CN106254164A - A kind of isomery many media networks system of selection - Google Patents

Abstract

The present invention relates to a kind of isomery many media networks system of selection, described method includes: update net-work parameter information；Set up information MAP table；Judge whether current ink quality deteriorates；Seek network function；Draw optimal communication modes；The present invention utilizes the characteristic of family eugenics (FEBE) Fast Convergent, the network communication parameters of various communication mode can be combined in real time, contrast, select communication mode optimum under real-time status, thus intelligent solution nets the terminal network select permeability that overlap covers more.

Description

A kind of isomery many media networks system of selection

Technical field

The present invention relates to the network selecting method in communication of power system field, a kind of many media networks of isomery System of selection.

Background technology

Improving constantly of the communication of power system technology that the development of power system brings, increasing communication mode It is used on powerline network.The numerous wired and communication being currently present in power communication system, they It is respectively arranged with quality: fiber optic communication has that capacity is big, repeater span length, electromagnetism interference, transmission performance are stable, not by radio frequency The advantage such as restriction, but it is laid and maintenance work has certain limitation；Power line carrier communication utilizes existing joining Electric line transmission is not required to separately spread dedicated communication line, and is easy to management, and cost of investment is low, but message transmission rate is relatively low, easy Be interfered, non-linear distortion and the impact of cross channel modulation, it is difficult to block point by transformer station etc., and when circuit damages Communication is interrupted immediately.The network construction cost of public wireless communication network is low, but real-time communication and reliability can not get ensureing Deng.

And current same network node (electric power website) there will be communication alternate covering in any case, I.e. there is communication from a node (website) to another website, the most just face the problem that a network selects. Currently, there are two kinds of solutions in powerline network, first be exactly in advance when connecting network by artificially to likely Communication mode compare, thus the priority of fixed communication mode.Second is through network selection algorithm Real-time and Dynamic Analyze the communication conditions of heterogeneous networks, thus carry out judging the priority of different communication mode.Clearly first scheme is the most logical Have great advantage for letter quality.

Current network selection algorithm has a lot, including using close to the ordinal number preference methods of ideal scheme, based on combination The network selection algorithm of weight, network selection algorithm based on fuzzy neural network and network selection algorithm based on preferred value etc. Multiple method, but its great majority are all mainly for for wireless domain, and do not meet the requirement of powerline network, and Part is not combined related service situation in power communication with wired algorithm combined and is analyzed, so cause cannot be straight Scoop out use.

In the case, according to the coverage condition of the communication mode in the presence of current powerline network, design one The method being suitable for the selection of powerline network interior joint communication mode is the most necessary；And the realization of the method, can be intelligent Solve the terminal network select permeability that many net overlaps cover.

Summary of the invention

In order to solve the above-mentioned deficiency in the presence of prior art, the present invention provides a kind of isomery many media networks selecting party Method.

Described method comprises the steps:

I, renewal net-work parameter information；

II, set up information MAP table；

III, judge whether current ink quality deteriorates；

IV, seek network utility function；

V, draw optimal communication modes.

Preferably, the renewal of described step I net-work parameter information includes: in network coverage area, terminal node connects in real time The all possible communication mode network information of continuous collection, the information MAP table of described step II includes: according to the network ginseng extracted Number information, heterogeneous networks is corresponding with the performance parameter of each automatic network, form corresponding form.

Preferably, the judgement of described step III link-quality includes: if current ink quality occurs to deteriorate, carry out step IV, otherwise returns step I.

Preferably, described network utility function Z (i) is shown below:

In formula, X_ij: evaluation index sequence, i=1,2 ..., n:i heterogeneous network, V_j(j=1,2,3 ..., p) evaluation refers to The weight that mark is corresponding, j=1 therein, 2,3 ..., p:j respectively adjudicates index.

Preferably, the acquisition of described step V optimal communication modes includes: seek maximum utility value with network utility function Z (i) Max{z (i) }, wherein i network is optimal communication modes.

Preferably, evaluation index sample set normalization described evaluation index sequence X is obtained_ij；

Described evaluation index sample set is shown below:

{x_ij| i=1,2,3...n；J=1,2,3...p} (2)

Described evaluation index sequence X_ijIt is shown below:

Or

In formula, x_jmaxRepresent the maximum in jth judgement index, x_jminRepresent the minima in jth judgement index, x_ijRepresent evaluation index sample, formula (3) is used for the cost type index such as network blocking probability, time delay, to evaluation index sample Collection normalization, uses formula (4) for profit evaluation model indexs such as remaining bandwidth ratio, received signal strengths, returns evaluation index sample set One changes.

Preferably, described evaluation criterion weight V_jConstraints be shown below:

In formula, j=1,2,3 ..., p represents that j respectively adjudicates index.

Preferably, based on family eugenics to V under described constraints_jThe determination of optimal value includes:

(1) initial solution colony U is generated；

(2) adaptive value of viable communication mode is sought；

(3) viable communication mode matched group is set up；

(4) composition family goes forward side by side expert front yard internal competition two-by-two；

(5) adaptive value generation random variation；

(6) judge whether to meet restrictive condition；

(7) V is drawn_jOptimal value.

Preferably, described step (6) judges whether that meeting restrictive condition includes: export result when the condition is satisfied, otherwise Return step (2).

Preferably, described step (1) initial solution colony U is to have i the initial population solved, and with matrix X corresponding thereto Should, wherein initial solution colony U such as following formula represents:

U={1,2 ... i} (6)

Matrix X is shown below:

Described step (4) family uses orthogonal operators that formed family is carried out copulation thus produces the side of lawful issue Formula.

Compared with prior art, the present invention has a following excellent beneficial effect:

The present invention fast convergence by family eugenics, it is the most right to carry out for the communication mode of multiple different media Than selecting, thus the terminal network select permeability that intelligent solution many nets overlap covers；

The present invention uses orthogonal operators that formed family is carried out copulation thus produces the mode of lawful issue, more comprehensively Effectively various communication modes are carried out real time contrast, thus quickly produces the current optimum communication mode accessed.

Accompanying drawing explanation

Fig. 1 is that the many media networks of isomery of the present invention select overall flow figure；

Fig. 2 is the family eugenics method flow diagram of the present invention.

Detailed description of the invention

In order to be more fully understood that the present invention, below in conjunction with Figure of description and example, present disclosure is done further Explanation.

The present invention provides a kind of isomery many media networks system of selection, and described method is by network delay, Netowrk tape The factorial analysis of the aspects such as width, service speed, network packet loss rate and service price processes and carries out optimal communication mode selection Computational methods, described method includes:

1, net-work parameter information is updated:

In the electric power terminal node that various kinds of media accesses, each terminal unit has multiple network interface, can be at network The area of coverage is collected the performance parameter of available network, such as handling capacity, time delay, packet loss, safety and cost etc., i.e. terminal node Real-time continuous collect the network information of likely communication mode.

2, information MAP table is set up:

According to the net-work parameter information extracted, heterogeneous networks is corresponding with the performance parameter of each automatic network, form information Mapping table, as shown in the table:

Table 1

3, judge whether current ink quality deteriorates:

When terminal detects that current ink quality deterioration maybe cannot meet the service quality (Quality of user's new business OfService, QoS) when requiring, terminal will automatically trigger network selection mechanism, utilize alternative network parameter and different business Demand to network, selects optimal network insertion by network selection algorithm, otherwise enters step 1 and updates network parameter letter Breath.

4, utilize family eugenics method that current network parameter is calculated:

Concrete algorithm model is U={1,2 ... i} represents the alternative network scheme set that i heterogeneous network is constituted, often One network needs integration test j to adjudicate index, i.e. judgement index set is V={1, and 2 ... j}.Assume that power communication is eventually End is the multimode terminal having multiple network interface, and can automatically collect the performance parameter obtaining each network.Each terminal The relative priority of each judgement index artificially can be set by equipment by interactive interface.Different types of service also may be used With according to each the relative priority of the requirement judgement index of QoS being set or revises.

In order to consider j judgement index of network, this is judged by this patent by weight, and weight is the biggest then The evaluation index of its correspondence is the most important, therefore correctly the solving of weight, and accesses selection have very for making optimal network Important meaning.

If the sample set of each assessing network index is

{x_ij| i=1,2,3...n；J=1,2,3...p} (1)

It is constituted initially adjudicates matrix X:

For selecting evaluation index to there is multiformity, for making evaluation index variable form consistency metric, it is simple to point Analysis.Need initial data is normalized, cost type (the smaller the better) index such as network blocking probability, time delay is used Formula (3) is normalized, and uses formula (4) for profit evaluation model (being the bigger the better) indexs such as remaining bandwidth ratio, received signal strengths It is normalized.X is obtained after evaluation index sample set normalization_ijSequence:

Weight corresponding to jth evaluation index is set to V_j(j=1,2 ..., p), z (i) (i=1,2 ..., n) it is network effect With function, by evaluation criterion weight and evaluation index sequence X_ijWeighted sum obtains:

By each evaluation index being carried out rational weight distribution V=(V₁, V₂, V₃..., V_p), obtain all alternative nets Value of utility z (i) of network.By find out maximum z (i) (i=1,2 ..., n) carry out network selection, i.e. max{z (i) } corresponding Network i using the first-selection as service access.Here, V=(V₁, V₂, V₃..., V_p), it is a unit vector, V_jRepresent and respectively comment Valency index weights need to meet in [0,1] is interval, constraints such as formula (6):

Go out can be obtained fom the above equation, in order to correctly assess each alternative network performance level, need to find the evaluation index power of optimum Reassignment scheme.

As in figure 2 it is shown, it is family eugenics algorithm basic procedure,

(1) the communication set U={1 that first we can connect according to existing possibility, 2 ... i} forms one and has i solution Initial population, and can be by utilizing matrix X corresponding thereto；

{x_ij| i=1,2,3...n；N=1,2,3...p} (1)

(2) adaptive value of each viable communication mode is calculated；

Adaptive value described here is for evaluating the good and bad degree of individuality, and the biggest individuality of adaptive value is the best, on the contrary adaptive value The least then individuality is the poorest；Individuality is selected by the size according to adaptive value, and to ensure conformability, good individuality has more Chance raises up seed, and makes good characteristic be able to heredity. and therefore, genetic algorithm requires that adaptive value functional value must be nonnegative number.? In the application because select be network objectives as solving colony, so its adaptive value is the normalized value of network parameter, And accomplish the change of single direction, be the most all the bigger the better, if the smaller the better, then to its inverted process.

(3) select probability of each communication mode, probability according to this in corresponding feasible network colony are set up based on adaptive value Distribution, carries out i time randomly selecting to colony, the individuality selected is put people " network pairing pond ".

(4) in feasible network pairing pond every two kinds combine two-by-two and form M2/ family (high adaptive value individuality in each family For male parent, another is maternal)；

(5) to each family, parents produce some " middle children " in the way of orthogonal copulation, competing by household internal Striving, only 2 offsprings grow into follow-on legal person；The i.e. heterogeneous networks parameter of competition in this application is different, and it is showed Performance out is different, so in competition i.e. certain period of time, the comparison of the superiority-inferiority of heterogeneous networks, to various communication mode phases Corresponding performance parameter uses the mode of orthogonal copulation operator.Illustrating concrete mode is:

Set A, B as two parameter factors in communication standard.Wherein factor A has r level, and B has s level, then altogether There is r*s horizontal combination.

Matrix H=(h_ij)_nmIt is a L_n(r₁,r₂,...r_m) type orthogonal table.Ifh_ij∈{0, 1,...,r_j-1}；At h_ij(i=1,2 ..., n) in each different nonnegative integers k (k ∈ 0,1 ..., r_j-1}) occurrence number is n/r_j；By the jth of H₁,j₂In n × 2 submatrix of row composition, every pair of difference Nonnegative integer < k₁,k₂> number of times that occurs is equal to n/ (r_j1,r_j2),k₁∈{0,1,...,r_j1-1},k₂∈{0,1,...,r_j2- 1}, wherein n, m, r₁,r₂…,r_mIt is integer, r_j>=2 (j=1,2,3 ..., m).

The most corresponding orthogonal number of times of n, m and the orthogonal factor number related to, r₁,r₂…,r_mCorrespond respectively to each factor Number of levels.Understood in orthogonal table by definition, for any 2 factors, it is achieved that include the experiment completely of all horizontal combination.Often The number of times that in string, each level occurs is identical, and arranges, to any 2, the horizontal centring constituted, and each level is to repeating phase Same number of times.So orthogonal table can obtain preferable or best result of the test with less test.

(6) adaptive value generation random variation；

Variation is to make a variation according to heterogeneous networks parameter comparative result, and random compared result preferably network is fitted The amendment that should be worth so that it is advantageously in competition, accelerates arithmetic speed.

(7) if having obtained satisfactory solution, the weight that adaptive value after output modifications is corresponding；Otherwise, continue step 3 and enter next Secondary iteration.

Optimum selecting can be carried out according to the current network parameter of communication according to above-mentioned steps, reach real-time shape Under state, the optimization of communication quality.

5, the current optimal communication modes of selection access:

The network insertion select permeability studied is converted to a contained optimization problem, i.e. gives at initial judgement matrix And on the premise of meeting formula (6) constraints, obtain each evaluation criterion weight V by certain many attribute optimizations algorithm_j, then count Calculate and obtain value of utility z (i) of n alternative network, select max{z (i) the network i of correspondence accesses, and draws max{z (i) Middle network i is optimal communication modes.

These are only embodiments of the invention, be not limited to the present invention, all in the spirit and principles in the present invention Within, any modification, equivalent substitution and improvement etc. done, be all contained in applying for the scope of the presently claimed invention that awaits the reply it In.

Claims (10)

1. isomery many media networks system of selection, it is characterised in that described method includes:

I, renewal net-work parameter information；

II, set up information MAP table；

III, judge whether current ink quality deteriorates；

IV, seek network utility function；

V, draw optimal communication modes.

2. isomery many media networks system of selection as claimed in claim 1, it is characterised in that described step I network parameter is believed The renewal of breath includes: in network coverage area, and terminal node real-time continuous collects all possible communication mode network information；

Described step II information MAP table includes: according to the net-work parameter information extracted, by the property of heterogeneous networks Yu each automatic network Energy parameter is corresponding, forms corresponding form.

3. isomery many media networks system of selection as claimed in claim 1, it is characterised in that described step III link-quality Judgement include: if current ink quality occurs to deteriorate, carry out step IV, otherwise return step I.

4. isomery many media networks system of selection as claimed in claim 1, it is characterised in that described step IV network utility letter Number Z (i) is shown below:

$z\left(i\right)={\mathrm{\&Sigma;}}_{j=1}^p{V}_j{X}_{ij},i=1,2,...,n---\left(1\right)$

In formula, X_ij: evaluation index sequence, i=1,2 ..., n:i heterogeneous network, V_j(j=1,2,3 ..., p) evaluation index pair The weight answered, j=1 therein, 2,3 ..., p:j respectively adjudicates index.

5. isomery many media networks system of selection as claimed in claim 1, it is characterised in that described step V optimal communication side The acquisition of formula includes: seek maximum utility value max{z (i) with network utility function Z (i) }, wherein i network is optimal communication side Formula.

6. isomery many media networks system of selection as claimed in claim 4, it is characterised in that by evaluation index sample set normalizing Change and obtain described evaluation index sequence X_ij；

Described evaluation index sample set is shown below:

{x_ij| i=1,2,3...n；J=1,2,3...p} (2)

Described evaluation index sequence X_ijIt is shown below:

$X_{ij}=\frac{x_{jmax}-x_{ij}}{x_{j\max }-x_{jmin}}---\left(3\right)$

Or

In formula, x_jmaxRepresent the maximum in jth judgement index, x_jminRepresent the minima in jth judgement index, x_ijTable Show evaluation index sample.

7. isomery many media networks system of selection as claimed in claim 4, it is characterised in that described evaluation criterion weight V_j's Constraints is shown below:

$\left\{\begin{array}{c}0\&le;V_j\&le;1,\left(j=1,2,\mathrm{...},p\right)\\ {\mathrm{\&Sigma;}}_{j=1}^p{V}_j^2=1\end{array}\right.---\left(5\right)$

In formula, j=1,2,3 ..., p represents that j respectively adjudicates index.

8. isomery many media networks system of selection as claimed in claim 7, it is characterised in that based on family eugenics to described V under constraints_jThe determination of optimal value includes:

(1) initial solution colony U is generated；

(2) adaptive value of viable communication mode is sought；

(3) viable communication mode matched group is set up；

(4) composition family goes forward side by side expert front yard internal competition two-by-two；

(5) adaptive value generation random variation；

(6) judge whether to meet described constraints；

(7) V is drawn_jOptimal value.

9. isomery many media networks system of selection as claimed in claim 8, it is characterised in that described step (6) judges whether Meet restrictive condition to include: enter step (7) when the condition is satisfied, otherwise return step (2).

10. isomery many media networks system of selection as claimed in claim 8, it is characterised in that described step (1) initial solution group Body U is to have i the initial population solved, and with matrix X corresponding thereto, wherein initial solution colony U such as following formula represents:

U={1,2 ... i} (6)

By evaluation index sample set { x_ij| i=1,2,3...n；J=1,2,3...p} constitutes initially judgement matrix X such as following formula institute Show:

$X=\left[\begin{array}{ccc}{x}_{11}& x_{12}...& x_{1i}\\ ...& ...& ...\\ x_{j1}& x_{j2}...& x_{ij}\end{array}\right]---\left(7\right)$

Described step (2) adaptive value is the normalized value of evaluation index；

Described step (3) viable communication mode matched group includes: set up communication party in corresponding feasible network colony based on adaptive value The select probability of formula, according to this probability distribution, carry out i time randomly selecting to colony, and the individuality selected is put into network pairing pond, can In row network pairing pond, every two kinds combine formation matched group two-by-two；

Described step (4) family uses orthogonal mating pattern that formed family is carried out copulation thus produces the side of lawful issue Formula, uses the mode of orthogonal copulation operator to the performance parameter that communication mode is corresponding.