CN107124308A - A kind of tactical data link hierarchical topology method - Google Patents

Abstract

The invention belongs to communication technical field, a kind of tactical data link hierarchical topology method is particularly related to.It is an object of the invention to：For a class tactical data link of non-stop layer node, in the case of node random distribution, based on Apriori data relation analysis algorithm principles, propose a kind of tactical data link hierarchical topology method for digging, the hierarchical relationship between different nodes is excavated, the hierarchical relationship between different nodes in data link is disclosed.

Description

A kind of tactical data link hierarchical topology method

Technical field

The invention belongs to communication technical field, a kind of tactical data link hierarchical topology method for digging is particularly related to.

Background technology

Tactical data link is typical case application of the data communication in military aspect, and the link of combat data is transmitted with wireless channel Tactical data link tissue under referred to as tactical data link, battlefield surroundings forms one according to command relation and troop operation feature Hierarchical network, therefore, for hierarchical network, the research of topological model is the important foundation that Routing Protocol is assessed, It is to influence the key factor of networking efficiency.The research counted greatly absolutely at present about tactical data link network topology concentrates on movement certainly In networking (mobile ad hoc network, abbreviation MANET), it is a kind of independent of the wireless of any static infrastructure Network, there is larger difference compared with tactical data chain network, most all to use section on network topology in the document of network performance The mode of point random distribution, the characteristics of not embodying stratification, therefore the present invention propose a kind of based on " Apriori data correlations The generating algorithm of the tactical data link hierarchical topology excavation of analysis ", can embody battlefield surroundings lower node stratification feature, realize The hierarchical relationship and correlation degree between different nodes are excavated in the case of network node random distribution.

Apriori algorithm is a kind of frequent item set algorithm of Mining Association Rules, and wherein item destination aggregation (mda) is referred to as item collection, item Element number is concentrated to be referred to as the length of item collection, length is referred to as k- item collections for k item collection, and a support is not less than some threshold value Item collection is called frequent item set.The core concept of the algorithm is by candidate generation and closing downwards detects two stages to excavate Frequent item set finds out all frequency collection first, has been widely used the every field such as business, network security.Apriori Algorithm excavates the correlation degree between correlation degree and project between data using bulk information.Detectd for tactical data link Examine size of message big, the characteristics of node random distribution, on the basis of the basic priori of reconnaissance, excavate different nodes it Between correlation degree, disclose the hierarchical relationship between different nodes in data link.

The content of the invention

It is an object of the invention to：For a class tactical data link of non-stop layer node, in the situation of node random distribution Under, based on Apriori data relation analysis algorithm principles, it is proposed that a kind of tactical data link hierarchical topology method for digging, excavate The hierarchical relationship gone out between different nodes, discloses the hierarchical relationship between different nodes in data link.

Technical scheme comprises the following steps：

S1, in multi-layer network, be each node distribution time slot, and set minimum support and min confidence；

S2, each node of statistics send the frequency of message, and the node by frequency not less than minimum support constitutes frequent k-1 Item collection；

S3, the frequent k-1 centralized node set to acquisition are scanned counting, count each node set and send message Frequency, the node by frequency not less than minimum support constitutes frequent k-2 item collections；

S4, the method iteration according to step S2-S3, until obtaining k dimension Frequent Item Sets；

S5, Frequent Item Sets are tieed up according to the k of acquisition, calculate confidence level, subordinate between egress is obtained according to confidence level and is closed System, obtains the hierarchical relationship between different nodes.

Further, the specific method of the step S2 is：

The time slot of a node is selected from multiple nodes as the condition of separation, so that the transmission for counting each node disappears The frequency of breath, then the frequent k-1 item collections of node composition by frequency not less than minimum support.

Further, in addition to step：

After S6, the time slot of a node of selection are as separation condition, the section of minimum support is less than for those frequencies Point, it is impossible to excavate their hierarchical relationship with separation condition again, it is necessary to reduce the hierarchy levels of cut-point and remove ratio The high node of new spliting node level, goes out not determining subordinate relation node in step s 5 according to step S2-S3 method iteration Composition k dimension Frequent Item Sets, calculate confidence level and subordinate relation between egress；

S7, repeat step S6 are until draw the hierarchical relationship in whole network between all nodes.

Beneficial effects of the present invention are that the present invention can excavate the hierarchical relationship between the different nodes of topological network, take off Show the correlation degree between different nodes, method is simple, works well.

Brief description of the drawings

Fig. 1 is 8 node level relationship topology figures；

Fig. 2 ties up the schematic flow sheet of Frequent Item Sets for acquisition k；

Fig. 3 is node A message sequence；

Fig. 4 is the message time sequence figure of 8 nodes；

Fig. 5 is the support of each platform；

Fig. 6 is the frequent k-2 item collections support of candidate；

Fig. 7 is the frequent k-2 item collections confidence level of candidate；

Fig. 8 is platform B-H message time sequence figure；

Fig. 9 is platform B-H support；

Figure 10 is the frequent k-2 item collections support of candidate.

Embodiment

Below in conjunction with the accompanying drawings and example, technical scheme is described in detail：

The present invention mainly proposes that a kind of Data-Link hierarchical topology based on Apriori data relation analysis is excavated, in number Size of message is scouted according to chain greatly, in the case that Node distribution randomness is big, the priori obtained using Apriori algorithm thought and scouting Knowledge, excavates the hierarchical relationship between different nodes, discloses the correlation degree between different nodes.

By taking eight platforms in a certain operation cluster as an example：

Assuming that there is the hierarchical topology figure between hierarchical relationship, each node between eight platforms as shown in figure 1, agreement is used Tactical data link time slot allocation rule, using P responses, while the message of platform must be in the message transmission interval of upper platform Complete to exchange.

Assuming that I={ A, B, C, D, E, F, G, H } is platform numbering collection, carried out between each node according to network topology structure " answer-answering " formula is communicated, and the period of reception message is divided into isometric timeslice { t₁、t₂、t₃......t_m, each timeslice Size is δ, and m is the timeslice sum of data receipt time section；Define T_i={ i₁、i₂、i₃......i_kBe network in one Communication pattern, i.e., in timeslice t_iMiddle generation communication behavior；

Theoretical according to Apriori algorithm, the model related definition and model used in the present invention is as follows：

Model related definition:

1) correlation rule between definition node

If minimum support is Support (R), min confidence is Confidence (R), when the message sequence between node The minimum support and min confidence of set are met：

Support(R)≥sup_min

Confidence(R)≥conf_min

Then there is certain subordinate relation between node.

2) frequent item set is defined

A) frequency that each node sends message is counted, frequency constitutes item collection as frequency not less than the node of minimum support Numerous k-1 item collections；

B) to second of scanning and counting of carry out of frequent k-1 centralized node set, statistics node sends the frequency of message Find frequent k-2 item collections；

C) (k-1) successively generated after search iteration according to the step of kth -1 ties up maximum Item Sets and produces k Wei Xuanxiangmuji, obtains Frequent Item Sets are tieed up to k；As shown in Figure 2；

D) obtain after target k dimension item collections, calculating confidence level obtains the subordinate relation between egress；

(2) model is described：

Assuming that in hierarchical topology figure (such as Fig. 1), being met in minimum support

Support(R)≥sup_min

In the case of, finally determine association between node by calculating message sequence set frequent item set min confidence Degree, as shown in table 1：

Correlation degree between the frequent item set min confidence of table 1 and node

The final goal of model is that the hierarchical topology relation excavated between eight nodes is：

{A->B->C }, { A->B->D }, { C->E }, { C->F }, { D->G }, { D->H}

Based on model as shown in Figure 1, the side of the present invention is explained in further detail with reference to following real data Case

(1) parameter setting：P=0.8, sup_min=0.6, conf_min=0.6；

(2) according to tactical data link time slot allocation rule, to eight node distribution time slots.Node A message sequences obey pool Pine distribution, obtains node A message arrival time (such as Fig. 3) and 8 node messages timing diagrams (such as Fig. 4)；

(3) node A-D hierarchical relationships；

A) frequent item set k-1 is found：The condition of separation is used as with the message of A nodes, it is possible to obtain following tables of data 2：

The tables of data that table 2 is obtained with the message of A nodes as the condition of separation

In table 2,1：Represent and send signal 0：Represent no signal；

The frequency of the transmission message of each node is counted, frequent k-1 item collections are obtained：{ { A }, { B }, { C }, { D } }.

B) frequent k-2 item collections are found：

There is { { AB }, { AC }, { AD }, { BC }, { BD }, { CD } } by (a) frequent k-2 item collections that candidate can be obtained, calculate them Support (such as Fig. 6) and confidence level (such as Fig. 7)；Frequent Set is selected in the case where confidence level and support meet condition {{AB},{BC},{BD}}。

C) frequent k-3 item collections are found：

There are the frequent k-3 item collections of candidate { { ABC }, { ABD } } on the basis of (b).Calculate the support of the frequent k-3 item collections of candidate Degree and confidence level are as shown in table 3：

The support and confidence level of the frequent k-3 item collections of table 3

Because support and confidence level are all met：

Support(R)≥sup_min

Confidence(R)≥conf_min

So it is considered that having A- in the hierarchical topology of the Data-Link>B->C and A->B->D relation.

(4) node E-H hierarchical relationships：

Because level is unsatisfactory for than relatively low node in the case of using A time slots as cut-point：

Support(R)≥sup_min

So now wanting to obtain the hierarchical relationship of { { E }, { F }, { G }, { H } }, it is necessary to make the hierarchy levels of cut-point Reduce and remove the node higher than new spliting node level.For example, the time slot of reselection B node is used as time slot dividing point, then It is accomplished by first removing the gap information of A nodes, it is to avoid the complexity of frequent item set is selected in increase.A gap information is first removed, with B time slot obtains the time series (such as Fig. 8) of remaining node as time slot dividing point：

A) frequent item set k-1 is found：

The frequency of the transmission message of each node is counted, frequent item set k-1 (such as Fig. 9) is obtained：{ B }, { C }, { D }, { E }, { F }, { G }, { H } }；

B) frequent item set k-2 is found：

The support for knowing all platforms by (a) is all met in the requirement of minimum support, and first time solution procedure Node B and node CD relation are tried to achieve, so the relation of B node and node CD can not be sought during current iteration, is obtained Curve map (such as Fig. 9).As can be seen from Figure 9 { CE }, { CF }, { DG }, the support of { DH } meet condition so will { CE }, { CF }, { DG }, { DH } } it is used as the frequent k-2 item collections of candidate.Obtained confidence level such as table 4 below：

The frequent k-2 item collections confidence level of table 4

Subordinate relation	Confidence level
		{C->E}	0.6875
{C->F}	0.675
		{D->G}	0.654
{D->H}	0.667

Because support and confidence level are all met：

Support(R)≥sup_min

Confidence(R)≥conf_min

Confidence level meets condition as can be seen from the table, that is, there is hierarchical relationship { C->E }, { C->F }, { D->G }, { D-> H}。

Verify whether result of calculation is correct：

It is strong and weak as shown in table 5 by calculating the subordinate relation between each node：

Subordinate relation between each node of table 5 is strong and weak

Subordinate relation	Confidence level
		{AB->C}	0.8682
{AB->D}	0.7984
		{C->E}	0.6875
{C->F}	0.675
		{D->G}	0.654
{D->H}	0.667

The residing level in hierarchical topology figure is substantially can be seen that from curve map 7 higher, k-1, k-2 item collection supports It is higher；More close two nodes (such as AB, BC) in hierarchical topology, confidence level is higher；Confidence level reflects correlation rule Power, confidence level is higher, and subordinate relation is more obvious.When two node astride hierarchies (such as：AC), confidence level declines and works as node on the contrary The networking under specific protocol, then message may downwards be transmitted along the side in topological diagram, forms belong to the collection in time The message sequence set of group, because response and the transmission of message have certain probability P, result in closest on hierarchical relationship Node correlation degree Support highests, the degree of association Support between level reduces therewith, and this also complies with actual operation Situation.

Claims (3)

1. a kind of tactical data link hierarchical topology method, comprises the following steps：

S1, in multi-layer network, be each node distribution time slot, and set minimum support；

S2, each node of statistics send the frequency of message, and the node by frequency not less than minimum support constitutes frequent k-1 Collection；

S3, the frequent k-1 centralized node set to acquisition are scanned counting, count the frequency that each node set sends message Rate, the node by frequency not less than minimum support constitutes frequent k-2 item collections；

S4, the method iteration according to step S2-S3, until obtaining k dimension Frequent Item Sets；

S5, according to the k of acquisition tie up Frequent Item Sets, calculate confidence level, the subordinate relation between egress is obtained according to confidence level, is obtained Obtain the hierarchical relationship between different nodes.

2. a kind of tactical data link hierarchical topology method according to claim 1, it is characterised in that the tool of the step S2 Body method is：

The time slot of a node is selected from multiple nodes as the condition of separation, so that count the transmission message of each node Frequency, then the frequent k-1 item collections of node composition by frequency not less than minimum support.

3. a kind of tactical data link hierarchical topology method according to claim 2, it is characterised in that also including step：

After S6, the time slot of a node of selection are as separation condition, the node of minimum support, nothing are less than for those frequencies Method excavates their hierarchical relationship with separation condition again, it is necessary to the hierarchy levels of cut-point is reduced and is removed than new segmentation The high node of node level, goes out not determining the composition k of subordinate relation node in step s 5 according to step S2-S3 method iteration Frequent Item Sets are tieed up, confidence level is calculated and obtains the subordinate relation between egress；

S7, repeat step S6 are until draw the hierarchical relationship in whole network between all nodes.