CN114239831A - Cross-entity attribute association-based satellite frequency-orbit resource information vector representation method - Google Patents

Cross-entity attribute association-based satellite frequency-orbit resource information vector representation method Download PDF

Info

Publication number
CN114239831A
CN114239831A CN202111569839.3A CN202111569839A CN114239831A CN 114239831 A CN114239831 A CN 114239831A CN 202111569839 A CN202111569839 A CN 202111569839A CN 114239831 A CN114239831 A CN 114239831A
Authority
CN
China
Prior art keywords
entity
orbit
satellite
attribute
cross
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN202111569839.3A
Other languages
Chinese (zh)
Other versions
CN114239831B (en
Inventor
何元智
闫迪
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Institute of Network Engineering Institute of Systems Engineering Academy of Military Sciences
Original Assignee
Institute of Network Engineering Institute of Systems Engineering Academy of Military Sciences
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Institute of Network Engineering Institute of Systems Engineering Academy of Military Sciences filed Critical Institute of Network Engineering Institute of Systems Engineering Academy of Military Sciences
Priority to CN202111569839.3A priority Critical patent/CN114239831B/en
Publication of CN114239831A publication Critical patent/CN114239831A/en
Application granted granted Critical
Publication of CN114239831B publication Critical patent/CN114239831B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N5/00—Computing arrangements using knowledge-based models
    • G06N5/02—Knowledge representation; Symbolic representation

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Computing Systems (AREA)
  • Data Mining & Analysis (AREA)
  • Evolutionary Computation (AREA)
  • Physics & Mathematics (AREA)
  • Computational Linguistics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Artificial Intelligence (AREA)
  • Radio Relay Systems (AREA)

Abstract

The invention discloses a method for representing satellite frequency-orbit resource information vectors based on cross-entity attribute association, which specifically comprises the following steps: defining entity types, entity attributes and entity relations of satellite communication frequency orbit resource knowledge, and representing the satellite communication frequency orbit resource knowledge in a triple form; representing the satellite communication frequency-orbit resource triples into numerical vectors, and defining loss functions of the numerical vectors; and establishing cross-entity attribute correlation characteristics, constructing a multi-objective optimization model, and solving the optimal representation of the satellite communication frequency-orbit resource information vector by using a random gradient descent optimization algorithm. The invention can uniformly and efficiently carry out numerical expression on satellite frequency-orbit resource information with huge data volume, complex association relation and fuzzy characteristics, and can enable the numerical vector representation to be closer to the true value of the satellite frequency-orbit resource information by establishing the cross-entity attribute association characteristic, thereby obtaining the optimal numerical vector representation suitable for the satellite frequency-orbit resource knowledge characteristic.

Description

Cross-entity attribute association-based satellite frequency-orbit resource information vector representation method
Technical Field
The invention belongs to the field of satellite communication, and particularly relates to a cross-entity attribute association-based satellite frequency-orbit resource information vector representation method.
Background
The satellite communication frequency orbit resource has the characteristics of large data volume, complex data association relation, fuzzy characteristics and the like, the satellite communication frequency orbit resource data is uniformly and efficiently expressed, relevant workers can be helped to quickly master relevant information and knowledge of the satellite communication frequency orbit resource, the subsequent information mining work efficiency is improved, the satellite communication frequency orbit resource can be managed, utilized and mined in satellite communication engineering practice, and further the deep mining of the satellite frequency orbit resource and the efficient design of a satellite constellation diagram are helped. In the process of information representation, related workers need to process heterogeneous satellite communication frequency-orbit related data acquired by various types, layers and channels, and if the data is only performed by combining working experience, the data is difficult to be comprehensively, uniformly and efficiently expressed. Meanwhile, the knowledge map is successfully applied to a big data environment as knowledge engineering, and the expression of satellite communication frequency-orbit resource data information can be realized in a more standard form; in addition, the expression of the satellite communication frequency-orbit resource information in the form of the numerical vector is a premise and a basis for applying the artificial intelligence technology to the comprehensive control of the satellite communication frequency-orbit resources. Considering the specific cross-entity attribute correlation characteristic of the satellite communication frequency and orbit resource information, how to more efficiently and pertinently carry out the numerical vector expression is an important problem in the construction and application of the satellite communication frequency and orbit resource knowledge graph.
Disclosure of Invention
Aiming at the characteristics of large data volume, complex data association relation, fuzzy characteristics and the like of satellite communication frequency and orbit resource data, the invention provides an attribute association-based cross-entity satellite frequency and orbit resource knowledge vector representation method, which can carry out machine learning-oriented expression on the satellite communication frequency and orbit resource knowledge with cross-entity attribute association characteristics through a numerical vector optimization model, thereby closely organizing the scattered satellite communication frequency and orbit resource knowledge and being beneficial to the application of artificial intelligence in the field of comprehensive control of satellite communication frequency and orbit resources.
The invention discloses a method for representing satellite frequency-orbit resource information vectors based on cross-entity attribute association, which comprises the following specific steps:
s1, defining the knowledge entity type of satellite communication frequency orbit resources according to the information related to the satellite communication frequency orbit resources in the SRS database, the website and the technical text of the international telecommunication union, wherein the entity type specifically comprises 3 types: the system comprises a satellite entity, an orbit bit resource entity and a frequency resource entity, wherein the satellite entity comprises a satellite and related attributes thereof, the orbit bit resource entity comprises all orbit bit resources of the satellite and related attribute information thereof, and the frequency resource entity comprises satellite use frequency, frequency band and related information thereof;
s2, defining the attribute of the satellite communication frequency orbit resource knowledge entity;
the step S2 specifically includes:
defining a satellite entity attribute set as { satellite name, satellite type, on-orbit condition, country, satellite network number and satellite network starting time }, and recording the satellite entity attribute set as an attribute set S;
defining the attribute set of the rail resource entity as { track name, track type, track position and track height }, and recording the attribute set of the rail resource entity as an attribute set G;
defining the attribute set of the frequency resource entity as { beam name, frequency band type, frequency range, service type and beam quantity }, and marking the attribute set of the frequency resource entity as an attribute set F.
S3, defining the knowledge entity relationship of satellite communication frequency orbit resources;
the step S3 specifically includes:
defining a relationship set among all satellite entities as a { same gateway system, interference relationship and same orbit relationship }, and recording the relationship set among all satellite entities as a relationship set S-S;
defining a relation set among all frequency band entities as { same-frequency uplink and downlink relation }, and recording the relation set among all frequency band entities as a relation set F-F;
defining a set of relations among all the rail position entities as { same rail relation and adjacent rail position relation }, and recording the set of relations among all the rail position entities as a relation set G-G;
defining a relation set between the satellite entity and the orbit entity as { inclusion relation, composition relation and one-to-many correspondence }, and recording the relation set between the satellite entity and the orbit entity as a relation set S-G;
defining a relation set between the satellite entity and the frequency band entity as a { use relation, one-to-many correspondence }, and recording the relation set between the satellite entity and the frequency band entity as a relation set S-F;
defining the relationship set between the rail position entity and the frequency band entity as { inclusion relationship, one-to-many correspondence }, and recording the relationship set between the rail position entity and the frequency band entity as a relationship set G-F.
S4, representing the satellite communication frequency orbit resource knowledge in the form of a triplet;
the step S4 includes the following steps:
describing the satellite communication frequency and orbit resource knowledge in a triple form, and expressing a triple set of the satellite communication frequency and orbit resource knowledge as follows:
Z＝{(h1,r1,t1),(h2,r2,t2),…,(hi,ri,ti),…,(hn,rn,tn)}，
where n represents the total number of triples in the triplet set, i represents the triplet number, and i is 1,2, …, n, hiHead entity representing the ith triplet, tiRepresenting the tail entity of the ith triplet, riRepresenting the directional relation of the ith triple from the head entity to the tail entity, hiHead entity representing the ith triplet, riIs the directed relationship between the head entity and the tail entity of the ith triplet.
S5, translating the satellite communication frequency-orbit resource knowledge triples into numerical vectors based on the translation model, and constructing loss functions of the triples; the step S5 includes the following steps: according to a TransE translation model, each satellite communication frequency-orbit resource knowledge triple (h) is combinedi,ri,ti) Head entity h iniHead entity to tail entity directed relationship riTail entity tiRespectively expressed as three numerical vectors
Figure BDA0003423251670000031
Thereby combining the triplets (h)i,ri,ti) Translation to a de novo entity vector
Figure BDA0003423251670000032
Vector of entities to tail
Figure BDA0003423251670000033
Using the relationship vector
Figure BDA0003423251670000034
Representing all satellite communication frequency orbit resource knowledge entities into corresponding numerical vectors; according to head entity vector
Figure BDA0003423251670000041
Adding a relationship vector
Figure BDA0003423251670000042
Sum, and tail entity vector
Figure BDA0003423251670000043
Distance of (d), constructing a triplet (h)i,ri,ti) Loss function of
Figure BDA0003423251670000044
Is composed of
Figure BDA0003423251670000045
Wherein
Figure BDA0003423251670000046
Representing the L2 norm.
S6, establishing a multi-objective optimization model based on the cross-entity attribute association characteristics;
the step S6 includes the following steps:
according to the correlation characteristics among various entity attributes in the satellite communication frequency and orbit resource knowledge, in all the satellite communication frequency and orbit resource knowledge triad sets Z, any two triads (h) are selectedi,ri,ti)、(hj,rj,tj) If it is the head entity hiAnd tail entity tiMiddle and head entity hjAnd tail entity tjThere is an association relationship on the attribute between them, and tiAnd hjWhen the same entity is designated, the entity h is determinediWith entity tjThere exists a cross-entity attribute association property between them, called (h)i,ri,rj,tj) Associating entity groups for a cross-entity attribute; finding all cross-entity attribute associated entity groups in all satellite communication frequency-orbit resource knowledge triple sets Z, describing the cross-entity attribute associated entity groups in a digitized vector form, and constructing a set of the cross-entity attribute associated entity groups, wherein the expression formula is as follows:
S＝{(H1,R11,R12,T1),(H2,R21,R22,T2),…,(Hj,Rj1,Rj2,Tj),…,(Hm,Rm1,Rm2,Tm)}，
where m represents the total number of entity groups associated across entity attributes, j represents the number of entity groups associated across entity attributes and j is 1,2, …, m, HjRepresenting the head entity, T, of the jth cross-entity-attribute associated entity groupjRepresenting the jth Tail entity, R, of the Cross-entity Attribute Association entity groupj1Representing a directed relationship, R, of a first attribute associated triplet of a jth cross-entity attribute associated entity groupj2Representing a directed relationship of a second attribute association triple of the jth cross-entity attribute association entity group;
constructing the jth cross-entity attribute association entity group (H)j,Rj1,Rj2,Tj) The loss function of (2), expressed as:
Figure BDA0003423251670000047
wherein
Figure BDA0003423251670000048
The norm of L2 is shown,
Figure BDA0003423251670000049
a vector representation representing the head entity of the jth cross-entity-attribute associated entity group,
Figure BDA0003423251670000051
a vector representation representing the tail entity of the jth cross-entity-attribute associated entity group,
Figure BDA0003423251670000052
a vector representation of the directional relationship representing the first attribute association triple of the jth cross-entity attribute association entity group,
Figure BDA0003423251670000053
a vector representation of a directed relationship representing a second attribute association triple of the jth cross-entity attribute association entity group;
establishing a multi-objective optimization model by taking the sum of the loss functions of all the triples as a first objective and the sum of the loss functions of all the cross-entity attribute association entity groups as a second objective:
Figure BDA0003423251670000054
wherein Z is all triple sets of the satellite communication frequency and orbit resource knowledge, and S is all cross-entity attribute association entity set in the satellite communication frequency and orbit resource knowledge.
S7, solving the multi-objective optimization model by using a random gradient descent optimization algorithm to obtain an optimal solution, and using the optimal solution as the optimal vector representation of the satellite communication frequency-orbit resource knowledge;
the step S7 includes the following steps:
s71, according to the importance of the two targets in the multi-target optimization model, giving different weight values lambda to the two targets1And λ2And 0 < lambda1＜1,0＜λ2＜1,λ1+λ2The loss function for obtaining the knowledge of the satellite communication frequency orbit resources is 1:
Figure BDA0003423251670000055
where θ ═ θ1,θ2,...θk,...θKThe resource knowledge entity represents a set of numerical vectors represented by all satellite communication frequency orbit resource knowledge entities, wherein K is 1, 2.
S72, initializing the stochastic gradient descent optimization algorithm, initializing each numerical vector in theta to a default value, and setting the termination distance epsilon and the step length alpha of the algorithm;
s73, traversing all the numerical vectors in the set theta, calculating the gradient of each numerical vector, and aiming at the numerical vector thetakGradient thereof
Figure BDA0003423251670000056
The expression of (a) is:
Figure BDA0003423251670000061
wherein K is 1,2, the.. and K, l and o are sequence numbers of randomly selected triples and cross-entity attribute associated entity groups in the triplet set and the cross-entity attribute associated entity group respectively, l belongs to {1,2, …, n }, and o belongs to {1,2, …, m };
s74, multiplying the step length alpha by the gradient of the numerical vector to obtain the descending distance of the numerical vector;
s75, judging whether the descending distance of each numerical vector in the set theta is smaller than epsilon, if so, selecting the current set theta as the final result of the algorithm, wherein each numerical vector in the current set theta is the vector optimal representation of the satellite communication frequency-orbit resource knowledge, ending the algorithm, otherwise, entering the next step;
s76, updating all the numerical vectors in the set theta, and updating the formula as
Figure BDA0003423251670000062
And the process returns to step S73.
The invention has the following advantages:
(1) the satellite frequency orbit resource information is expressed as a numerical vector, so that the satellite frequency orbit resource information with huge data volume, complex incidence relation and fuzzy characteristics can be uniformly and efficiently numerically expressed, the satellite frequency orbit resource information is conveniently analyzed and calculated numerically, and the satellite frequency orbit resource information is favorably utilized to carry out effective satellite frequency orbit resource excavation and satellite constellation diagram design;
(2) the invention provides the cross-entity attribute correlation characteristic of the satellite frequency-orbit resource information, and the loss function represented by the satellite frequency-orbit resource information triple is minimized as much as possible by utilizing the characteristic, namely, the numerical vector representation is closer to the true value of the satellite frequency-orbit resource information, so that the representation error is effectively reduced, and the satellite resource state can be more truly represented.
Drawings
Fig. 1 is a schematic diagram of a method for representing a satellite frequency-orbit resource information vector based on cross-entity attribute association in the present invention.
Detailed Description
The present invention will be described in detail with reference to examples.
Fig. 1 is a schematic diagram of a method for representing a satellite frequency-orbit resource information vector based on cross-entity attribute association in the present invention.
The invention discloses a method for representing satellite frequency-orbit resource information vectors based on cross-entity attribute association, which comprises the following specific steps:
s1, defining a satellite communication frequency and orbit resource knowledge entity type according to information related to satellite communication frequency and orbit resources in an SRS (space radio communications states) database, a website and a technical text of the international telecommunication union, wherein the entity type specifically comprises 3 types: the system comprises a satellite entity, an orbit bit resource entity and a frequency resource entity type, wherein the satellite entity comprises a satellite and related attributes thereof, the orbit bit resource entity comprises all orbit bit resources of the satellite and related attribute information thereof, and the frequency resource entity comprises satellite use frequency, frequency band and related information thereof;
s2, defining the attribute of the satellite communication frequency orbit resource knowledge entity;
the step S2 specifically includes:
defining a satellite entity attribute set as { satellite name, satellite type, on-orbit condition, country, satellite network number and satellite network starting time }, and recording the satellite entity attribute set as an attribute set S;
defining the attribute set of the rail resource entity as { track name, track type, track position and track height }, and recording the attribute set of the rail resource entity as an attribute set G;
defining the attribute set of the frequency resource entity as { beam name, frequency band type, frequency range, service type and beam quantity }, and marking the attribute set of the frequency resource entity as an attribute set F.
S3, defining the knowledge entity relationship of satellite communication frequency orbit resources;
the step S3 specifically includes:
defining a relationship set among all satellite entities as a { same gateway system, interference relationship and same orbit relationship }, and recording the relationship set among all satellite entities as a relationship set S-S;
defining a relation set among all frequency band entities as { same-frequency uplink and downlink relation }, and recording the relation set among all frequency band entities as a relation set F-F;
defining a set of relations among all the rail position entities as { same rail relation and adjacent rail position relation }, and recording the set of relations among all the rail position entities as a relation set G-G;
defining a relation set between the satellite entity and the orbit entity as { inclusion relation, composition relation and one-to-many correspondence }, and recording the relation set between the satellite entity and the orbit entity as a relation set S-G;
defining a relation set between the satellite entity and the frequency band entity as a { use relation, one-to-many correspondence }, and recording the relation set between the satellite entity and the frequency band entity as a relation set S-F;
defining the relationship set between the rail position entity and the frequency band entity as { inclusion relationship, one-to-many correspondence }, and recording the relationship set between the rail position entity and the frequency band entity as a relationship set G-F.
S4, representing the satellite communication frequency orbit resource knowledge in the form of a triplet;
the step S4 includes the following steps:
describing the satellite communication frequency and orbit resource knowledge in a triple form, and expressing a triple set of the satellite communication frequency and orbit resource knowledge as follows:
Z＝{(h1,r1,t1),(h2,r2,t2),…,(hi,ri,ti),…,(hn,rn,tn)}，
where n represents the total number of triples in the triplet set, i represents the triplet number, and i is 1,2, …, n, hiHead entity representing the ith triplet, tiRepresenting the tail entity of the ith triplet, riRepresenting the directional relation of the ith triple from the head entity to the tail entity, hiHead entity representing the ith triplet, riIs the directed relationship between the head entity and the tail entity of the ith triplet.
S5, translating the satellite communication frequency-orbit resource knowledge triples into numerical vectors based on the translation model, and constructing loss functions of the triples; the step S5 includes the following steps: according to a TransE translation model, each satellite communication frequency-orbit resource knowledge triple (h) is combinedi,ri,ti) Head entity h iniHead entity to tail entity directed relationship riTail entity tiRespectively expressed as three numerical vectors
Figure BDA0003423251670000081
Thereby combining the triplets (h)i,ri,ti) Translation to a de novo entity vector
Figure BDA0003423251670000082
Vector of entities to tail
Figure BDA0003423251670000083
Using the relationship vector
Figure BDA0003423251670000084
Representing all satellite communication frequency orbit resource knowledge entities into corresponding numerical vectors; according to head entity vector
Figure BDA0003423251670000091
Adding a relationship vector
Figure BDA0003423251670000092
Sum, and tail entity vector
Figure BDA0003423251670000093
Distance of (d), constructing a triplet (h)i,ri,ti) Loss function of
Figure BDA0003423251670000094
Is composed of
Figure BDA0003423251670000095
Wherein
Figure BDA0003423251670000096
Representing the L2 norm.
S6, establishing a multi-objective optimization model based on the cross-entity attribute association characteristics;
the step S6 includes the following steps:
according to the correlation characteristics among various entity attributes in the satellite communication frequency and orbit resource knowledge, in all the satellite communication frequency and orbit resource knowledge triad sets Z, any two triads (h) are selectedi,ri,ti)、(hj,rj,tj) If it is the head entity hiAnd tail entity tiMiddle and head entity hjAnd tail entity tjThere is an association relationship on the attribute between them, and tiAnd hjWhen the same entity is designated, the judgment is madeEntity hiWith entity tjThere exists a cross-entity attribute association property between them, called (h)i,ri,rj,tj) Associating entity groups for a cross-entity attribute; finding all cross-entity attribute associated entity groups in all satellite communication frequency-orbit resource knowledge triple sets Z, describing the cross-entity attribute associated entity groups in a digitized vector form, and constructing a set of the cross-entity attribute associated entity groups, wherein the expression formula is as follows:
S＝{(H1,R11,R12,T1),(H2,R21,R22,T2),…,(Hj,Rj1,Rj2,Tj),…,(Hm,Rm1,Rm2,Tm)}，
where m represents the total number of entity groups associated across entity attributes, j represents the number of entity groups associated across entity attributes and j is 1,2, …, m, HjRepresenting the head entity, T, of the jth cross-entity-attribute associated entity groupjRepresenting the jth Tail entity, R, of the Cross-entity Attribute Association entity groupj1Representing a directed relationship, R, of a first attribute associated triplet of a jth cross-entity attribute associated entity groupj2Representing a directed relationship of a second attribute association triple of the jth cross-entity attribute association entity group;
constructing the jth cross-entity attribute association entity group (H)j,Rj1,Rj2,Tj) The loss function of (2), expressed as:
Figure BDA0003423251670000097
wherein
Figure BDA0003423251670000098
The norm of L2 is shown,
Figure BDA0003423251670000099
a vector representation representing the head entity of the jth cross-entity-attribute associated entity group,
Figure BDA0003423251670000101
a vector representation representing the tail entity of the jth cross-entity-attribute associated entity group,
Figure BDA0003423251670000102
a vector representation of the directional relationship representing the first attribute association triple of the jth cross-entity attribute association entity group,
Figure BDA0003423251670000103
a vector representation of a directed relationship representing a second attribute association triple of the jth cross-entity attribute association entity group;
establishing a multi-objective optimization model by taking the sum of the loss functions of all the triples as a first objective and the sum of the loss functions of all the cross-entity attribute association entity groups as a second objective:
Figure BDA0003423251670000104
wherein Z is all triple sets of the satellite communication frequency and orbit resource knowledge, and S is all cross-entity attribute association entity set in the satellite communication frequency and orbit resource knowledge.
S7, solving the multi-objective optimization model by using a random gradient descent optimization algorithm to obtain an optimal solution, and using the optimal solution as the optimal vector representation of the satellite communication frequency-orbit resource knowledge;
the step S7 includes the following steps:
s71, according to the importance of the two targets in the multi-target optimization model, giving different weight values lambda to the two targets1And λ2And 0 < lambda1＜1,0＜λ2＜1,λ1+λ2The loss function for obtaining the knowledge of the satellite communication frequency orbit resources is 1:
Figure BDA0003423251670000105
where θ ═ θ1,θ2,...θk,…θKThe resource knowledge entity represents a set of numerical vectors represented by all satellite communication frequency orbit resource knowledge entities, wherein K is 1, 2.
S72, initializing the stochastic gradient descent optimization algorithm, initializing each numerical vector in theta to a default value, and setting the termination distance epsilon and the step length alpha of the algorithm;
s73, traversing all the numerical vectors in theta, calculating the gradient of each numerical vector, and aiming at the numerical vectors thetakGradient thereof
Figure BDA0003423251670000106
The expression of (a) is:
Figure BDA0003423251670000111
wherein K is 1,2, the.. and K, l and o are sequence numbers of randomly selected triples and cross-entity attribute associated entity groups in the triplet set and the cross-entity attribute associated entity group respectively, l belongs to {1,2, …, n }, and o belongs to {1,2, …, m };
s74, for each numerical vector, multiplying the step length alpha by the gradient of the numerical vector to obtain the descending distance of the numerical vector;
s75, judging whether the descending distance of each numerical vector in the set theta is smaller than epsilon, if so, selecting the current set theta as the final result of the algorithm, wherein each numerical vector in the current set theta is the vector optimal representation of the satellite communication frequency-orbit resource knowledge, ending the algorithm, otherwise, entering the next step;
s76, updating all the numerical vectors in the theta, and updating the formula as
Figure BDA0003423251670000112
And the process returns to step S73.
The above description is only an example of the present application and is not intended to limit the present application. Various modifications and changes may occur to those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the scope of the claims of the present application.

Claims (7)

1. A method for representing satellite frequency-orbit resource information vectors based on cross-entity attribute association is characterized by comprising the following specific steps:
s1, defining the knowledge entity type of satellite communication frequency orbit resources according to the information related to the satellite communication frequency orbit resources in the SRS database, the website and the technical text of the international telecommunication union, wherein the entity type specifically comprises 3 types: a satellite entity, an orbit resource entity and a frequency resource entity;
s2, defining the attribute of the satellite communication frequency orbit resource knowledge entity;
s3, defining the knowledge entity relationship of satellite communication frequency orbit resources;
s4, representing the satellite communication frequency orbit resource knowledge in the form of a triplet;
s5, translating the satellite communication frequency-orbit resource knowledge triples into numerical vectors based on the translation model, and constructing loss functions of the triples;
s6, establishing a multi-objective optimization model based on the cross-entity attribute association characteristics;
and S7, solving the multi-objective optimization model by using a random gradient descent optimization algorithm to obtain an optimal solution, and using the optimal solution as the optimal vector representation of the satellite communication frequency-orbit resource knowledge.
2. The method as claimed in claim 1, wherein the step S2 specifically includes:
defining a satellite entity attribute set as { satellite name, satellite type, on-orbit condition, country, satellite network number and satellite network starting time }, and recording the satellite entity attribute set as an attribute set S;
defining the attribute set of the rail resource entity as { track name, track type, track position and track height }, and recording the attribute set of the rail resource entity as an attribute set G;
defining the attribute set of the frequency resource entity as { beam name, frequency band type, frequency range, service type and beam quantity }, and marking the attribute set of the frequency resource entity as an attribute set F.
3. The method of claim 1, wherein the cross-entity attribute association based satellite frequency-orbit resource information vector representation,
the step S3 specifically includes:
defining a relationship set among all satellite entities as a { same gateway system, interference relationship and same orbit relationship }, and recording the relationship set among all satellite entities as a relationship set S-S;
defining a relation set among all frequency band entities as { same-frequency uplink and downlink relation }, and recording the relation set among all frequency band entities as a relation set F-F;
defining a set of relations among all the rail position entities as { same rail relation and adjacent rail position relation }, and recording the set of relations among all the rail position entities as a relation set G-G;
defining a relation set between the satellite entity and the orbit entity as { inclusion relation, composition relation and one-to-many correspondence }, and recording the relation set between the satellite entity and the orbit entity as a relation set S-G;
defining a relation set between the satellite entity and the frequency band entity as a { use relation, one-to-many correspondence }, and recording the relation set between the satellite entity and the frequency band entity as a relation set S-F;
defining the relationship set between the rail position entity and the frequency band entity as { inclusion relationship, one-to-many correspondence }, and recording the relationship set between the rail position entity and the frequency band entity as a relationship set G-F.
4. The method of claim 1, wherein the cross-entity attribute association based satellite frequency-orbit resource information vector representation,
the step S4 includes the following steps:
describing the satellite communication frequency and orbit resource knowledge in a triple form, and expressing a triple set of the satellite communication frequency and orbit resource knowledge as follows:
Z＝{(h1,r1,t1),(h2,r2,t2),…,(hi,ri,ti),…,(hn,rn,tn)}，
where n represents the total number of triples in the triplet set, i represents the triplet number, and i is 1,2, …, n, hiHead entity representing the ith triplet, tiRepresenting the tail entity of the ith triplet, riRepresenting the directional relation of the ith triple from the head entity to the tail entity, hiHead entity representing the ith triplet, riIs the directed relationship between the head entity and the tail entity of the ith triplet.
5. The method of claim 1, wherein the cross-entity attribute association based satellite frequency-orbit resource information vector representation,
the step S5 includes the following steps: according to a TransE translation model, each satellite communication frequency-orbit resource knowledge triple (h) is combinedi,ri,ti) Head entity h iniHead entity to tail entity directed relationship riTail entity tiRespectively expressed as three numerical vectors
Figure FDA0003423251660000031
Thereby combining the triplets (h)i,ri,ti) Translation to a de novo entity vector
Figure FDA0003423251660000032
Vector of entities to tail
Figure FDA0003423251660000033
Using the relationship vector
Figure FDA0003423251660000034
Characterization performed will beThe knowledge entity of the satellite communication frequency orbit resource is expressed into a corresponding numerical vector; according to head entity vector
Figure FDA0003423251660000035
Adding a relationship vector
Figure FDA0003423251660000036
Sum, and tail entity vector
Figure FDA0003423251660000037
Distance of (d), constructing a triplet (h)i,ri,ti) Loss function of
Figure FDA0003423251660000038
Is composed of
Figure FDA0003423251660000039
Wherein
Figure FDA00034232516600000310
Representing the L2 norm.
6. The method of claim 1, wherein the cross-entity attribute association based satellite frequency-orbit resource information vector representation,
the step S6 includes the following steps:
according to the correlation characteristics among various entity attributes in the satellite communication frequency and orbit resource knowledge, in all the satellite communication frequency and orbit resource knowledge triad sets Z, any two triads (h) are selectedi,ri,ti)、(hj,rj,tj) If it is the head entity hiAnd tail entity tiMiddle and head entity hjAnd tail entity tjThere is an association relationship on the attribute between them, and tiAnd hjWhen the same entity is designated, the entity h is determinediWith entity tjThere exists a cross-entity attribute association property between them, called (h)i,ri,rj,tj) Associating entity groups for a cross-entity attribute; finding all cross-entity attribute associated entity groups in all satellite communication frequency-orbit resource knowledge triple sets Z, describing the cross-entity attribute associated entity groups in a digitized vector form, and constructing a set of the cross-entity attribute associated entity groups, wherein the expression formula is as follows:
S＝{(H1,R11,R12,T1),(H2,R21,R22,T2),…,(Hj,Rj1,Rj2,Tj),…,(Hm,Rm1,Rm2,Tm)}，
where m represents the total number of entity groups associated across entity attributes, j represents the number of entity groups associated across entity attributes and j is 1,2, …, m, HjRepresenting the head entity, T, of the jth cross-entity-attribute associated entity groupjRepresenting the jth Tail entity, R, of the Cross-entity Attribute Association entity groupj1Representing a directed relationship, R, of a first attribute associated triplet of a jth cross-entity attribute associated entity groupj2Representing a directed relationship of a second attribute association triple of the jth cross-entity attribute association entity group;
constructing the jth cross-entity attribute association entity group (H)j,Rj1,Rj2,Tj) The loss function of (2), expressed as:
Figure FDA0003423251660000041
wherein
Figure FDA0003423251660000042
The norm of L2 is shown,
Figure FDA0003423251660000043
a vector representation representing the head entity of the jth cross-entity-attribute associated entity group,
Figure FDA0003423251660000044
a vector representation representing the tail entity of the jth cross-entity-attribute associated entity group,
Figure FDA0003423251660000045
a vector representation of the directional relationship representing the first attribute association triple of the jth cross-entity attribute association entity group,
Figure FDA0003423251660000046
a vector representation of a directed relationship representing a second attribute association triple of the jth cross-entity attribute association entity group;
establishing a multi-objective optimization model by taking the sum of the loss functions of all the triples as a first objective and the sum of the loss functions of all the cross-entity attribute association entity groups as a second objective:
Figure FDA0003423251660000047
s.t.(hi,ri,ti)∈Z
(Hj,Rj1,Rj2,Tj)∈S
wherein Z is all triple sets of the satellite communication frequency and orbit resource knowledge, and S is all cross-entity attribute association entity set in the satellite communication frequency and orbit resource knowledge.
7. The method of claim 1, wherein the cross-entity attribute association based satellite frequency-orbit resource information vector representation,
the step S7 includes the following steps:
s71, according to the importance of the two targets in the multi-target optimization model, giving different weight values lambda to the two targets1And λ2And 0 < lambda1＜1,0＜λ2＜1,λ1+λ2Obtaining a loss function of knowledge of satellite communication frequency-orbit resourcesComprises the following steps:
Figure FDA0003423251660000048
where θ ═ θ1,θ2,...θk,...θKThe resource knowledge entity represents a set of numerical vectors represented by all satellite communication frequency orbit resource knowledge entities, wherein K is 1, 2.
S72, initializing the stochastic gradient descent optimization algorithm, initializing each numerical vector in the set theta to a default value, and setting the termination distance epsilon and the step length alpha of the algorithm;
s73, traversing all the numerical vectors in the set theta, calculating the gradient of each numerical vector, and aiming at the numerical vector thetakGradient thereof
Figure FDA0003423251660000051
The expression of (a) is:
Figure FDA0003423251660000052
wherein K is 1,2, the.. and K, l and o are sequence numbers of randomly selected triples and cross-entity attribute associated entity groups in the triplet set and the cross-entity attribute associated entity group respectively, l belongs to {1,2, …, n }, and o belongs to {1,2, …, m };
s74, multiplying the step length alpha by the gradient of the numerical vector to obtain the descending distance of the numerical vector;
s75, judging whether the descending distance of each numerical vector in the set theta is smaller than epsilon, if so, selecting the current set theta as the final result of the algorithm, wherein each numerical vector in the current set theta is the vector optimal representation of the satellite communication frequency-orbit resource knowledge, ending the algorithm, otherwise, entering the next step;
s76, updating all the numerical vectors in the set theta, and updating the formula as
Figure FDA0003423251660000053
And the process returns to step S73.
CN202111569839.3A 2021-12-21 2021-12-21 Cross-entity attribute association-based satellite frequency-orbit resource information vector representation method Active CN114239831B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202111569839.3A CN114239831B (en) 2021-12-21 2021-12-21 Cross-entity attribute association-based satellite frequency-orbit resource information vector representation method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202111569839.3A CN114239831B (en) 2021-12-21 2021-12-21 Cross-entity attribute association-based satellite frequency-orbit resource information vector representation method

Publications (2)

Publication Number Publication Date
CN114239831A true CN114239831A (en) 2022-03-25
CN114239831B CN114239831B (en) 2022-06-10

Family

ID=80760256

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202111569839.3A Active CN114239831B (en) 2021-12-21 2021-12-21 Cross-entity attribute association-based satellite frequency-orbit resource information vector representation method

Country Status (1)

Country Link
CN (1) CN114239831B (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114880484A (en) * 2022-05-11 2022-08-09 军事科学院系统工程研究院网络信息研究所 Satellite communication frequency-orbit resource map construction method based on vector mapping

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111753098A (en) * 2020-06-23 2020-10-09 陕西师范大学 Teaching method and system based on cross-media dynamic knowledge graph
CN112131404A (en) * 2020-09-19 2020-12-25 哈尔滨工程大学 Entity alignment method in four-risk one-gold domain knowledge graph
FR3099673A1 (en) * 2019-07-31 2021-02-05 Thales PROCESS FOR DETERMINING THE CONSTRAINTS OF A NON-GEOSTATIONARY SYSTEM WITH RESPECT TO ANOTHER NON-GEOSTATIONARY SYSTEM
US20210201015A1 (en) * 2019-12-27 2021-07-01 Robert Bosch Gmbh System and method for enhancing neural sentence classification
CN113194084A (en) * 2021-04-27 2021-07-30 中国人民解放军国防科技大学 Satellite network security analysis method, device, system and storage medium
CN113220894A (en) * 2021-02-07 2021-08-06 国家卫星气象中心(国家空间天气监测预警中心) Intelligent satellite remote sensing data acquisition method based on perception calculation

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
FR3099673A1 (en) * 2019-07-31 2021-02-05 Thales PROCESS FOR DETERMINING THE CONSTRAINTS OF A NON-GEOSTATIONARY SYSTEM WITH RESPECT TO ANOTHER NON-GEOSTATIONARY SYSTEM
US20210201015A1 (en) * 2019-12-27 2021-07-01 Robert Bosch Gmbh System and method for enhancing neural sentence classification
CN111753098A (en) * 2020-06-23 2020-10-09 陕西师范大学 Teaching method and system based on cross-media dynamic knowledge graph
CN112131404A (en) * 2020-09-19 2020-12-25 哈尔滨工程大学 Entity alignment method in four-risk one-gold domain knowledge graph
CN113220894A (en) * 2021-02-07 2021-08-06 国家卫星气象中心(国家空间天气监测预警中心) Intelligent satellite remote sensing data acquisition method based on perception calculation
CN113194084A (en) * 2021-04-27 2021-07-30 中国人民解放军国防科技大学 Satellite network security analysis method, device, system and storage medium

Non-Patent Citations (4)

* Cited by examiner, † Cited by third party
Title
ZHEN WANG等: "Knowledge Graph Embedding by Translating on Hyperplanes", 《TWENTY-EIGHTH AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE》 *
ZIKANG WANG等: "Multimodal Data Enhanced Representation Learning for Knowledge Graphs", 《2019 INTERNATIONAL JOINT CONFERENCE ON NEURAL NETWORKS (IJCNN)》 *
刘舆等: "基于知识图谱的卫星情报分析方法研究", 《情报探索》 *
廖琼: "卫星数传调度优化算法及可视化仿真技术研究", 《中国优秀硕士学位论文全文数据库 (工程科技Ⅱ辑)》 *

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114880484A (en) * 2022-05-11 2022-08-09 军事科学院系统工程研究院网络信息研究所 Satellite communication frequency-orbit resource map construction method based on vector mapping

Also Published As

Publication number Publication date
CN114239831B (en) 2022-06-10

Similar Documents

Publication Publication Date Title
CN104794195B (en) A kind of data digging method found for the potential renewal user of telecommunications
CN109117560A (en) Automobile typical machined part three-dimensional process design method and platform based on MBD
CN112364908B (en) Longitudinal federal learning method oriented to decision tree
CN105740245A (en) Frequent item set mining method
CN103593400A (en) Lightning activity data statistics method based on modified Apriori algorithm
CN106021560A (en) Data-flow-oriented construction method for low-time-delay memory B+ tree index
CN114239831B (en) Cross-entity attribute association-based satellite frequency-orbit resource information vector representation method
CN109787821B (en) Intelligent prediction method for large-scale mobile client traffic consumption
Kiran et al. Resource allocation optimization in LTE-A/5G networks using big data analytics
CN115525038A (en) Equipment fault diagnosis method based on federal hierarchical optimization learning
Wang et al. A simulation approach to the process planning problem using a modified particle swarm optimization.
CN116050540A (en) Self-adaptive federal edge learning method based on joint bi-dimensional user scheduling
CN106202347A (en) A kind of device excavated with useful data for data quality management
US20220188649A1 (en) Decision tree-oriented vertical federated learning method
CN109302734B (en) Access mode selection method for wireless access network planning
Xu et al. Federated traffic synthesizing and classification using generative adversarial networks
CN113703482B (en) Task planning method based on simplified attention network in large-scale unmanned aerial vehicle cluster
CN106649203A (en) Method for improving processing quality of big data
CN112020085B (en) Node failure sweep effect analysis method for aviation ad hoc network
CN113378514B (en) Multi-label data feature selection processing method and device
Li et al. Virtual network embedding based on multi-objective group search optimizer
CN117574779B (en) Groundwater monitoring network optimization method for improving quantum particle swarm
CN108876648A (en) Social network influence power maximization approach and system based on coverage range
CN117635218B (en) Business district flow prediction method based on six-degree separation theory and graph annotation network
Lyu et al. Petri nets for performance modelling study of client-server systems

Legal Events

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