CN101976290B - Navigation constellation optimization design platform and method based on decomposition thought and particle swarm fusion method - Google Patents

Abstract

The invention discloses navigation constellation optimization design platform and method based on a decomposition thought and a particle swarm fusion method, wherein the platform comprises a man-machine interaction module, a constellation configuration module, a constellation performance module, a constellation cost module, a multi-target integration module, a Combined algorithm module, a constellation performance analyzing module, a visualization module and a report generation module. The method comprises the following steps of: firstly putting forward a navigation constellation design scheme by a user; then setting a performance index and calculating by utilizing a multi-target improved method based on the decomposition thought and particle swarm fusion or multi-method combination; and finally analyzing and displaying a processing result. The method analyzes the characteristics of various constellation designs, overcomes the shortages of the traditional method, optimizes the constellations under the condition of virtual simulation and simultaneously and creatively adopts the decomposition based multi-target improved method to the navigation constellation optimization designs, thereby simplifying the design process, also avoiding the problems of overlong calculation time and partial optimization; in addition, the platform has expandability.

Description

Based on the navigation constellation optimal design platform and the method for decomposing thought and population fusion method

Technical field

The invention belongs to constellation optimal design field, be specifically related to based on the navigation constellation optimal design platform and the method for decomposing thought and population fusion method.

Background technology

Along with the fast development of airmanship, national navigation positioning system is all fallen over each other to study in countries in the world at present, and China is also in " two generations of the Big Dipper " GPS of stepping up to set up oneself.But along with the navigation positioning system globalization covers; Single satellite independent working mode can't have been satisfied and covered requirement over the ground; Navsat must could satisfy the covering requirement with the collaborative work of constellation network's mode, and therefore navigation constellation reasonable in design is the key that realizes global location.Meanwhile; Along with the sequential transmissions of Navsat is used, the constellation scale is increasing, and it is complicated more that the configuration of constellation also becomes thereupon; Not only need consider the various parameter indexs of satellite constellation simultaneously, also will take into account the various performances and the relation of the mutual restriction between the design cost of satellite system.In order to obtain best constellation design effect, must carry out COMPREHENSIVE CALCULATING to the design object of multidimensional such as parameter index, performance requirement and design cost.It is too extensive that but the navigation constellation optimal design relates to content, uses the multidimensional goal approach and assess the cost and consume excessively, simultaneously domesticly lacks a complete design platform and realize the requirement of constellation optimal design.Propose a high performance Multipurpose Optimal Method, and set up the constellation optimal design platform of a system, to design processes simplified, it is crucial improving counting yield.

The navigation constellation optimal design comprises design of navigation constellation configuration and navigation constellation system design.The design of navigation constellation configuration refers to the constellation space geometry of the multiple design criteria of consideration and the optimal design of task track, and the navigation constellation system design refers to the overall system design that comprises constellation configuration parameters and satellite key technical index.

At present, the traditional design method of constellation optimal design mainly is based on methods of numerical, and like linear weighted function method, leash law, simplicial method and complex evolutionary method, these methods have occupied main status in actual engineering and theoretical analysis process.These methods mainly rely on the initial point position, and explore unknown space indifferent is absorbed in local optimum easily; Owing to the relation between optimization variable in the constellation design and the covering performance is blured, can't describe with concrete analytical expression simultaneously, some utilize the optimization method of gradient information to be difficult to utilization.Because hereditary computing method have good optimizing characteristic; In recent years; Chinese scholars releases one after another and is applicable to the classical evolutionary computation method of constellation optimal design, as: NSGA-II, SPEA2, though these methods have solved the indeterminable problem of general optimization method; Obtained very big achievement, consumed big, the too much deficiency of calculating iterations but also exist to assess the cost.

Through above-mentioned analysis; Can find that traditional optimization methods generally all is to handle more effective for the constellation design of low target dimension, low complex degree; Promptly be lower than 6 for objective function; The variable number is lower than 30 constellation and designs, and has certain limitation for the calculating of high target problem of dimension, simultaneously improving constantly along with dimension; Calculation consumption and design cost also increase thereupon, and these target and comprehensive national conditions of country with the globalization of the realization of China's quickening at present navigator fix are inconsistent.The research of constellation optimal design rests in the optimization of method mostly at present in addition, does not make up a complete technology platform, therefore provides a public constellation Optimization Platform to help to save research cost, is convenient to the application of various design philosophys.

Improve one's methods based on the multiple goal of decomposing and to convert multi-objective optimization question into the single goal optimization problem exactly; It is the elementary tactics of finding the solution multi-objective optimization question with mathematic programming methods; Typical conversion method comprises weight and method, Chai Beichefufa, border common factor method etc.; Research in nearly 2 years combines this traditional multiple goal solution strategies and has constructed a kind of multi-target evolution method (as: MOEA/D) based on decomposition of novelty with evolvement method; The PROBLEM DECOMPOSITION that this method will be approached whole Pareto leading surface is the single goal optimization problem of some, finds the solution these single goal optimization problems simultaneously with evolvement method then.Method is kept a current optimal by each subproblem and is separated the population of forming, and the neighbor relationships between the subproblem is defined as the distance between the subproblem weight vectors, the optimizing process of each subproblem through and its neighbour subproblem between evolution operate and accomplish.This method successfully is incorporated into evolution multiple goal field with decomposition method commonly used in the mathematical programming, and the fitness can directly adopt evolvement method to find the solution the single goal optimization problem time distributes with diversity and keeps tactful.

Summary of the invention

The objective of the invention is in order to solve the calculating defective of conventional constellation method for designing in the constellation optimal design problem of high target dimension, configuration complicacy; Realize building the target of unified China's navigation constellation Optimization Platform simultaneously, propose based on the navigation constellation optimal design platform and the method for decomposing thought and population fusion method.

Of the present invention based on the navigation constellation optimal design platform that decomposes thought and population fusion method, comprise human-computer interaction module, constellation configuration module, constellation performance module, constellation cost module, multiple-target integration module, unified algorithm module, constellation analysis-by-synthesis module, visualization model and report generation module.

Human-computer interaction module obtains the user and proposes constellation design and require data, and data is sent out give constellation configuration module, constellation performance module and constellation cost module respectively; Constellation configuration module will go out the configured requirement from the data qualification that human-computer interaction module obtains, and generate constellation configuration model, obtain constellation configuration pattern function analytic expression, and the constellation configuration pattern function analytic expression that generates is sent in the multiple-target integration module; The constellation performance module will go out the performance design requirement from the data qualification that human-computer interaction module obtains; Generate the constellation performance model; Obtain the form of constellation performance model function analytic expression, and the constellation performance model function analytic expression that generates is sent in the multiple-target integration module; Constellation cost module will go out the cost designing requirement from the data qualification that human-computer interaction module obtains, and generate the constellation cost model, obtain constellation cost model function analytic expression, and the constellation cost model function analytic expression that generates is sent in the multiple-target integration module; The model analytic expression that the multiple-target integration module will obtain from constellation configuration module, constellation performance module and constellation cost module; To obtain the model analytic expression respectively and carry out the mathematics simultaneous with the form of objective function; And then be integrated into a column vector that comprises a plurality of objective functions; And ask the optimal solution set of column vector, finally generate constellation optimal design multi-objective problem, and the constellation optimal design multi-objective problem that generates is sent into the unified algorithm module; The unified algorithm module provides different types of multi-target evolution method; The user selects or utilizes the several different methods Combined Treatment that walks abreast according to concrete application; Seek optimal solution, and then obtain the optimization disaggregation, and the optimization disaggregation is sent into constellation analysis-by-synthesis module; Constellation analysis-by-synthesis module receives the compute optimal disaggregation data that sent by the unified algorithm module; And the optimal solution set data are carried out the mathematical statistics analysis; Do not separate with uncorrelated and separate rejecting meeting actual edge, the disaggregation data after will filtering are again sent into visualization model; Visualization model receives the constellation optimal solution set data of constellation analysis-by-synthesis module gained, demonstrates constellation 3D configuration, generates simulated performance testing journal sheet data simultaneously; The report generation module receives the report data that is transmitted by visualization model, selects according to the user, generates corresponding report, accomplishes emulation.

Navigation constellation Optimization Design based on decomposing thought and population fusion method comprises that step is following:

Step 1: set up the constellation mathematical optimization models;

Step 2: set up constellation configuration model;

According to user's requirement, generate constellation configuration pattern function analytic expression;

Step 3: constellation configuration model optimization;

According to the singularity of designing requirement that the user provides, on constellation configuration model based, increase N _AddtionalSupplementary functions satellite, and satellite type separately, N '=N+N are set _Addtional, wherein N ' is the population of satellite;

Step 4: obtain constellation performance model function analytic expression;

The main performance index of constellation performance model function analytic expression comprises: the bearing accuracy factor, covering tuple, percentage of coverage, effective grid point take ratio;

Step 5: obtain constellation cost model function analytic expression;

Step 6: generate multi-objective problem;

Step 7: seek optimal solution, and then obtain the result;

Step 8: the optimal solution set data are carried out mathematical analysis, separate rejecting with part is uncorrelated;

Step 9: the simulated performance test generates corresponding report.

The invention has the advantages that:

(1) not only is applicable to low target dimension, the simple constellation optimal design of configuration, in the constellation optimal design of high target dimension, configuration complicacy, has advantage equally;

When (2) carrying out navigation constellation optimization, only constellation index requirement for restriction need be provided, just can set up navigation constellation and design a model;

(3) when using the multi-method combined calculation, has processing power fast equally;

(4) can carry out aspect assessments such as performance and cost to constellation design, be convenient to the quality that the user judges constellation design;

(5) utilization is carried out the navigation constellation optimal design based on decomposing thought and population fusion method, has initiative.

(6) modelling has extensibility, designability to this platform for navigation constellation.

Description of drawings

Fig. 1 is a platform structure synoptic diagram of the present invention;

Fig. 2 is a method flow diagram of the present invention;

Fig. 3 is the method flow diagram of step 7 of the present invention;

Among the figure:

1. human-computer interaction module 2. constellation configuration modules 3. constellation performance module 4. constellation cost modules

5. multiple-target integration module 6. unified algorithm modules 7. constellation analysis-by-synthesis modules 8. visualization model

9. report generation module

Embodiment

To combine accompanying drawing and embodiment that the present invention is done further detailed description below.

The present invention is a kind of based on the navigation constellation optimal design platform that decomposes thought and population fusion method; The structure of this platform is as shown in Figure 1, comprises human-computer interaction module 1, constellation configuration module 2, constellation performance module 3, constellation cost module 4, multiple-target integration module 5, unified algorithm module 6, constellation analysis-by-synthesis module 7, visualization model 8 and report generation module 9.

Human-computer interaction module 1 is connected respectively with constellation configuration module 2, constellation performance module 3 and constellation cost module 4 respectively, and sends data to constellation configuration module 2, constellation performance module 3 and constellation cost module 4; Human-computer interaction module 1 is the platform terminal interface, can obtain the user simultaneously and propose constellation design and require data, and data sent out give constellation configuration module 2, constellation performance module 3 and constellation cost module 4 respectively.

Constellation configuration module 2, constellation performance module 3 and constellation cost module 4 are connected with multiple-target integration module 5, and send data to multiple-target integration module 5 respectively;

Constellation configuration module 2 will go out the configured requirement from the data qualification that human-computer interaction module 1 obtains, the generation of the planet seat configuration model of going forward side by side.Constellation configuration model comprises the basic configuration of constellation (basic configuration comprises single configuration and combination configuration again) and two parts of additional satellite; Generate constellation configuration pattern function analytic expression, and the constellation configuration pattern function analytic expression that generates is sent in the multiple-target integration module 5.

Constellation performance module 3 will go out performance design requirement, the generation of the planet seat performance model of going forward side by side from the data qualification that human-computer interaction module 1 obtains.The constellation performance model comprises that the bearing accuracy factor, covering tuple, percentage of coverage, effective grid point take than four parts; Generate the form of constellation performance model function analytic expression, and the constellation performance model function analytic expression that generates is sent in the multiple-target integration module 5.

Constellation cost module 4 will go out the cost designing requirement from the data qualification that human-computer interaction module 1 obtains; The go forward side by side generation of planet seat cost model; Generate constellation cost model function analytic expression, and the constellation cost model function analytic expression that generates is sent in the multiple-target integration module 5.

Multiple-target integration module 5 is connected with unified algorithm module 6, and sends data to unified algorithm module 6; The model analytic expression that multiple-target integration module 5 will obtain from constellation configuration module 2, constellation performance module 3 and constellation cost module 4; To obtain the model analytic expression respectively and carry out the mathematics simultaneous with the form of objective function; And then be integrated into a column vector that comprises a plurality of objective functions; And ask the optimal solution set of column vector, finally generate constellation optimal design multi-objective problem, and the constellation optimal design multi-objective problem that generates is sent into unified algorithm module 6.

Unified algorithm module 6 is connected with constellation analysis-by-synthesis module 7, and sends data to constellation analysis-by-synthesis module 7; Unified algorithm module 6 provides different types of multi-target evolution method; The user can select according to concrete application; Also can utilize simultaneously the several different methods Combined Treatment that walks abreast; Seeking optimal solution, and then obtain the optimization disaggregation, and the optimization disaggregation is sent into constellation analysis-by-synthesis module 7.

Constellation analysis-by-synthesis module 7 is connected with visualization model 8, and sends data to visualization model 8.Constellation analysis-by-synthesis module 7 receives the compute optimal disaggregation data that sent by unified algorithm module 6; And the optimal solution set data are carried out the mathematical statistics analysis; Do not separate with uncorrelated and separate rejecting meeting actual edge, the disaggregation data after refiltering are sent into visualization model 8.

Visualization model 8 is connected with report generation module 9, and sends data to report generation module 9; Visualization model 8 receives the constellation optimal solution set data of constellation analysis-by-synthesis module 7 gained; Its data are sent into interface to be connected in the satellite tool box STK software; Demonstrate constellation 3D configuration, use its powerful analysis ability simultaneously, generate simulated performance testing journal sheet data.

Report generation module 9 receives the report data that is transmitted by visualization model 8, selects according to the user, generates corresponding report, accomplishes emulation.

Of the present invention a kind of based on the navigation constellation Optimization Design of decomposing thought and population fusion method, as shown in Figure 2, comprise that step is following:

Step 1: set up the constellation mathematical optimization models

Provide under the prerequisite that constellation design optimization requires data the user, the navigation constellation Optimization Platform generates constellation optimal design multi-objective problem, promptly sets up in the multiple goal constellation optimization problem that complicated constraint condition is being arranged, and its mathematics is described below basically:

min?F(x)＝[f ₁(x)，f ₂(x)，…，f _m(x)] ^T

g _i(x)≤0，i＝1，2，…，p

h _j(x)＝0，j＝1，2，…，q

In the formula, x=(x ₁..., x _i..., x _n), x wherein _i, i ∈ (1,2 ..., n) being known variables, n is the variable number, m is the objective function number, f _m(x) be m objective function, g _i(x)≤0 be the constraint inequality, p is constraint inequality number, h _j(x)=0 be the constraint equality, q is constraint equality number.Objective function F (x) has defined m by the mapping function of decision space to object space; And g _i(x), h _j(x) be constraint condition, it has limited the span of decision vector.In constellation design, the component x of x ₁, x ₂..., x _nThe corresponding base case constellation number of satellite N of difference, orbital plane P, semi-major axis a, eccentric ratio e, orbit altitude h, orbit inclination α, right ascension of ascending node Ω, ascending node angular distance u, argument of perigee ω, mean anomaly β, supplementary functions number of satellite are N _Addtional

Step 2: set up constellation configuration model

The foundation of constellation configuration model is to do the basis for the optimization of next step constellation configuration; Requirement according to the user; Generate constellation configuration pattern function analytic expression; Classical constellation configuration (like Walker constellation and rose constellation) commonly used in several navigation constellation designs is provided in the constellation configuration module, has added popular now combination configuration (GEO+IGEO, geosynchronous satellite and near-synchronous orbit satellite) simultaneously.For practical application, the user also can select the setting of basic configuration voluntarily.

Step 3: constellation configuration model optimization

According to the singularity of designing requirement that the user provides, on constellation configuration model based, increase n supplementary functions satellite, and satellite type separately is set.

N '=N+N _Addtional, wherein N ' is the population of satellite.

Step 4: obtain constellation performance model function analytic expression

Influence several main performance index of navigation constellation overall performance among the present invention, comprising: the bearing accuracy factor, covering tuple, percentage of coverage, effective grid point take ratio.

I. the bearing accuracy factor

GDOP _max(w，j)＝max(GDOP(w，j，t))；

Wherein: (w, j are that the earth surface longitude is j t) to GDOP, and latitude is that the point of w is at t GDOP value constantly, GDOP _Max(w is that the earth surface longitude is j j), and latitude is the GDOP value maximal value of point in whole observation time of w,

Then be illustrated among the selected earth surface zone θ each latitude circle

The mean value of maximum GDOP value in last the having a few, Be the total latitude circle number in the network, Be desired bearing accuracy factor target, σ is a customer requirements

The maximal value of value;

Ii. cover tuple

$N_{\mathrm{cov}}^{\min }=\min N_{\mathrm{cov}}(j,w,t)\≥\mathrm{\ξ};$

Wherein: N _Cov(j, w are that the earth surface longitude is j t), latitude be the point of w at t covering tuple constantly,

Be that the earth surface longitude is j, latitude is the covering tuple minimum value of point in whole observation time of w, and ξ is a customer requirements

The minimum value of value;

Iii. percentage of coverage

$\mathrm{\β}=\arccos \left(\frac{R\cos E}{R+h}\right)-E$

$A_s=4\mathrm{\&pi;}{\mathrm{<figure-callout\; id="2"\; label="R"\; filenames="HDA0000030446040000011.png"\; state="\{\{state\}\}">R</figure-callout>}}^2{\mathrm{<figure-callout\; id="2"\; label="sin"\; filenames="HDA0000030446040000011.png"\; state="\{\{state\}\}">sin</figure-callout>}}^2\frac{\mathrm{\&beta;}}{2}$

$A=\frac{A_s}{A_{\mathrm{earth}}}={\mathrm{<figure-callout\; id="2"\; label="sin"\; filenames="HDA0000030446040000011.png"\; state="\{\{state\}\}">sin</figure-callout>}}^2\frac{\mathrm{\&beta;}}{2}\&times;100\%\&GreaterEqual;\mathrm{\&epsiv;}$

Wherein: β representes the satellite cone of coverage, and R representes earth radius, and h representes the height on satellite distance ground, and E representes minimum view angle, A _sExpression satellite coverage area area, A _EarthThe expression earth surface is long-pending, and A representes that satellite coverage area accounts for the number percent of global area, and ε is the minimum value of the A value of customer requirements.

Iv. the effective grid point takies ratio

Wherein: rel representes that the effective grid point takies and compares minimum value;

is λ for the earth surface longitude; The area summation that the GDOP value of latitude point for

is not more than the effective grid point of μ accounts for the number percent of all net point area summations; K is the minimum value of the GDOP value of user's qualification, and R is the real number set.

Step 5: obtain constellation cost model function analytic expression

Min{C _ICO}＝N·[(C _power，D+C _payload，D+C _bus，D+C _launch)+φ(N)(C _power，T+C _payload，T+C _bus，T)]

+N _addtional·[(C′ _power，D+C _payload，D+C′ _bus，D+C′ _launch)+φ(N _addtional)(C′ _power，T+C′ _payload，T+C′ _bus，T)]

Wherein:

φ(N)＝N ^B，φ(N _addtional)＝N _addtional ^B

$B=1-\frac{\ln (100\%/S)}{\mathrm{<figure-callout\; id="2"\; label="ln"\; filenames="HDA0000030446040000011.png"\; state="\{\{state\}\}">ln</figure-callout>}2}$

Wherein: C _{Power, D}, C _{Payload, D}, C _{Bus, D}, C _LaunchBe respectively power supply, the useful load of constellation basis configuration satellite and remove the operating cost and the launching costs of power supply rear platform; C _{Power, T}, C _{Payload, T}, C _{Bus, T}Be respectively power supply, useful load and the platform building cost of constellation basis configuration satellite; φ (N) is a constellation scale cost multiplier of having considered learning curve; Corresponding C ' _{Power, D}, C ' _{Payload, D}, C ' _{Bus, D}, C ' _LaunchBe respectively power supply, useful load of replenishing satellite and operating cost and the launching costs of removing the power supply rear platform; C ' _{Power, T}, C ' _{Payload, T}, C ' _{Bus, T}Be respectively the power supply, useful load and the platform building cost that replenish satellite.N supplementary functions satellite; N is a number of satellite, and φ (N) makes N satellite cost for considering after the learning curve, and S is the number percent slope of learning curve; When unit number gets 95% less than 10 the time; Unit number gets 90% in the time of 10～15, unit number gets 85% at 15～50 o'clock S, and unit number surpasses at 50 o'clock and gets 80%.

Step 6: generate multi-objective problem

$\min F\left(x\right)={[N^{\&prime;},1/{\mathrm{GDOP}}_{\mathrm{ave}}^{\mathrm{\&theta;}}(w,j),N_{\mathrm{cov}}^{\min },A,1/\mathrm{rel},\mathrm{Min}\{C_{\mathrm{ICO}}\left\}\right]}^T;$

Wherein: N ' is the population of satellite,

Then be illustrated among the selected earth surface zone θ each latitude circle

The mean value of maximum GDOP value in last the having a few,

The earth surface longitude is j, and latitude is that the point of w covers tuple in t minimum constantly, and A representes that satellite coverage area accounts for the number percent of global area, and rel representes

Longitude is λ at the earth's surface, and latitude does

The effective grid point that the GDOP value of point is not more than μ accounts for the number percent of all net points, Min{C _ICOThe navigation constellation least cost.

Step 7: seek optimal solution, and then obtain the result

To generate multi-objective problem and send into unified algorithm module 5, utilization is calculated multi-objective problem based on decomposing thought and population fusion method, and the particular flow sheet of this method is as shown in Figure 3, and following mask body calculation procedure is following:

1. initializing variable

1) the note optimal solution set is EP, and

2) calculate and i T the weight indexed set that weight vector is nearest, wherein indexed set is designated as B (i)={ i ₁..., i _T, note λ ⁱBe i weighted value in the equally distributed N weight vector, i ∈ [1, N],

Be λ ⁱT nearest weighted value, N is for based on the number that decomposes the subproblem of considering in thought and the population fusion method, T be the quantity of the weight vector nearest apart from whenever single weight vector;

3) produce initial population at random and be designated as x ¹..., x ^N, and make the value of separating of the corresponding target of each population be F ⁱ=F (x ⁱ), i ∈ [1, N] wherein;

4) note cycle index t=0, preestablishing cycle index is t _p

2. the population method is found out the single goal Function Optimization and is separated

Utilize the population method to solve each objective function f that a step decomposites _i(x) temporary transient optimum solution z _i, initialization optimal solution set z=(z ₁..., z _m) ^T

3. upgrade the EP disaggregation

1) from weight vector B (i), select two indexs at random, be designated as k respectively, l uses genetic operator from x ^kAnd x ^lThe middle new explanation y that produces;

2) improve: the quality according to gained new explanation y after the heredity is made amendment to new explanation y, if f _i(y)＞f _i(x ^k) and f _i(y)＞f _i(x ^l), y '=y then; If f _i(y)≤max (f _i(x ^k), f _i(x ^l)), and f _i(x ^k)>=f _i(x ^l) y '=x then ^kIf f _i(y)≤max (f _i(x ^k), f _i(x ^l)), and f _i(x ^l)>=f _i(x ^k) y '=x then ^lY ' is separating after improving;

3) upgrade z: to any j=1 ..., if m is z _j＜f _j(y '), then assignment z _j=f _j(y '), z _jBe any optimum solution;

4) upgrading adjacent problem separates: to j ∈ B (i), if g ^Te(y ' | λ ^j, z)≤g ^Te(x ^j| λ ^j, z), arbitrary initial population x then ^j=y ', F (x ^j)=F (y '); Wherein defining RP is z _iThe target function value of j subproblem do

λ wherein ^jBe weight vector in the equally distributed weight vector group.

5) upgrade optimal solution set EP: deletion is modified the vector of back target function value F (y ') domination from EP.If the vector of target function value F (y ') then will not improve back target function value F (y ') and add EP after domination improves among the EP.

6) cycle index t=t+1;

4. stop condition is judged

Preestablish cycle index and promptly work as t=t if satisfy _pThe time, then stop circulation, change 5 over to; Otherwise change 3 over to.

5. output optimal solution set

After finishing to calculate, with of the form output of the optimal solution set of calculating gained with matrix.

Step 8: the optimal solution set data are carried out mathematical analysis, separate rejecting with part is uncorrelated

According to the target function value scope that the user proposes, rejecting is separated at the uncorrelated edge of disaggregation, and then reject the satellite orbit that has been taken by existing satellite.Separate concentratedly remaining, choose and have the orbit altitude that returns characteristic, promptly operate in D days around earth N circle, wherein the span of D is 2～10, and the span of N is 4～20.Last disaggregation is sent into visualization model 8.

Step 9: the simulated performance test generates corresponding report

With each single the separating in the optimal solution set that receives, to send into respectively in the satellite tool box STK software, simulation generates respective constellation; Utilization STK software covering analyzing module is to this constellation function; The gained data are sent into report generation module 9 generate respective list, finally present to the user, end operation.

Platform according to the invention can merge present various countries navigation constellation design characteristic, has considered the singularity of navigation constellation design simultaneously emphatically, therefore has meaning widely with respect to traditional function property constellation design; The utilization of novelty simultaneously applies in the navigation constellation optimal design based on the multi-target method that decomposes; And specific aim incorporates this method with the population method; Proposed based on decomposing thought and population fusion method; This method also be not applied in the navigation constellation optimization field before this, so the development of constellation design is had initiative meaning.Platform also provides the function of multi-method combined calculation in addition; The processing procedure that the method for system of selection is concrete can be different, and diverse ways also can be different to the treatment effect of different multi-objective problems simultaneously, therefore selects the method for combined calculation; With the multi-objective optimization question parallel processing; Remedying the shortcoming of this respect greatly, though on time and processing cost, increase to some extent, still is helpful for the exploitation of unknown constellation design problem; Effectively saved the scientific research cycle, for the user provides diversified selection.System of the present invention both had been applicable to the processing of traditional constellation optimization problem; Just lower constellation optimal design to configuration simple target dimension; More be applicable to the processing of the navigation constellation optimal design problem of present China, promptly to the higher constellation optimal design of the complicated high target dimension of configuration; Can carry out systematization assessment and visual reproduction according to result simultaneously.

Described navigation constellation optimal design platform based on decomposition thought and population fusion method comprises two kinds of mode of operations:

A kind of is to utilize based on decomposing thought and population fusion method directly to carry out the navigation constellation optimal design; At first the user proposes requirement for restriction to each index of constellation in human-computer interaction interface; It is required to send into respectively constellation configuration module, constellation performance module and constellation cost module; Generate constellation configuration model, constellation performance model and constellation cost model, and will generate constellation configuration model, constellation performance model and constellation cost model data and send into the multiple-target integration module; After getting into the multiple-target integration module, the utilization mathematical method is integrated constellation configuration model, constellation performance model and constellation cost model and is generated constellation optimization problem model, and the constellation optimization problem model after will generating is sent into the unified algorithm module; After getting into the unified algorithm module; Utilization is carried out mathematical computations based on decomposing thought and population fusion method; And result of calculation sent into constellation analysis-by-synthesis module; Send into visual the result who analyzes and the report generation module, the constellation 3D configuration of last visual and report generation module after with optimization shows, and on man-machine interface, demonstrates the form of constellation analysis-by-synthesis.

A kind of in addition user of being can take the mode of multi-method combined calculation that navigation constellation is optimized design; In human-computer interaction interface, requirement for restriction is proposed each index of constellation equally; It is required to send into respectively constellation configuration module, constellation performance module and constellation cost module; Generate constellation configuration model, constellation performance model and constellation cost model, and will generate constellation configuration model, constellation performance model and constellation cost model data and send into the multiple-target integration module; After getting into the multiple-target integration module, the utilization mathematical method is integrated constellation configuration model, constellation performance model and constellation cost model and is generated constellation optimization problem model, and the constellation optimization problem model after will generating is sent into the unified algorithm module.Different is after the constellation optimization problem model that generates is sent in the unified algorithm module; Each method in the unified algorithm module is carried out parallel computation to problem model; The optimal solution set that several different methods is obtained merges; Final public optimal solution set, the last visual and report generation module constellation 3D configuration after with optimization shows, and on man-machine interface, demonstrates the form of constellation analysis-by-synthesis.

Claims (7)

1. one kind based on the navigation constellation optimal design device that decomposes thought and population fusion method; It is characterized in that, comprise human-computer interaction module, constellation configuration module, constellation performance module, constellation cost module, multiple-target integration module, unified algorithm module, constellation analysis-by-synthesis module, visualization model and report generation module;

Human-computer interaction module obtains the user and proposes constellation design requirement data, and data are sent to constellation configuration module, constellation performance module and constellation cost module respectively;

Constellation configuration module will go out the configured requirement from the data qualification that human-computer interaction module obtains, and generate constellation configuration model, obtain constellation configuration pattern function analytic expression, and the constellation configuration pattern function analytic expression that generates is sent in the multiple-target integration module;

The constellation performance module will go out the performance design requirement from the data qualification that human-computer interaction module obtains; Generate the constellation performance model; Obtain the form of constellation performance model function analytic expression, and the constellation performance model function analytic expression that generates is sent in the multiple-target integration module;

Described constellation performance model comprises that the bearing accuracy factor, covering tuple, percentage of coverage, effective grid point take ratio;

The bearing accuracy factor is specially:

GDOP _max(w,j)＝max(GDOP(w,j,t))；

Wherein: (w, j are that the earth surface longitude is j t) to GDOP, and latitude is that the point of w is at t GDOP value constantly, GDOP _Max(w is that the earth surface longitude is j j), and latitude is the GDOP value maximal value of point in whole observation time of w,

Then be illustrated among the selected earth surface zone θ each latitude circle The mean value of maximum GDOP value in last the having a few,

Be the total latitude circle number in the network, Be desired bearing accuracy factor target, σ is a customer requirements

The maximal value of value;

The covering tuple is specially:

$N_{\mathrm{cov}}^{\min }=\min N_{\mathrm{cov}}(j,w,t)\&GreaterEqual;\mathrm{\&xi;};$

Wherein: N _Cov(j, w are that the earth surface longitude is j t), latitude be the point of w at t covering tuple constantly,

Be that the earth surface longitude is j, latitude is the covering tuple minimum value of point in whole observation time of w, and ξ is a customer requirements

The minimum value of value;

Percentage of coverage is specially:

$\mathrm{\&beta;}=\arccos \left(\frac{R\cos E}{R+h}\right)-E$

$A_s=4\mathrm{\&pi;}{R}^2{\sin }^2\frac{\mathrm{\&beta;}}{2}$

$A=\frac{A_s}{A_{\mathrm{earth}}}={\sin }^2\frac{\mathrm{\&beta;}}{2}\&times;100\%\&GreaterEqual;\mathrm{\&epsiv;}$

Wherein: β representes the satellite cone of coverage, and R representes earth radius, and h representes the height on satellite distance ground, and E representes minimum view angle, A _sExpression satellite coverage area area, A _EarthThe expression earth surface is long-pending, and A representes that satellite coverage area accounts for the number percent of global area, and ε is the minimum value of the A value of customer requirements;

The effective grid point takies than is specially:

Wherein: rel representes that the effective grid point takies and compares minimum value;

is λ for the earth surface longitude; The area summation that the GDOP value of latitude point for

is not more than the effective grid point of μ accounts for the number percent of all net point area summations; K is the minimum value of the GDOP value of user's qualification, and R is the real number set;

Constellation cost module will go out the cost designing requirement from the data qualification that human-computer interaction module obtains, and generate the constellation cost model, obtain constellation cost model function analytic expression, and the constellation cost model function analytic expression that generates is sent in the multiple-target integration module; Described constellation cost model function analytic expression is:

Min{C _ICO}＝N·[(C _power，D+C _payload，D+C _bus,D+C _launch)+φ(N)(C _power，T+C _payload,T+C _bus，T)]+N _addtional·[(C′ _power，D+C′ _payload,D+C′ _bus,D+C′ _launch)+φ(N _addtional)(C′ _power，T+C′ _payload,T+C′ _bus,T)]

Wherein:

φ(N)＝N ^B,φ(N _addtional)＝N _addtional ^B

$B=1-\frac{\ln (100\%/S)}{\ln 2}$

Wherein: C _{Power, D}, C _{Payload, D}, C _{Bus, D}, C _LaunchBe respectively the operating cost of the power supply of constellation basis configuration satellite, the operating cost of useful load, the operating cost of removing the power supply rear platform, launching costs, C _{Power, T}, C _{Payload, T}, C _{Bus, T}The power supply that is respectively constellation basis configuration satellite is built cost, useful load is built cost, platform building cost; φ (N) is a constellation scale cost multiplier of having considered learning curve; Corresponding C ' _{Power, D}, C ' _{Payload, D}, C ' _{Bus, D}, C ' _LaunchBe respectively the operating cost of the power supply of supplementary functions satellite, the operating cost of useful load, the operating cost of removing the power supply rear platform, launching costs; C ' _{Power, T}, C ' _{Payload, T}, C ' _{Bus, T}The power supply that is respectively the supplementary functions satellite is built cost, useful load is built cost, platform building cost, N _AddtionalSupplementary functions satellite, N is the base case constellation number of satellite, and φ (N) makes N satellite cost for considering after the learning curve, and S is the number percent slope of learning curve;

The multiple-target integration module obtains the model analytic expression from constellation configuration module, constellation performance module and constellation cost module; The model analytic expression that obtains is carried out the mathematics simultaneous with the form of objective function; And then be integrated into a column vector that comprises a plurality of objective functions; And ask the optimal solution set of column vector, finally generate constellation optimal design multi-objective problem, and the constellation optimal design multi-objective problem that generates is sent into the unified algorithm module; Described constellation optimal design multi-objective problem is specially:

$\min F\left(x\right)={[N^{\&prime;},1/{\mathrm{GDOP}}_{\mathrm{ave}}^{\mathrm{\&theta;}}(w,j),N_{\mathrm{cov}}^{\min },A,1/\mathrm{rel},\mathrm{Min}\{C_{\mathrm{ICO}}\left\}\right]}^T;$

Wherein: N ' is the population of satellite,

Then be illustrated among the selected earth surface zone θ each latitude circle

The mean value of maximum GDOP value in last the having a few,

The earth surface longitude is j, and latitude is that the point of w covers tuple in t minimum constantly, and A representes that satellite coverage area accounts for the number percent of global area, and rel representes Longitude is λ at the earth's surface, and latitude does

The effective grid point that the GDOP value of point is not more than μ accounts for the number percent of all net points, Min{C _ICOIt is the navigation constellation least cost;

The unified algorithm module provides different types of multi-target evolution method; The user selects or utilizes the several different methods Combined Treatment that walks abreast according to concrete application; Seek optimal solution, and then obtain the optimization disaggregation, and the optimization disaggregation is sent into constellation analysis-by-synthesis module;

Wherein, the multi-target evolution method is specially:

Initializing variable: the note optimal solution set is EP, and

Calculate and i T the weight indexed set that weight vector is nearest, wherein indexed set is designated as B (i)={ i ₁... I _T,, note λ ⁱBe i weighted value in the equally distributed K weight vector, i ∈ [1, K],

Be λ ⁱT nearest weighted value, K is for based on the number that decomposes the subproblem of considering in thought and the population fusion method, T be the quantity of the weight vector nearest apart from whenever single weight vector; Produce initial population at random and be designated as x ¹..., x ^K, and make the value of separating of the corresponding target of each population be F ⁱ=F (x ⁱ), i ∈ [1, K] wherein; Note cycle index t=0, preestablishing cycle index is t _p

The population method is found out the single goal Function Optimization and is separated: utilize the population method to solve each objective function f that a step decomposites _i(x) temporary transient optimum solution z _i, initialization optimal solution set z=(z ₁..., z _m) ^T

Upgrade the EP disaggregation: from weight vector B (i), select two indexs at random, be designated as k respectively, l uses genetic operator from x ^kAnd x ^lThe middle new explanation y that produces; Improve: the quality according to gained new explanation y after the heredity is made amendment to new explanation y, if f _i(y)＞f _i(x ^k) and f _i(y)＞f _i(x ^l), y '=y then; If f _i(y)≤max (f _i(x ^k), f _i(x ^l)), and f _i(x ^k)>=f _i(x ^l) y '=x then ^kIf f _i(y)≤max (f _i(x ^k), f _i(x ^l)), and f _i(x ^l)>=f _i(x ^k) y '=x then ^lY ' is separating after improving; Upgrade z: to any j=1 ..., if m is z _j＜f _j(y '), then assignment z _j=f _j(y '), z _jBe any optimum solution; Upgrading adjacent problem separates: to j ∈ B (i), if g ^Te(y ' | λ j, z)≤g ^Te(x ^j| λ ^j, z), arbitrary initial population x then ^j=y ', F (x ^j)=F (y '); Wherein defining RP is z _iThe target function value of j subproblem do

λ wherein _jBe weight vector in the equally distributed weight vector group; Upgrade optimal solution set EP: deletion is modified the vector of back target function value F (y ') domination from EP; If the vector of target function value F (y ') then will not improve back target function value F (y ') and add EP after domination improves among the EP; Cycle index t=t+1;

Stop condition is judged: preestablish cycle index and promptly work as t=t if satisfy _pThe time, then stop circulation, the output optimal solution set; Otherwise continue to upgrade the EP disaggregation;

Output optimal solution set: after finishing to calculate, with of the form output of the optimal solution set of calculating gained with matrix;

Constellation analysis-by-synthesis module receives the compute optimal disaggregation data that sent by the unified algorithm module; And the optimal solution set data are carried out the mathematical statistics analysis; Do not separate with uncorrelated and separate rejecting meeting actual edge, the disaggregation data after will filtering are again sent into visualization model;

Visualization model receives the constellation optimal solution set data of constellation analysis-by-synthesis module gained, demonstrates constellation 3D configuration, generates simulated performance testing journal sheet data simultaneously; The report generation module receives the report data that is transmitted by visualization model, selects according to the user, generates corresponding report, accomplishes emulation.

2. according to claim 1 a kind of based on the navigation constellation optimal design device that decomposes thought and population fusion method; It is characterized in that; Described constellation configuration model comprises the basic configuration of constellation and two parts of supplementary functions satellite, and the basic configuration of constellation comprises single configuration and combination configuration.

3. according to claim 1 a kind of based on the navigation constellation optimal design device that decomposes thought and population fusion method; It is characterized in that; Described visualization model is sent the optimal solution set data into interface and is connected in the satellite tool box STK software; Demonstrate constellation 3D configuration, generate simulated performance testing journal sheet data.

4. the navigation constellation Optimization Design based on decomposition thought and population fusion method is characterized in that, comprises that step is following:

Step 1: set up the constellation mathematical optimization models

Provide under the prerequisite that constellation design optimization requires data the user, generate constellation optimal design multi-objective problem, promptly establish the multiple goal constellation optimization problem of complicated constraint condition, its mathematics is described below basically:

minF(x)＝[f ₁(x),f ₂(x),…,f _m(x)] ^T

g _i(x)≤0,i＝1，2,…,p

h _j(x)＝0,j＝1,2,…,q

In the formula, x=(x ₁..., x _i..., x _n), x wherein _iBe known variables, i ∈ (1,2 ..., n), n is the variable number, m is the objective function number, f _m(x) be m objective function, g _i(x)≤0 be the constraint inequality, p is constraint inequality number, h _j(x)=0 be the constraint equality, q is constraint equality number, and objective function F (x) has defined m by the mapping function of decision space to object space, and g _i(x), h _j(x) be constraint condition, be used to limit the span of decision vector; The component x of x ₁, x ₂..., x _nThe corresponding base case constellation number of satellite N of difference, orbital plane P, semi-major axis a, eccentric ratio e, orbit altitude h, orbit inclination α, right ascension of ascending node Ω, ascending node angular distance u, argument of perigee ω, mean anomaly β, supplementary functions number of satellite N _Addtional

Step 2: set up constellation configuration model

According to user's requirement, generate constellation configuration pattern function analytic expression;

Step 3: constellation configuration model optimization

According to the singularity of designing requirement that the user provides, on constellation configuration model based, increase N _AddtionalSupplementary functions satellite, and satellite type separately, N '=N+N are set _Addtional, wherein N ' is the population of satellite;

Step 4: obtain constellation performance model function analytic expression

The performance index of constellation performance model function analytic expression comprise: the bearing accuracy factor, covering tuple, percentage of coverage and effective grid point take ratio;

I. the bearing accuracy factor

GDOP _max(w,j)＝max(GDOP(w,j,t))；

Wherein: (w, j are that the earth surface longitude is j t) to GDOP, and latitude is that the point of w is at t GDOP value constantly, GDOP _Max(w is that the earth surface longitude is j j), and latitude is the GDOP value maximal value of point in whole observation time of w,

Then be illustrated among the selected earth surface zone θ each latitude circle

The mean value of maximum GDOP value in last the having a few,

Be the total latitude circle number in the network,

Be desired bearing accuracy factor target, σ is a customer requirements

The maximal value of value;

Ii. cover tuple

$N_{\mathrm{cov}}^{\min }=\min N_{\mathrm{cov}}(j,w,t)\&GreaterEqual;\mathrm{\&xi;};$

Wherein: N _Cov(j, w are that the earth surface longitude is j t), latitude be the point of w at t covering tuple constantly,

Be that the earth surface longitude is j, latitude is the covering tuple minimum value of point in whole observation time of w, and ξ is a customer requirements

The minimum value of value;

Iii. percentage of coverage A

$\mathrm{\&beta;}=\arccos \left(\frac{R\cos E}{R+h}\right)-E$

$A_s=4\mathrm{\&pi;}{R}^2{\sin }^2\frac{\mathrm{\&beta;}}{2}$

$A=\frac{A_s}{A_{\mathrm{earth}}}={\sin }^2\frac{\mathrm{\&beta;}}{2}\&times;100\%\&GreaterEqual;\mathrm{\&epsiv;}$

Wherein: β representes the satellite cone of coverage, and R representes earth radius, and h representes the height on satellite distance ground, and E representes minimum view angle, A _sExpression satellite coverage area area, A _EarthThe expression earth surface is long-pending, and A representes that satellite coverage area accounts for the number percent of global area, and ε is the minimum value of the A value of customer requirements;

Iv. the effective grid point takies ratio

Wherein: rel representes that the effective grid point takies and compares minimum value;

is λ for the earth surface longitude; The area summation that the GDOP value of latitude point for is not more than the effective grid point of μ accounts for the number percent of all net point area summations; K is the minimum value of the GDOP value of user's qualification, and R is the real number set;

Step 5: obtain constellation cost model function analytic expression

Min{C _ICO}＝N·[(C _power，D+C _payload,D+C _bus,D+C _launch)+φ(N)(C _power，T+C _payload,T+C _bus，T)]+N _addtional·[(C′ _power,D+C′ _payload,D+C′ _bus,D+C′ _launch)+φ(N _addtional)(C′ _power,T+C′ _payload,T+C′ _bus,T)]

Wherein:

φ(N)＝N ^B,φ(N _addtional)＝N _addtional ^B

$B=1-\frac{\ln (100\%/S)}{\ln 2}$

Wherein: C _{Power, D}, C _{Payload, D}, C _{Bus, D}, C _LaunchBe respectively the operating cost of the power supply of constellation basis configuration satellite, the operating cost of useful load, the operating cost of removing the power supply rear platform, launching costs, C _{Power, T}, C _{Payload, T}, C _{Bus, T}The power supply that is respectively constellation basis configuration satellite is built cost, useful load is built cost, platform building cost; φ (N) is a constellation scale cost multiplier of having considered learning curve; Corresponding C ' _{Power, D}, C ' _{Payload, D}, C ' _{Bus, D}, C ' _LaunchBe respectively the operating cost of the power supply of supplementary functions satellite, the operating cost of useful load, the operating cost of removing the power supply rear platform, launching costs; C ' _{Power, T}, C ' _{Payload, T}, C ' _{Bus, T}The power supply that is respectively the supplementary functions satellite is built cost, useful load is built cost, platform building cost, N _AddtionalSupplementary functions satellite, N is the base case constellation number of satellite, and φ (N) makes N satellite cost for considering after the learning curve, and S is the number percent slope of learning curve;

Step 6: generate multi-objective problem

$\min F\left(x\right)={[N^{\&prime;},1/{\mathrm{GDOP}}_{\mathrm{ave}}^{\mathrm{\&theta;}}(w,j),N_{\mathrm{cov}}^{\min },A,1/\mathrm{rel},\mathrm{Min}\{C_{\mathrm{ICO}}\left\}\right]}^T;$

Wherein: N ' is the population of satellite,

Then be illustrated among the selected earth surface zone θ each latitude circle

The mean value of maximum GDOP value in last the having a few, The earth surface longitude is j, and latitude is that the point of w covers tuple in t minimum constantly, and A representes that satellite coverage area accounts for the number percent of global area, and rel representes

Longitude is λ at the earth's surface, and latitude does The effective grid point that the GDOP value of point is not more than μ accounts for the number percent of all net points, Min{C _ICOIt is the navigation constellation least cost;

Step 7: seek optimal solution, and then obtain the result

To generate multi-objective problem and send into the unified algorithm module, utilization is calculated multi-objective problem based on decomposing thought and population fusion method, and following mask body calculation procedure is following:

(1) initializing variable

1) the note optimal solution set is EP, and

2) calculate and i T the weight indexed set that weight vector is nearest, wherein indexed set is designated as B (i)={ i ₁..., i _T, note λ ⁱBe i weighted value in the equally distributed K weight vector, i ∈ [1, K],

Be λ ⁱT nearest weighted value, K is for based on the number that decomposes the subproblem of considering in thought and the population fusion method, T be the quantity of the weight vector nearest apart from whenever single weight vector;

3) produce initial population at random and be designated as x ¹..., x ^K, and make the value of separating of the corresponding target of each population be F ⁱ=F (x ⁱ), i ∈ [1, K] wherein;

4) note cycle index t=0, preestablishing cycle index is t _p

(2) the population method is found out the single goal Function Optimization and is separated

Utilize the population method to solve each objective function f that a step decomposites _i(x) temporary transient optimum solution z _i, initialization optimal solution set z=(z ₁..., z _m) ^T

(3) upgrade the EP disaggregation

1) from weight vector B (i), select two indexs at random, be designated as k respectively, l uses genetic operator from x ^kAnd x ^lThe middle new explanation y that produces;

2) improve: the quality according to gained new explanation y after the heredity is made amendment to new explanation y, if f _i(y)＞f _i(x ^k) and f _i(y)＞f _i(xl), y '=y then; If f _i(y)≤max (f _i(x ^k), f _i(x ^l)), and f _i(x ^k)>=f _i(x ^l) y '=x then ^k

If f _i(y)≤max (f _i(x ^k), f _i(x ^l)), and f _i(x ^l)>=f _i(x ^k) y '=x then ^lY ' is separating after improving;

3) upgrade z: to any j=1 ..., if m is z _j＜f _j(y '), then assignment z _j=f _j(y '), z _jBe any optimum solution;

4) upgrading adjacent problem separates: to j ∈ B (i), if g ^Te(y ' | λ ^j, z)≤g ^Te(x ^j| λ ^j, z), arbitrary initial population x then ^j=y ', F (x ^j)=F (y '); Wherein defining RP is z _iThe target function value of j subproblem do $g^{\mathrm{Te}}\left(x\right|{\mathrm{\&lambda;}}^j,z)=\underset{1\&le;i\&le;m}{\mathrm{Max}}\left\{{\mathrm{\&lambda;}}_i^j\right|f_i\left(x\right)-z_i\left|\right\};$ λ wherein ^jBe weight vector in the equally distributed weight vector group;

5) upgrade optimal solution set EP: deletion is modified the vector of back target function value F (y ') domination from EP; If the vector of target function value F (y ') then will not improve back target function value F (y ') and add EP after domination improves among the EP;

6) cycle index t=t+1;

(4) stop condition is judged

Preestablish cycle index and promptly work as t=t if satisfy _pThe time, then stop circulation, change step (5) over to; Otherwise change step (3) over to;

(5) output optimal solution set

After finishing to calculate, with of the form output of the optimal solution set of calculating gained with matrix;

Step 8: the optimal solution set data are carried out mathematical analysis, separate rejecting with part is uncorrelated

According to the target function value scope that the user proposes, rejecting is separated at the uncorrelated edge of disaggregation, and then reject the satellite orbit that has been taken by existing satellite; Separate concentratedly remaining, choose and have the orbit altitude that returns characteristic, promptly operate in D days, last disaggregation is sent into visualization model around earth Q circle;

Step 9: the simulated performance test generates corresponding report

According to each single the separating in the optimal solution set that receives, simulation generates respective constellation, and simulated performance testing journal sheet data are sent into the report generation module, generates respective list, finally presents to the user, end operation.

5. according to claim 4 a kind of based on the navigation constellation Optimization Design of decomposing thought and population fusion method; It is characterized in that; Get 95% as satellite number S less than 10 time in the described step 5; Satellite number S in the time of 10 ~ 15 gets 90%, and the satellite number gets 85% at 15 ~ 50 o'clock S, and the satellite number surpasses 50 o'clock S and gets 80%.

6. a kind of navigation constellation Optimization Design based on decomposition thought and population fusion method according to claim 4 is characterized in that in the described step 8, the span of D is 2～10, and the span of Q is 4～20.

7. according to claim 4 a kind of based on the navigation constellation Optimization Design of decomposing thought and population fusion method, it is characterized in that, in the described step 9; Visualization model is with each single the separating in the optimal solution set that receives; Send into respectively in the satellite tool box STK software, simulation generates respective constellation, and utilization STK software covering analyzing module is to this constellation function; Generate simulated performance testing journal sheet data, and the gained data are sent into the report generation module.

Images