum Clique Q in disjunct summit all, all summits, this summit is dyed and color T identical in steps A 2, and be added to can be concurrent link set;

A4 repeating step A3, if cannot find again qualified summit be added to can be concurrent link set, execution step A5;

A5 is from local link conflict graph M _{λ, c}=(Λ, l _{λ, c}) in leave out the aforementioned limit of having dyeed summit and being connected with aforementioned summit, then perform step A1;

Step3 genetic algorithm:

B1gen=0, generates initial population Pop (0), has N bar chromosome in initial population Pop (0);

B2 utilizes fitness function (7) to calculate gen for the chromosomal fitness value of N bar in population;

If B3 meets the end condition of algorithm genetic algorithm, output meets the chromosome of algorithm end condition, and the gene order on this chromosome is the dyeing sequence that represents the optimum dyeing of local link conflict graph; Otherwise carry out B4;

B4 adopts the definite heredity of formula (8) to select probability γ _ithe N bar chromosome copying in initial population obtains new N bar chromosome;

B5 to new N bar chromosome according to crossover probability γ _cselect in pairs N _cbar chromosome, the definite crossover algorithm of applying step S234 intersects (selected multipair chromosome, between two chromosomes between every pair of chromosome, adopted crossover algorithm to intersect); Again according to variation probability γ _mselect N _rbar chromosome, the definite mutation algorithm of applying step S325 make a variation (adopting respectively mutation algorithm to make a variation to every chromosome selecting); And then the random N selecting _nbar chromosome, the reverse Operators Algorithm in applying step S326 is to the random N selecting in population _nbar chromosome reverses (to every chromosome selecting, adopting respectively reverse algorithm to reverse), obtains population of future generation, establishes gen=gen+1, returns to B3;

Step 5: based on step 4, obtain representing the non-competing MAC agreement of the optimum dyeing sequence foundation day blind ultraviolet non line of sight Ad-hoc communication network dyeing of local link conflict graph, specific as follows:

The node that dyes same color in day blind ultraviolet non line of sight Ad-hoc communication network is distributed in to same time slot, and the node that dyes different colours is distributed in different time-gap, and the number of time slot equals the number that dyes;

In described non-competing MAC agreement, several time slots form a frame, frame structure as shown in figure 10, each frame is divided into controls subframe and data burst, starting of each frame is to control subframe, utilize control subframe that each node transport service is assigned to each data slot in data burst, all nodes sharing within the scope of collision domain are controlled scheduling and the data slot assignment information of subframes.

Time slot in data burst, each node is according to controlling the time slot transceiving data of stipulating in cooperative scheduling table in subframe.At each data slot, can there is multipair node to receive and dispatch simultaneously.Data burst consists of a plurality of data slots, and each node carries out the distribution of data slot according to the cooperative scheduling of controlling subframe, at each data slot, can have multipair node to carry out transceiving data simultaneously.

Referring to Figure 11, data slot is assigned to respectively one or more nodes and sends or receive data.Each data slot is comprised of front protective time slot, transmitting time slot, data receiver/distribution period and rear protective time slot.Transmitting time slot is for forwarding from information other node reception, that need overall shared information, data receiver/distribution period to be used for receiving the information of neighbor node transmission or to distribute local generation.

For strengthening the practicality of day blind ultraviolet non line of sight Ad-hoc communication network, must consider that, in it applies especially Military Application, in network, some informational needs of each node generation diffuses to the whole network, all nodes sharing of confession.Although the directional beam in day blind ultraviolet non line of sight Ad-hoc communication network has larger network throughput, yet than omnidirectional's light beam, for the diffusion of information of certain node to the whole network being needed to long many times of cost.In fact, a lot of military informations all have ageing in communication process, and the propagation time is long, and the military information that causes destination node finally to receive was lost efficacy.For this reason, in UVAd-TDMA agreement, taked the mode incidentally forwarding, the transmitting time slot (seeing Figure 11) at each time slot incidentally forwards " local information " that paid-in other node needs the whole network to share, and realizes the efficient broadcast that the overall situation is shared information.

Embodiment 1: at Pentium (R) Dual-Core CPU, and E52002.50GHz1.21GHz, under the hardware environment of 2.00GB, to thering are 10 nodes, 16 nodes, the link conflict graph of 21 nodes dyes, and obtains following result:

By Fig. 7, Fig. 8, Fig. 9 and table 1, can be found out, 2 kinds of colors have only been used in optimum dyeing to the link conflict graph of 10 nodes, 21 nodes have only been used 3 kinds of colors, and the optimum dyeing of the link conflict graph of 16 nodes has also been used to 4 kinds of colors, this is that therefore chromatic number used increases relatively because the limit of link conflict graph of 16 nodes and the relation on summit are more complicated than the limit of link conflict graph of 21 nodes and the relation on summit.But on running time, 16 node 48.59420s, 21 node 98.9220s, in greedy algorithm operation when its spent time is mainly used in colony's initialization.

Table 1

Link conflict graph nodes	Chromatic number	Algebraically	Running time (s)
				10	2	2	1.375020
16	4	6	48.59402
				21	3	7	98.92200

Embodiment 2: adopting NS2 simulated scenario is that 10 nodes are randomly dispersed in the region of 1000m*1000m, and the range of transmission of node is 200 meters, and node motion is assumed to random walk.The major parameter of simulation is as shown in table 2.

Table 2

Parameter	Value
		Nodes
	10
		Emulation scope	1000m×1000m
Data package size	216bit
		Simulated time	1000s
Physical layer rate	1028b/s
		Node maximum speed of motion	20m/s

Figure 12 has compared the throughput of system of UVAd-TDMA agreement and IEEE802.11.Under identical load, the throughput of system of UVAd-TDMA agreement is significantly improved than IEEE802.11, and improving reason is that UVAd-TDMA agreement utilizes shared channel optimum allocation algorithm and the oriented light beam of ultraviolet to increase reusing of time slot, has improved channel utilization.

Figure 13 has compared the end-to-end delay of UVAd-TDMA agreement and IEEE802.11.Competition Characteristics due to IEEE802.11 under identical load makes it have lower delay, and UVAd-TDMA agreement presents higher delay owing to moving shared channel optimum allocation algorithm, but still at Internet Transmission tolerance interval (<0.000055s).

Finally explanation is, above embodiment is only unrestricted in order to technical scheme of the present invention to be described, although the present invention is had been described in detail with reference to preferred embodiment, the those of ordinary skill of this neighborhood is to be understood that, can modify or be equal to replacement technical scheme of the present invention, and not departing from aim and the scope of technical solution of the present invention, it all should be encompassed in the middle of claim scope of the present invention.

Claims (1)

1. day blind ultraviolet non line of sight Ad-hoc communication network shared channel optimization method, is characterized in that, comprises the steps:

Step 1: model is avoided in the conflict of having set up under day three kinds of channelling modes of blind ultraviolet non line of sight Ad-hoc communication network:

The conflict of S11: omnidirectional---directional communication mode avoids model suc as formula (1):

|N1-N2|≥(1+Δ)(r ₁+r ₂) （1）；

Wherein, Δ >0, represents boundary belt region, and N1 and N2 all represent the node of day blind ultraviolet non line of sight Ad-hoc communication network, r ₁and r ₂represent respectively node N1, effective propagation path during N2 omnidirectional transfer of data;

S12: omnidirectional---conflict under beam communication mode avoids model suc as formula (2):

|N1-N2|≥(1+Δ)(r ₁'+r ₂) （2）；

Wherein, Δ >0, represents boundary belt region, and N1 and N2 all represent the node of day blind ultraviolet non line of sight Ad-hoc communication network, r ₁' the effective propagation path of expression node N1 when directional data transmits, r ₂effective propagation path while representing node N2 omnidirectional transfer of data;

S13: orientation---conflict under beam communication mode avoids model suc as formula (3):

|N1-N2|≥(1+Δ)(r ₁'+r ₂') （3）；

Wherein, Δ >0, represents boundary belt region, and N1 and N2 all represent the node of day blind ultraviolet non line of sight Ad-hoc communication network, r ₁' the effective propagation path of expression node N1 when directional data transmits, r ₂' the effective propagation path of expression node N2 when directional data transmits;

Step 2: build the local link conflict graph of day blind ultraviolet non line of sight Ad-hoc communication network, concrete steps are as follows:

S21: structure link interference approximate model:

Every link neighbours' link in only jumping with self ξ exists and disturbs,

r=min{r wherein _i, r _i' represent the transmission radius of node, R'=max{r _i, r _i' represent the interference radius of node;

S22: utilize the computing capability of node in day blind ultraviolet non line of sight Ad-hoc communication network to calculate, each node produces the broadcast data packet that includes own business transmission and suffered interfere information, with omnidirectional's light beam, this broadcast data packet is sent to the neighbours' link nodes in its ξ jumping;

S23: the ξ that each node is preserved it jumps the broadcast packet that interior nodes sends, and generate the local link conflict graph of day blind ultraviolet non line of sight Ad-hoc communication network;

Step 3: the local link conflict graph of day blind ultraviolet non line of sight Ad-hoc communication network of setting up according to step 2 calculates optimum dyeing sequence, specific as follows:

S31: the local link conflict graph of establishing day blind ultraviolet non line of sight Ad-hoc communication network of step 2 foundation is M _{λ, c}=(Λ, l _{λ, c}), the μ-Vertex Coloring of this local link conflict graph is from local link conflict graph M _{λ, c}=(Λ, l _{λ, c}) vertex set Λ to a mappings of chromatic number collection PC (μ); And if only if λ _i, λ _j∈ Λ, and λ _iλ _j∈ l _{λ, c}time, σ (λ _i) ≠ σ (λ _j), local link conflict graph M _{λ, c}=(Λ, l _{λ, c}) the set of μ-Vertex Coloring be designated as Ξ _μ(M _{λ, c}), if | Ξ _μ(M _{λ, c}) | ≠ 0, claim local link conflict graph M _{λ, c}=(Λ, l _{λ, c}) be μ-stainable, wherein, λ _i, λ _jrepresent respectively link conflict graph M _{λ, c}=(Λ, l _{λ, c}) vertex set Λ in i and j summit, l _{λ, c}represent conflict link set, σ (λ _i) expression λ _idye number, σ (λ _j) expression λ _jdye number, chromatic number collection PC (μ) represents the set that dye number forms;

S32: adopt genetic algorithm to combine and carry out the dyeing of link conflict graph with greedy algorithm, concrete steps are as follows:

S321: build chromosome space:

With A (M _{λ, c})=(ai _j) _{k * k}represent local link conflict graph M _{λ, c}=(Λ, l _{λ, c}) adjacency matrix, specifically suc as formula (4):

$a_{\mathrm{ij}}\left\{\begin{array}{ccc}1& \mathrm{if}& {\mathrm{\&lambda;}}_i{\mathrm{\&lambda;}}_j\&Element;l_{\mathrm{\&lambda;},c}\\ 0& \mathrm{if}& {\mathrm{\&lambda;}}_i{\mathrm{\&lambda;}}_j\&NotElement;l_{\mathrm{\&lambda;},c}\end{array}\right.---\left(4\right);$

Wherein k represents link conflict graph M _{λ, c}=(Λ, l _{λ, c}) exponent number, a _ijrepresent adjacency matrix A (M _{λ, c})=(ai _j) _{k * k}in element, to local link conflict graph M _{λ, c}=(Λ, l _{λ, c}) all dyeing forecast scheme configuration chromosomes space on summit;

Following coding is made in chromosome space:

If M _{λ, c}=(Λ, l _{λ, c}) a kind of order on k summit be n ₁, n ₂... n _k, n wherein ₁, n ₂... n _kbe natural 1,2 ..., a kind of arrangement of k}, the sequence x that is k corresponding to a length ₁x ₂... x _i... x _k, x wherein _iexpression is to M _{λ, c}=(Λ, l _{λ, c}) summit λ _icolor, x _i∈ PC (μ);

After being encoded in chromosome space, the initial population that the N bar chromosome that utilizes greedy algorithm generation to be comprised of integer forms, the sequence x that is k corresponding to a length ₁x ₂... x _i... x _kbecome item chromosome x, x _ibecome this chromosomal i gene;

S322: set fitness function:

Definition penalty function is suc as formula (5):

$p\left(x\right)=\underset{i=1.j=i+1}{\overset{k}{\mathrm{\&Sigma;}}}\mathrm{\&eta;}({\mathrm{\&lambda;}}_i,{\mathrm{\&lambda;}}_j)---\left(5\right);$

And

η (x) represents that in chromosome space, chromosome x violates the number on the limit of constraint;

Definition fitness function is suc as formula (7):

$\mathrm{\&rho;}(x_1{x}_2...x_k)=\frac{1}{p\left(x\right)+|\mathrm{\&chi;}\left(x\right)-\mathrm{\&epsiv;}-1|}---\left(7\right);$

Wherein χ (x) is chromosome x=(x ₁x ₂... x _k) in number of colours used, ε is color minimum value used in the μ that tried to achieve-can dyeing scheme;

S333: probability γ is selected in heredity _idetermine:

For fitness value, be ρ _ii bar chromosome x, probability γ is selected in its heredity _iavailable formula (8) calculates:

${\mathrm{\&gamma;}}_i=\frac{{\mathrm{\&rho;}}_i}{\underset{j=1}{\overset{\mathrm{pop}\_\mathrm{size}}{\mathrm{\&Sigma;}}}{\mathrm{\&rho;}}_j}---\left(8\right);$

The value of the upper limit of wherein suing for peace pop_size is population scale N;

S324: the determining of crossover algorithm:

If x=is (x ₁x ₂..x _i..x _k) and z=(z ₁z ₂... z _i... z _k) be two parent chromosomes participating in intersection, successively to gene x _i, (i=1,2 ..., k) and z _i, (i=1,2 ..., k) divide following two class situations to intersect:

1) as gene x _i< ε and z _iduring < ε, if

make (λ _i, λ _j) ∈ l _{λ, c}, x _i=x _j, make x _i=z _i;

2) otherwise, make x _i=min (x _i, z _i);

S325: the determining of mutation algorithm:

At chromosome x=(x ₁x ₂... x _k) in random selection two positions, the genic value between these two positions is resequenced, the gene of other position keeps original value constant;

S326: reverse determining of algorithm:

At chromosome x=(x ₁x ₂... x _k) in choose at random two positions, then genic value between these two positions is reversed;

S327: the end condition of genetic algorithm:

The average fitness value of every generation population

can calculate by (9):

$\stackrel{\&OverBar;}{\mathrm{\&rho;}}=\frac{\underset{j=1}{\overset{\mathrm{pop}\_\mathrm{size}}{\mathrm{\&Sigma;}}}{\mathrm{\&rho;}}_j}{N}---\left(9\right);$

The end condition of genetic algorithm is

0< θ <1;

Step 4: solve the dyeing sequence that represents the optimum dyeing of local link conflict graph, concrete steps are as follows:

Input: local link conflict graph M _{λ, c}=(Λ, l _{λ, c}), adjacency matrix A (M _{λ, c})=(ai _j) _{k * k};

Output: the dyeing sequence that represents the optimum dyeing of local link conflict graph; The initialization of Step1 colony, arranges control parameter, and population size is N, and maximum evolutionary generation is gen, and crossover probability is γ _c, variation probability is γ _m;

Step2 greedy algorithm:

Use integer string encoding mode, adopt greedy algorithm to obtain N local suboptimal solution, make colony's initialization, concrete greedy algorithm is as follows:

If A1 local link conflict graph M _{λ, c}=(Λ, l _{λ, c}) in the summit of being unstained in addition, execution step A2, otherwise greedy algorithm finishes, and carries out genetic algorithm;

A2 is arbitrary summit in arbitrary Maximum Clique Q, finds the summit with the shortest another Maximum Clique of its topology distance P, dyes identical color T, with form can be concurrent link set;

A3 continue to find with steps A 2 in Maximum Clique Q topology distance the shortest and and this Maximum Clique Q in disjunct summit all, all summits, this summit is dyed and color T identical in steps A 2, and be added to can be concurrent link set;

A4 repeating step A3, if cannot find again qualified summit be added to can be concurrent link set, execution step A5;

A5 is from local link conflict graph M _{λ, c}=(Λ, l _{λ, c}) in leave out the aforementioned limit of having dyeed summit and being connected with aforementioned summit, then perform step A1;

Step3 genetic algorithm:

B1gen=0, generates initial population Pop (0), has N bar chromosome in initial population Pop (0);

B2 utilizes fitness function (7) to calculate gen for the chromosomal fitness value of N bar in population;

If B3 meets the end condition of algorithm genetic algorithm, output meets the chromosome of algorithm end condition, and the gene order on this chromosome is the dyeing sequence that represents the optimum dyeing of local link conflict graph; Otherwise carry out B4;

B4 adopts the definite heredity of formula (8) to select probability γ _ithe N bar chromosome copying in initial population obtains new N bar chromosome;

B5 to new N bar chromosome according to crossover probability γ _cselect in pairs N _cbar chromosome, the definite crossover algorithm of applying step S234 intersects between two paired dyeing; Again according to variation probability γ _mselect N _rbar chromosome, the definite mutation algorithm of applying step S325 makes a variation; And then the random N selecting _nbar chromosome,

Reverse algorithm in applying step S326 reverses, and obtains population of future generation, establishes gen=gen+1, returns to B3;

Step 5: based on step 4, obtain representing the non-competing MAC agreement of the optimum dyeing sequence foundation day blind ultraviolet non line of sight Ad-hoc communication network dyeing of local link conflict graph, specific as follows:

The node that dyes same color in day blind ultraviolet non line of sight Ad-hoc communication network is distributed in to same time slot, and the node that dyes different colours is distributed in different time-gap, and the number of time slot equals with dyeing number;

In described non-competing MAC agreement, several time slots form a frame, each frame is divided into controls subframe and data burst, starting of each frame is to control subframe, utilize control subframe that each node transport service is assigned to each data slot in data burst, all nodes sharing within the scope of collision domain are controlled scheduling and the data slot assignment information of subframes.

Images