Frequency assignment for satellite communication systems - PowerPoint Presentation

Slide1

Frequency assignment for satellite communication systemsKata KIATMANAROJSupervisors: Christian ARTIGUES, Laurent HOUSSIN

1Slide2

ContentsProblem definitionCurrent state of the artContributionsConclusions and perspectives2Slide3

Problem definition3Slide4

Problem definitionTo assign a limited number of frequencies to as many users as possible within a service area4Slide5

Problem definitionTo assign a limited number of frequencies to as many users as possible within a service areaFrequency is a limited resource!Frequency reuse -> co-channel interference

Intra-system interference

5Slide6

Problem definitionSimplified beamSDMA: Spatial Division Multiple Access6

j

k

iSlide7

Problem definitionTo assign a limited number of frequencies to as many users as possible within a service areaFrequency is a limited resource!Frequency reuse -> co-channel interference

Intra-system interference

Graph coloring problem

NP-hard

7Slide8

Problem definitionInterference constraints8

i

j

i

j

k

Binary interference

Cumulative interference

Acceptable interference threshold

Interference coefficientsSlide9

Problem definitionAssignmentLogical boxes (superframes)Demand = |F|x|T|No overlapping within the

superframe

Overlapping

between

superframes (simultaneous)



may create interference

9

0 ≤

o

ij

≤ 1

1

2Slide10

Problem definitionSuperframe structure10Slide11

Problem definitionFrames and satellite beams11Slide12

Problem definition12Slide13

Current state of the art13Slide14

Current state of the art - FAPDistance FAPsMaximum Service FAPMinimum Order FAPMinimum Span FAPMinimum Interference FAPSolving methodsExact methodHeuristics and

metaheuristics

14Slide15

Current state of the art – satellite FAPTwo branchesInter-system interferenceIntra-system interferenceInter-system interferenceTwo or more adjacent satellitesMinimize co-channel interference (multiple carriers)Intra-system interference

Multi-spot beams

Geographical zones assuming the same propagation condition

15Slide16

Contributions16Slide17

ContributionsPart 1: Single carrier modelsPart 2: Multiple carrier modelsPart 3: Industrial application17Slide18

Single carrier models18K. Kiatmanaroj, C. Artigues, L. Houssin, and F. Messine

, Frequency assignment in a SDMA satellite communication system with beam

decentring

feature, submitted to Computational Optimization and Applications (COA)

K.

Kiatmanaroj

, C.

Artigues

, L.

Houssin

, and F.

Messine

, Frequency allocation in a SDMA satellite communication system with beam moving, IEEE International Conference on Communications (ICC), 2012

K.

Kiatmanaroj

, C.

Artigues

, L.

Houssin

, and F.

Messine

, Hybrid discrete-continuous optimization for the frequency assignment problem in satellite communication system, IFAC symposium on Information Control in Manufacturing (INCOM), 2012Slide19

Single carrier models1 frequency over the total durationSame frequency

+

located too close -> Interference

3 models (supplied by Thales

Alenia

Space)

19Slide20

Single carrier modelsModel 1 (fixed-beam binary interference)40 fixed-beams2 frequencies / beam even no userInterference matrix (binary interference)Graph coloring: DSAT algorithm -> 4 colors

20

8 frequencies in totalSlide21

Single carrier modelsModel 2 (fixed-beam varying frequency)40 fixed-beams8 frequencies (different within the same beam)Cumulative interferenceGreedy vs. ILP

21Slide22

Single carrier modelsModel 3 (SDMA-beam varying frequency)SDMA (beam-centered)8 frequencies (different within the same beam)Cumulative interferenceGreedy vs. ILP

22Slide23

Single carrier modelsGreedy algorithmsUser selection rulesFrequency selection rules23Slide24

Single carrier modelsGreedy algorithmsUser selection rulesFrequency selection rules24Slide25

Single carrier modelsInteger Linear Programming (ILP)25Slide26

Single carrier models26Performance comparison

ILP 60 secSlide27

Single carrier models27ILP performancesSlide28

Continuous optimization28* Collaboration with Frédéric Mezzine, IRIT, ToulouseSlide29

Continuous optimizationBeam moving algorithmFor each unassigned userContinuously move the interferers’ beams from their center positionsNon-linear antenna gainMinimize the moveNot violating interference constraints29Slide30

Continuous optimization30

i

j

k

x

User

i

Gain

α

i

Δ

ix

i

Δ

ix

+

j

Δ

jx

+

k

Δ

kx

+

x

0

-

Matlab’s

solver

fminconSlide31

Continuous optimization31

i

j

k

x

User

i

Gain

α

i

Δ

ix

i

↓

↓

↓

↓+

j

k

x

-

Matlab’s

solver

fminconSlide32

Continuous optimization32

i

j

k

x

User

i

Gain

α

i

Δ

ix

i

↓

↓

↓

↓

j

k

x

-

Matlab’s

solver

fminconSlide33

Continuous optimization33

i

j

k

x

User

i

Gain

α

i

Δ

ix

i

↓

↓

↓

↓-

j

k

x

-

Matlab’s

solver

fminconSlide34

Continuous optimization34

i

j

k

x

User

i

Gain

α

i

Δ

ix

i

↓

↓

↓

↓

j

↓

↓

↓

↓

k

↓

↓

↓

↓

x

+

Matlab’s

solver

fminconSlide35

Continuous optimization35Matlab’s solver fmincon

k: number of beams to be moved

MAXINEG: margin from the interference threshold

UTVAR: whether to include user x to the moveSlide36

Continuous optimization36Matlab’s solver fmincon

Parameters: k, MAXINEG, UTVARSlide37

Continuous optimization37

Beam moving results with k-MAXINEG-UTVAR = 7-2-0Slide38

Continuous optimization38

Beam moving results with k-MAXINEG-UTVAR = 7-2-0Slide39

Continuous optimization39

Closed-loop implementationSlide40

Conclusions and further study – Part 1Greedy algorithm: efficient and fastILP: optimal but long calculation timeBeam moving: performance improvementColumn generation for ILPFast heuristics for continuous problemNon-linear integer programming

40Slide41

Multiple carrier models41Slide42

Multiple carrier modelsBinary interferenceCumulative interference42Slide43

Multiple carrier modelsBinary interferenceLF: loading factor

43Slide44

Multiple carrier modelsBinary interferenceA user as a task or an operationUser demand (frequencies) as processing timeInterference pairs as non-overlapping constraintsDisjunctive scheduling problem without precedence constraints

Max. number of scheduled tasks with a

common deadline

44Slide45

Multiple carrier modelsBinary interferenceDisjunctive graph and clique{1,2}, {2,3}, {2,4}, {3,5}, {4,5,6} vs. 7 interference pairs

CP optimizer

45Slide46

Multiple carrier modelsBinary interference46Slide47

Multiple carrier modelsBinary interference47Slide48

Multiple carrier modelsBinary interference48Slide49

Multiple carrier modelsCumulative interferenceOverlapping duration should be considered49Slide50

Multiple carrier modelsCumulative interference: ILP150Slide51

Multiple carrier modelsCumulative interference: ILP251Slide52

Multiple carrier modelsCumulative interference: ILP352Slide53

Multiple carrier modelsScheduling (CP) vs. ILP (CPLEX)53Slide54

Multiple carrier modelsCumulative interference vs. binary interference54Slide55

Multiple carrier modelsCumulative interference vs. binary interference55Slide56

Conclusions and further study – Part 2FAP as scheduling problemOutperform ILPCumulative -> Binary interferencePattern-based ILP with column generationHeuristics based on interval graph coloringLocal search technique

56Slide57

Industrial application57K. Kiatmanaroj, C. Artigues, L. Houssin, and

E. Corbel,

Greedy

algorithms for time-frequency allocation in a SDMA satellite communication system, International conference on Modeling, Optimization and Simulation (MOSIM), 2012Slide58

Industrial applicationTerminal types50 dBW, 45 dBWMax. 24 Mbps, 10 MbpsTraffic typesGuaranteed, Non-guaranteed

User priority level and handling

58Slide59

Industrial applicationSymbol rate - Modulation - Coding scheme (RsModCod)16 ModCod4 symbol rates (Rs) corr. to 5, 10, 15 and 20 MHzSupport bitrate (Mbps)

Different acceptable interference thresholds (alpha)

59Slide60

Industrial applicationBeam positioning methodsFixed-beamSDMA beams60Slide61

Greedy algorithms61Slide62

Greedy algorithmsFastFlexibleExtensive hierarchical searchMI (Minimum Interference)MB (Minimum Bandwidth)No performance guarantee

62Slide63

Greedy algorithms: MIMinimum Interference (MI)Superframe 1

Superframe 2

63

MI

New superframe when the old one is utilized.Slide64

Greedy algorithmsMinimum Bandwidth (MB)64

New superframe before increasing bandwidthSlide65

Experimental results65Slide66

Computational experimentsTest instances66Slide67

Experimental resultsAssignment time (seconds)67

BC longer time than FB

BC30 longer than BC25

MI about the same time as MBSlide68

Experimental resultsNumber of rejected users68

Largely depended on demand / BWSlide69

Conclusions and further study – Part 3Highly complex problem and fast calculation time requirementILP impracticalMI: least interferenceMB: least bandwidthLower bounds on the number of rejected usersLocal search heuristics

69Slide70

Conclusions and further study70Slide71

Conclusions and further studySolved FAP in a satellite communication systemBinary and cumulative interferenceSingle, multiple carrier, realistic modelsGreedy algorithm, ILP, schedulingHyper-heuristicsNon-linear integer programmingColumn generationLocal search: math-heuristics

71Slide72

Thank you72Slide73

Problem definitionFrame structure constraints73Slide74

Experimental results74Slide75

Industrial applicationUser priority level and handling 0 - 3Weighted-Round-Robin ordering75Slide76

Industrial applicationUplink power controlAfter the resource assignmentPCMarginOverall interference reduction76Slide77

Greedy algorithms: MINbS (superframe)m-n (bin configurations) y1-y2 (low – high frequencies) x1-x2 (leftmost – rightmost time bin) Interference calculation repeats* Use control parameters to limit the search space

77Slide78

Experimental resultsNumber of optima for ILPs78Slide79

Experimental resultsFrequency utilization (MHz)79

Note: system maximum bandwidth 300 MHzSlide80

Experimental resultsTotal interference gap80