Method, device and storage medium for allocating beam resources between low-orbit constellations

Abstract

The present invention discloses a method, device and storage medium for allocating beam resources between low-orbit constellations. The method includes: a population initialization step, a coloring scheme fitness calculation and a best individual storage step, a roulette selection step, a crossover operation step, a first local search step, a second local search step, a mutation operation step, a population merging step, and an iteration termination judgment and result output step. The present invention constructs a beam spectrum resource allocation graph model between low-orbit constellations, designs a sub-constellation wave position graph model coloring scheme optimization in the overlapping coverage area between constellations based on a genetic algorithm, uses a mode combining global convergence and local search, proposes a double-layer local search method based on four-color principle compliance and based on reducing inter-constellation interference, accelerates the convergence speed, and improves the optimization efficiency of the sub-constellation beam resource allocation optimal solution under the main constellation interference constraint condition.

Description

Technical Field

[0001] The present invention relates to the field of satellite communication technology, and in particular to a method, device and storage medium for allocating beam resources between low-orbit constellations. Background Art

[0002] As the global demand for satellite communications continues to grow, more and more countries and companies are investing in the construction of low-orbit satellite constellations. However, due to the limited orbital space around the earth, especially in the low-orbit region, it is difficult for satellites in each constellation to completely avoid overlapping coverage when operating. The main challenge brought by overlapping coverage is that it may cause signal interference and conflict, affecting communication quality and network performance. How to ensure that the beams of each low-orbit constellation with overlapping coverage can provide services normally while minimizing interference has become an urgent problem to be solved.

[0003] In actual operation, if a new constellation needs to be deployed, it is necessary to actively negotiate with the constellation already in orbit according to the relevant coordination provisions of Article 9 of the ITU Radio Regulations. Only after the negotiation is successful can the new constellation be actually deployed and provide services. That is, the constellation already in orbit is in a negotiated position and has a higher priority than the new constellation. The constellation already in orbit can be regarded as the primary constellation and the new constellation can be regarded as the secondary constellation. Figure 1 As shown in the figure, there is an overlap in the service coverage area between the secondary constellation and the main constellation. The overlapping area contains several fixed beams, and the fixed beam coverage service area division of the secondary constellation satellite and the main constellation satellite in the overlapping area is consistent. On the basis of the existing main constellation in orbit, in order to accommodate the secondary constellation, it is necessary to consider the beam allocation of the secondary constellation in the service overlapping area, so as to achieve the purpose of minimizing interference to the main constellation and ensuring the normal operation of the main constellation service, so that the secondary constellation and the main constellation can coexist and provide services.

[0004] like Figure 2 As shown in the figure, the allocation of beam spectrum resources between the main and secondary constellations is equivalent to the coloring problem of the graph model based on the double-layer wave position diagram. When the main constellation beam allocation is determined, that is, the coloring of the main constellation wave position diagram in the overlapping coverage area between constellations is known, and the coloring of the main constellation wave position diagram conforms to the different colors of adjacent wave positions, the coloring scheme of the secondary constellation wave position diagram in the overlapping coverage area between constellations is optimized to obtain the best coloring scheme, which ensures that the communication needs of each wave position of the main constellation satellite are not affected by the secondary constellation, and each adjacent wave position of the secondary constellation is covered by beams of different frequency bands, and the communication needs of the secondary constellation satellite are maximized, thereby improving the spectrum utilization rate in the overlapping area of ​​the main and secondary constellation satellites. Summary of the invention

[0005] The present invention aims to solve one of the technical problems in the related art to at least a certain extent. To this end, one object of the present invention is to provide a method, device and storage medium for allocating beam resources between low-orbit constellations, which can reasonably and efficiently allocate beam spectrum resources between low-orbit constellations, so that the secondary constellation can share spectrum resources to the maximum extent without interfering with the primary constellation.

[0006] The technical solution adopted by the present invention is:

[0007] In a first aspect, the present invention provides a method for allocating beam resources in overlapping areas between low-orbit constellations, the method comprising the following steps: S11: Initializing a population: constructing a beam spectrum resource allocation graph model between low-orbit constellations according to known parameters of the low-orbit constellations, recording the number of wave positions in the overlapping areas between constellations as M, and for a sub-constellation wave position graph model G in the overlapping coverage area between constellations, randomly generating an initial coloring scheme set color_set with a population size of N that conforms to the four-color principle, wherein color_set is a two-dimensional array, each row corresponds to a coloring scheme, and each column corresponds to one of the M vertices in the graph model G, that is, color_set (x, y) represents the color of the yth wave position in the xth scheme, wherein 1≤x≤N, 1≤y≤M; S12: Calculating the fitness of each coloring scheme in the population, and saving the best individual separately: according to the known coloring of the main constellation wave position graph model in the overlapping coverage area between constellations, the fitness value corresponding to the coloring scheme of each sub-constellation wave position graph model in color_set is calculated by the fitness calculation formula, and the fitness value corresponding to the xth coloring scheme is , 1≤x≤N, and then use the elite mechanism to save the best individual; S13: Use the roulette wheel selection operator to select the coloring scheme as the next generation parent; S14: Use the intersection crossover operator to realize the generation of the offspring coloring scheme after the crossover operation according to the crossover rate; S15: Based on the local search of the four-color principle compliance, improve the crossover result and obtain a new offspring coloring scheme; S16: Based on the local search to reduce interference, further improve the crossover result and obtain a new offspring coloring scheme; S17: Use the mutation operator to realize the generation of the offspring coloring scheme after the mutation operation according to the mutation rate; S18: Merge the offspring population and form a new generation population based on the fitness calculation result; S19: Determine whether the current number of iterations has reached the maximum number of iterations. If it has not reached the maximum number of iterations, replace the color_set with the new generation population and jump to the step S12. Otherwise, if it has reached the maximum number of iterations, find the coloring scheme with the smallest fitness value from the new generation population, which is the optimal coloring scheme and output it as the result.

[0008] The fitness calculation formula is as follows: ,in, iA coloring scheme for the sub-constellation wave pattern model within the overlapping coverage area between constellations, For the plan i The corresponding fitness value is is the penalty factor for the violation of the four-color principle by the internal wave position of the sub-constellation satellite, which is set to 10000; The number of times the four-color principle is violated in a coloring scheme of a constellation wave diagram model; is the penalty factor for the same-wavelength-position-and-color interference between constellations, which is set to 100; is the number of times the same color interference exists in the same wave position between constellations; is the penalty factor for the presence of same-color interference in adjacent wave positions between constellations, set to 1; is the number of times the same-color interference exists in adjacent wave positions between constellations; The penalty factor for the failure of the secondary constellation to provide service is set to 0.0001; It is the sum of spectrum requirements corresponding to each closed service wave position in the secondary constellation.

[0009] Wherein, step S13 includes: the probability of the xth coloring scheme being selected is , the cumulative probability of the xth coloring solution , the process of selecting a coloring scheme is: generate a uniformly distributed random number r1 in the interval [0,1], compare it with the cumulative probability of each scheme in turn, and select scheme t so that True; repeat the above selection process, select N coloring schemes from color_set as the next generation parents, and store them row by row into the two-dimensional array selected_set.

[0010] Among them, step S14 includes: traversing each coloring scheme in the two-dimensional array selected_set, judging whether each coloring scheme needs to be crossover as a parent, if not, directly copying the parent to the next generation, if necessary, randomly selecting another coloring scheme from the two-dimensional array selected_set as another parent, and generating a next generation coloring scheme according to a crossover process; in this way, N next generation coloring schemes are finally obtained and stored in the two-dimensional array crossover_set; the step of judging whether each coloring scheme needs to be crossover as a parent includes: generating a uniformly distributed random number r2 in the interval [0,1], when the crossover rate δ>r2, the parent coloring scheme must be crossover; the step of one crossover includes: assuming a and b are two parent coloring schemes participating in one crossover, and the vertex number partition block of the graph model with color c (c=1,2,3,4) in a and b is and (c=1,2,3,4), initialization (c=1,2,3,4), in turn (c=1,2,3,4), find the partition block from b , so that and The intersection is maximum, then ,from and Delete Repeat the above process for c=1,2,3,4 until (c=1,2,3,4) are all generated, and then the remaining vertex numbers that have not yet been included in the partition block are randomly inserted in sequence (c=1,2,3,4), finally divide the blocks according to the color (c=1,2,3,4), and obtain a next-generation coloring scheme after crossover.

[0011] Among them, step S15 includes: performing search_times loops on each coloring scheme in the two-dimensional array crossover_set in turn, assuming that a coloring scheme is crossover_set(m,:) (m=1,2,…,N), in each loop, counting the number of times that the graph model vertex violates the four-color principle for the coloring scheme, calculating the number of times each vertex is the endpoint of an edge that violates the four-color principle, selecting the vertex with the largest number of times, exchanging the color with another randomly selected vertex, and generating a new coloring scheme. If the fitness value of the new scheme is less than the original scheme, the new scheme is used to replace the original scheme, otherwise, the original scheme is kept unchanged; then performing the next loop operation until the loop reaches search_times times, a coloring scheme can be obtained and stored in search_set1(m,:); after all the coloring schemes in crossover_set have completed the local search, the two-dimensional array search_set1 is finally obtained.

[0012] Among them, step S16 includes: (n=1,2,…,N), consider the case where each vertex causes interference between constellations. Assume that the coefficient of non-reusable same color is 10 and the coefficient of non-reusable neighboring region is 1. Then the cumulative value of inter-constellation interference caused by vertex d (d=1,2,…,M) is , where E is the same when the primary and secondary constellations have the same color at vertex d. s is 1, otherwise E s is 0; E nh is the number of points with the same color as the vertex d of the secondary constellation and the adjacent points of the vertex d of the primary constellation; i n For each vertex, assuming there are vertices with non-zero interference cumulative values, Two independent search paths are used to synchronize the solution. i n After the search process is completed, the coloring scheme obtained according to search path 1 is recorded as , the coloring solution obtained according to search path 2 is recorded as ,Compare and The fitness value of the smaller coloring scheme is stored in search_set2(n,:). When all coloring schemes in search_set1 complete the above local search, the two-dimensional array search_set2 is finally obtained; the method of the search path 1 is: set , sort by non-zero interference cumulative value from large to small, and get the corresponding vertex sorting , sort the vertices in this order (j=1,2,…, ) to operate Take 1 to 4 in turn, calculate the fitness values ​​of these four coloring schemes respectively, and Update to the coloring scheme corresponding to the minimum fitness value, and then Repeat the same procedure until All operations on vertices are completed, and the coloring scheme is obtained ; The method of searching path 2 is: , the non-zero interference cumulative values ​​are not sorted by size, and the original order of the vertices corresponding to the non-zero interference cumulative values ​​is , sort the vertices in this order (j=1,2,…, ) to operate Take 1 to 4 in turn, calculate the fitness values ​​of these four coloring schemes respectively, and Update to the coloring scheme corresponding to the minimum fitness value, and then Repeat the same procedure until All operations on vertices are completed, and the coloring scheme is obtained .

[0013] Wherein, step S17 includes: for a coloring scheme in search_set2 (p=1,2,…,N), determine whether the coloring scheme needs to be mutated. If mutation is required, randomly select i p Randomly change the color of a vertex in the image to generate a new coloring scheme , let mutation_set = , otherwise the coloring scheme Will be directly stored in mutation_set ; When all coloring schemes in search_set2 complete the above operation, a two-dimensional array mutation_set is finally obtained; wherein, the method for determining whether a coloring scheme needs to be mutated includes: generating a uniformly distributed random number r3 in the interval [0,1], when the mutation rate , the coloring scheme must be mutated; otherwise, no mutation is required.

[0014] Among them, the elite mechanism used in step S12 to save the best individual includes: retaining the best individual in color_set coloring schemes are saved as elite individuals in the two-dimensional array elite_set, where each row corresponds to a coloring scheme. The proportion reserved for the elite; step S18 includes: calculating the fitness value of each coloring scheme in mutation_set, and arranging the coloring schemes from small to large according to the fitness value, and using elite_set to replace the worst A coloring scheme is used to obtain the new generation population next_set.

[0015] In a second aspect, the present invention provides a device for allocating beam resources in overlapping areas between low-orbit constellations, the device comprising: an initialization population module, used to construct a low-orbit constellation beam spectrum resource allocation graph model according to known parameters of the low-orbit constellation, the number of wave positions in the overlapping area between constellations is recorded as M, and for the secondary constellation wave position graph model G in the overlapping coverage area between constellations, a set of initial coloring schemes color_set with a population size of N that conforms to the four-color principle is randomly generated, wherein color_set is a two-dimensional array, each row corresponds to a coloring scheme, and each column corresponds to one of the M vertices in the graph model G, that is, color_set (x, y) represents the color of the yth wave position in the xth scheme, wherein 1≤x≤N, 1≤y≤M; a coloring scheme fitness calculation and best individual storage module, used to calculate the fitness of each coloring scheme in the population and store the best individual: according to the known coloring of the main constellation wave position graph model in the overlapping coverage area between constellations, the fitness value corresponding to each coloring scheme in color_set is calculated by the fitness calculation formula, and the fitness value corresponding to the xth coloring scheme is , 1≤x≤N, using the elite mechanism to save the best individuals; a roulette wheel selection module, used to use the roulette wheel selection operator to select the coloring scheme as the next generation parent; a crossover operation module, used to use the intersection crossover operator, according to the crossover rate, to achieve the generation of the offspring coloring scheme after the crossover operation; a first local search module, used to improve the crossover result based on the local search of the four-color principle conformity, to obtain a new offspring coloring scheme; a second local search module, used to further improve the crossover result based on the local search of reducing interference, to obtain a new offspring coloring scheme; a mutation operation module, used to use the mutation operator, according to the mutation rate, to realize the generation of offspring coloring scheme after mutation operation; population merging module, used to merge offspring populations, and form a new generation population based on the fitness calculation results; iteration termination judgment and result output module, used to judge whether the current iteration number reaches the maximum iteration number, if not, then replace the color_set with the new generation population, jump to the coloring scheme fitness calculation and save the best individual module, on the contrary, if the maximum iteration number is reached, then find the coloring scheme with the smallest fitness value from the new generation population, that is, the optimal coloring scheme, and output it as the result.

[0016] In a third aspect, the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores computer-executable instructions, and the computer-executable instructions are used to enable a computer to execute the method as described above.

[0017] The beneficial effects of the present invention are:

[0018] Aiming at the scenario where there are overlapping coverage areas between low-orbit constellations, the present invention optimizes the allocation of secondary constellation beam resources under the interference constraint of the main constellation, transforms the secondary constellation beam allocation problem into a graph model coloring scheme optimization problem, and reduces the mutual interference in the overlapping coverage area as much as possible by reasonably allocating the satellite beams of the secondary low-orbit constellation while ensuring the normal operation of the main low-orbit constellation, thereby improving the utilization rate of orbital frequency resources.

[0019] The present invention constructs a low-orbit inter-constellation beam spectrum resource allocation graph model, which has a double-layer structure and includes a main constellation wave position graph model and a secondary constellation wave position graph model in an overlapping coverage area between constellations. On the premise that the main constellation beam allocation situation is known, that is, the coloring situation of the main constellation wave position graph model in the overlapping coverage area between constellations is determined, a sub-constellation wave position graph model coloring method based on a genetic algorithm is designed to optimize the overlapping coverage area between constellations. A mode combining global convergence and local search is used, and a double-layer local search method based on four-color principle compliance and based on reducing inter-constellation interference is proposed to accelerate the convergence speed and improve the optimization efficiency of the optimal solution for sub-constellation beam resource allocation under the main constellation interference constraint condition.

[0020] In addition, the present invention adopts an elite mechanism to retain the best individuals during the iteration process of the genetic algorithm, ensuring that these individuals will not be destroyed by crossover and mutation operations. By retaining the best individuals, the optimal solution is prevented from being lost during the evolution process. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 This is a schematic diagram of the overlapping coverage areas between low-orbit constellations;

[0022] Figure 2 It is a model of spectrum resource allocation diagram of beams between LEO constellations;

[0023] Figure 3 It is a flowchart of an embodiment of a method for allocating beam resources between low-orbit constellations of the present invention;

[0024] Figure 4 yes Figure 3 A schematic diagram of spectrum resources of a four-color graph model of an application embodiment;

[0025] Figure 5 yes Figure 3 A schematic diagram of beam resource allocation in the overlapping area of ​​a main constellation satellite according to an application embodiment;

[0026] Figure 6 yes Figure 3 A schematic diagram of beam resource allocation in an overlapping area of ​​a secondary constellation satellite according to an application embodiment;

[0027] Figure 7 yes Figure 3 A schematic diagram of a simulation effect of an application embodiment;

[0028] Figure 8 yes Figure 3 Another schematic diagram of a simulation effect of an application embodiment;

[0029] Fig. 9 It is a structural schematic diagram of an embodiment of a device for allocating beam resources between low-orbit constellations of the present invention. DETAILED DESCRIPTION

[0030] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application may be combined with each other. Embodiment 1

[0031] See also Figure 3 , Figure 3 FIG. 1 is a flow chart of an embodiment of a method for allocating beam resources between low-orbit constellations of the present invention. Figure 3 As shown, the method comprises the following steps:

[0032] S11: Initialize the population;

[0033] According to the known parameters of the low-orbit constellation, a beam spectrum resource allocation graph model between low-orbit constellations is constructed. The graph model has a double-layer structure, including a main constellation wave position graph model and a secondary constellation wave position graph model in the overlapping coverage area between constellations. The number of wave positions in the overlapping area between constellations is recorded as M. For the secondary constellation wave position graph model G in the overlapping coverage area between constellations, an initial coloring scheme set color_set with a population size of N that conforms to the four-color principle is randomly generated, where color_set is a two-dimensional array, each row corresponds to a coloring scheme, and each column corresponds to one of the M vertices in the graph model G, that is, color_set(x,y) represents the color of the yth wave position in the xth scheme, where 1≤x≤N, 1≤y≤M; let the color set used for coloring be , count the number of colors in the first column of the color_set array, which are c1_num, c2_num, c3_num, and c4_num respectively. The generated initial solution set should also satisfy the following: if N is divisible by 4, then c1_num=c2_num=c3_num=c4_num=N / 4; if N is not divisible by 4, then c1_num=c2_num=c3_num=t, t is the quotient of N divided by 4, and c4_num=N-3*t.

[0034] Among them, the known parameters of the low-orbit constellation include: the number of beams in the overlapping area between constellations, the spectrum width corresponding to each color, the coloring scheme of the main constellation wave position, the spectrum requirement of each wave position of the main constellation, and the spectrum requirement of each wave position of the secondary constellation.

[0035] S12: Calculate the fitness of each coloring scheme in the population and save the best individual;

[0036] According to the coloring of the main constellation wave pattern model in the known overlapping coverage area between constellations, the fitness value corresponding to each coloring scheme in color_set is calculated by the fitness calculation formula. The fitness value corresponding to the x-th coloring scheme is , 1≤x≤N;

[0037] Among them, the fitness calculation formula is: ,in, i A coloring scheme for the sub-constellation wave pattern model within the overlapping coverage area between constellations, For the plan i The corresponding fitness value is is the penalty factor for the violation of the four-color principle by the internal wave position of the sub-constellation satellite, which is set to 10000; The number of times the four-color principle is violated in a coloring scheme of a constellation wave diagram model; is the penalty factor for the same-wavelength-position-and-color interference between constellations, which is set to 100; is the number of times the same color interference exists in the same wave position between constellations; is the penalty factor for the presence of same-color interference in adjacent wave positions between constellations, set to 1; is the number of times the same-color interference exists in adjacent wave positions between constellations; The penalty factor for the failure of the secondary constellation to provide service is set to 0.0001; It is the sum of the spectrum requirements corresponding to each closed service wave position in the secondary constellation.

[0038] In step S12, the best individual is saved by using the elite mechanism: the best individual in color_set is retained. coloring schemes are saved as elite individuals in the two-dimensional array elite_set, where each row corresponds to a coloring scheme. A proportion is reserved for the elite.

[0039] S13: Use the roulette wheel selection operator to select the coloring scheme as the parent of the next generation;

[0040] The probability that the xth coloring scheme is selected is , the cumulative probability of the xth coloring solution , the process of selecting a coloring scheme is: generate a uniformly distributed random number r1 in the interval [0,1], compare it with the cumulative probability of each scheme in turn, and select scheme t so that True; repeat the above selection process, select N coloring schemes from color_set as the next generation parents, and store them row by row into the two-dimensional array selected_set.

[0041] S14: using the intersection crossover operator, based on the crossover rate, to realize the generation of the offspring coloring scheme after the crossover operation;

[0042] Traverse each coloring scheme in the two-dimensional array selected_set, and determine whether each coloring scheme needs to be crossed as a parent. If it does not need to be crossed, the parent is directly copied to the next generation. If it needs to be crossed, another coloring scheme is randomly selected from the two-dimensional array selected_set as another parent, and a next-generation coloring scheme is generated according to a crossover process. In this way, N next-generation coloring schemes are finally obtained and stored in the two-dimensional array crossover_set;

[0043] The step of judging whether each coloring scheme as a parent needs to perform a crossover operation comprises: generating a uniformly distributed random number r2 in the interval [0,1], and when the crossover rate δ>r2, the parent coloring scheme must perform a crossover;

[0044] The steps of a crossover include: let a and b be two parent coloring schemes participating in a crossover, and record the vertex number partition blocks of the graph model with color c (c=1,2,3,4) in a and b as and (c=1,2,3,4), initialization (c=1,2,3,4), in turn (c=1,2,3,4), find the partition block from b , so that and The intersection is maximum, then ,from and Delete Repeat the above process for c=1,2,3,4 until (c=1,2,3,4) are all generated, and then the remaining vertex numbers that have not yet been included in the partition block are randomly inserted in sequence (c=1,2,3,4), finally divide the blocks according to the color (c=1,2,3,4), and obtain a next-generation coloring scheme after crossover.

[0045] S15: Based on the local search of the four-color principle, the crossover result is improved to obtain a new offspring coloring scheme;

[0046] For each coloring scheme in the two-dimensional array crossover_set, search_times loops are performed in turn. Suppose a coloring scheme is crossover_set(m,:) (m=1,2,…,N). In each loop, the number of times the graph model vertices violate the four-color principle is counted for this coloring scheme, and the number of times each vertex is the endpoint of an edge that violates the four-color principle is calculated. The vertex with the largest number of violations is selected, and its color is swapped with another randomly selected vertex to generate a new coloring scheme. If the fitness value of the new scheme is less than the original scheme, the new scheme is used to replace the original scheme, otherwise, the original scheme is kept unchanged; then the next loop operation is performed until the loop reaches search_times times, and a coloring scheme can be obtained and stored in search_set1(m,:); after all the coloring schemes in crossover_set have completed the local search, the two-dimensional array search_set1 is finally obtained.

[0047] S16: Based on the local search with reduced interference, the crossover result is further improved to obtain a new offspring coloring scheme;

[0048] About the plan (n=1,2,…,N), consider the case where each vertex causes interference between constellations. Assume that the coefficient of non-reusable same color is 10 and the coefficient of non-reusable neighboring region is 1. Then the cumulative value of inter-constellation interference caused by vertex d (d=1,2,…,M) is , where E is the same when the primary and secondary constellations have the same color at vertex d. s is 1, otherwise E s is 0; E nh is the number of points with the same color as the vertex d of the secondary constellation and the adjacent points of the vertex d of the primary constellation;

[0049] About the plan i n For each vertex, assuming there are vertices with non-zero interference cumulative values, Two independent search paths are used to synchronize the solution. i n After the search process is completed, the coloring scheme obtained according to search path 1 is recorded as , the coloring solution obtained according to search path 2 is recorded as ,Compare and The fitness value of the smaller one is stored in search_set2(n,:). When all coloring schemes in search_set1 complete the above local search, the two-dimensional array search_set2 is finally obtained.

[0050] The method of searching path 1 is as follows: , sort by non-zero interference cumulative value from large to small, and get the corresponding vertex sorting , sort the vertices in this order (j=1,2,…, ) to operate Take 1 to 4 in turn, calculate the fitness values ​​of these four coloring schemes respectively, and Update to the coloring scheme corresponding to the minimum fitness value, and then Repeat the same procedure until All operations on vertices are completed, and the coloring scheme is obtained ;

[0051] The method of searching path 2 is: , the non-zero interference cumulative values ​​are not sorted by size, and the original order of the vertices corresponding to the non-zero interference cumulative values ​​is , sort the vertices in this order (j=1,2,…, ) to operate Take 1 to 4 in turn, calculate the fitness values ​​of these four coloring schemes respectively, and Update to the coloring scheme corresponding to the minimum fitness value, and then Repeat the same procedure until All operations on vertices are completed, and the coloring scheme is obtained .

[0052] S17: using a mutation operator to generate a coloring scheme for a subsequent generation after the mutation operation according to the mutation rate;

[0053] For a coloring scheme in search_set2 (p=1,2,…,N), determine whether the coloring scheme needs to be mutated. If mutation is required, randomly select Randomly change the color of a vertex in the image to generate a new coloring scheme , let mutation_set = , otherwise the coloring scheme Will be directly stored in mutation_set ; When all coloring schemes in search_set2 complete the above operations, the two-dimensional array mutation_set is finally obtained;

[0054] The method for determining whether a coloring scheme needs to be mutated includes: generating a uniformly distributed random number r3 in the interval [0,1], and when the mutation rate , the coloring scheme must be mutated; otherwise, no mutation is required.

[0055] S18: Merge the offspring populations and select the best ones to form a new generation population based on the fitness calculation results;

[0056] Calculate the fitness value of each coloring scheme in mutation_set, and arrange the coloring schemes from small to large according to the fitness value, and use elite_set to replace the worst one A coloring scheme is used to obtain the new generation population next_set.

[0057] S19: Determine whether the current number of iterations has reached the maximum number of iterations. If not, replace the color_set with the new generation population and jump to step S12. Otherwise, if the maximum number of iterations has been reached, find the coloring scheme with the smallest fitness value from the new generation population, which is the optimal coloring scheme and output it as the result.

[0058] The maximum number of iterations is recorded as max_gen, and it is determined whether the current number of iterations reaches the maximum number of iterations max_gen. If not, color_set = next_set is updated, and the process returns to step S12. If max_gen is reached, end_set = next_set is set, and then the fitness value of each coloring scheme in end_set is calculated, and the coloring scheme corresponding to the minimum fitness value is selected as the final result output.

[0059] The following are the simulation effects of the present invention:

[0060] Compare this method with the exhaustive traversal optimization method. The specific steps of the exhaustive traversal method are:

[0061] For the sub-constellation wave position graph model in the overlapping coverage area between constellations, according to the four-color principle, find all the graph model coloring schemes that meet the requirements, and these schemes all satisfy the requirement that adjacent wave positions have different colors;

[0062] By using an evaluation function (which is the same as the fitness calculation formula), according to the known coloring of the primary constellation wave graph model in the overlapping coverage area between constellations, the spectrum requirements of each wave graph of the primary constellation, and the spectrum requirements of each wave graph of the secondary constellation, the evaluation values ​​corresponding to all the graph model coloring schemes found in step (1) are calculated;

[0063] All evaluation values ​​are sorted, and the sub-constellation wave position diagram model coloring scheme in the overlapping coverage area between constellations corresponding to the minimum evaluation value is found. This scheme is the optimal scheme.

[0064] (I) Simulation conditions

[0065] When the number of beams in the overlapping area between constellations, the spectrum width corresponding to each color, the coloring scheme of the main constellation, and the spectrum requirements of each wave position of the main and secondary constellations are known, the beam resource allocation based on the four-color graph model coloring is simulated by exhaustive traversal and genetic algorithm to optimize the secondary constellation coloring scheme. The specific simulation parameters are shown in Table 1.

[0066]

[0067] Assume that the spectrum resources corresponding to the four colors used for coloring the graph model are as follows Figure 4 As shown in Figure 2, according to the main constellation wave position coloring scheme, the coloring schematic diagram of the main constellation satellite overlapping area wave position diagram model is as follows: Figure 5 As shown in the figure, each wave position shows the wave position number and the corresponding spectrum requirement.

[0068] (II) Simulation results

[0069] Through exhaustive traversal and genetic algorithm, the optimal beam resource allocation scheme for the overlapping area of ​​the secondary constellation satellite can be obtained, so that the spectrum requirements of the secondary constellation satellite can be met to the maximum extent without interfering with the main constellation. The optimal scheme is [1 3 1 3 4 2 4 2 1 3 1 2 3 2 4 3 1 3 2], as shown in Figure 6 Except for wave positions 3, 5, 6, 7, 11, and 12 which cannot provide services due to neighboring interference to the main constellation, the needs of other secondary constellation wave positions can be met.

[0070] This method was compared with the exhaustive traversal method to conduct a time cost test. Each method was tested 100 times. The results are as follows: Figure 7 .

[0071] like Figure 7 As shown in the figure, the exhaustive traversal method takes a long time, with an average time of about 258 seconds; the genetic algorithm is a non-traversal algorithm, and the average time consumed for optimization is about 11 seconds, which greatly reduces the optimization time. Compared with the exhaustive traversal method, the optimization efficiency of this method is increased by nearly 23 times.

[0072] like Figure 8 As shown, the optimization method based on genetic algorithm can find the optimal sub-constellation satellite beam resource allocation scheme, and this method has a good convergence speed. Embodiment 2

[0073] See also Fig. 9 , Fig. 9 It is a structural schematic diagram of an embodiment of the device for allocating beam resources between low-orbit constellations of the present invention. As shown in the figure, the device includes a population initialization module 11, a coloring scheme fitness calculation and best individual storage module 12, a roulette wheel selection module 13, a crossover operation module 14, a first local search module 15, a second local search module 16, a mutation operation module 17, a population merging module 18, and an iteration termination judgment and result output module 19.

[0074] The population initialization module 11 is used to construct a beam spectrum resource allocation diagram model between low-orbit constellations according to the known parameters of the low-orbit constellations, record the number of wave positions in the overlapping area between constellations as M, and randomly generate an initial coloring scheme set color_set with a population size of N that conforms to the four-color principle for the sub-constellation wave position diagram model G in the overlapping coverage area between constellations, where color_set is a two-dimensional array, each row corresponds to a coloring scheme, and each column corresponds to one of the M vertices in the diagram model G, that is, color_set(x,y) represents the color of the yth wave position in the xth scheme, where 1≤x≤N, 1≤y≤M.

[0075] The coloring scheme fitness calculation and best individual storage module 12 is used to calculate the fitness of each coloring scheme in the population and store the best individual: according to the coloring of the main constellation wave position diagram model in the known overlapping coverage area between constellations, the fitness value corresponding to each coloring scheme in color_set is calculated by the fitness calculation formula, and the fitness value corresponding to the x-th coloring scheme is , 1≤x≤N, and an elite mechanism is used to store the best individuals.

[0076] The roulette wheel selection module 13 is used to select a coloring scheme as a parent of the next generation by using a roulette wheel selection operator.

[0077] The crossover operation module 14 is used to use the intersection crossover operator to realize the generation of the offspring coloring scheme after the crossover operation according to the crossover rate.

[0078] The first local search module 15 is used for improving the crossover result and obtaining a new offspring coloring scheme based on the local search conforming to the four-color principle.

[0079] A second local search module 16, configured to further improve the crossover result based on the local search with reduced interference, and obtain a new offspring coloring scheme;

[0080] A mutation operation module 17, used to use a mutation operator to realize the generation of a coloring scheme of offspring after the mutation operation according to a mutation rate;

[0081] The population merging module 18 is used to merge the offspring populations and select the best ones to form a new generation population according to the fitness calculation results;

[0082] The iteration termination judgment and result output module 19 is used to judge whether the current number of iterations has reached the maximum number of iterations. If the maximum number of iterations has not been reached, the color_set is replaced by the new generation population, and the module 12 for calculating the fitness of the coloring scheme and storing the best individual is jumped to. On the contrary, if the maximum number of iterations has been reached, the coloring scheme with the smallest fitness value is found from the new generation population, which is the optimal coloring scheme and is output as the result.

[0083] Specifically, the working methods of each module in this embodiment have been described in detail in the first embodiment and will not be repeated here. Embodiment 3

[0084] The present invention further provides a computer-readable storage medium, which stores computer-executable instructions, and the computer-executable instructions are used to enable a computer to execute the method described in the first embodiment.

[0085] The above is a specific description of the preferred implementation of the present invention, but the invention is not limited to the embodiments. Those skilled in the art can make various equivalent modifications or substitutions without violating the spirit of the present invention. These equivalent modifications or substitutions are all included in the scope defined by the claims of this application.

Claims

1. A method for allocating beam resources in overlapping areas between low-orbit constellations, characterized in that: The steps include: S11: Initialize the population: construct a spectrum resource allocation graph model of beams between LEO constellations according to known parameters of LEO constellations, record the number of wave positions in the overlapping area between constellations as M, and for the sub-constellation wave position graph model G in the overlapping coverage area between constellations, randomly generate an initial coloring scheme set color_set with a population size of N that complies with the four-color principle, where color_set is a two-dimensional array, each row corresponds to a coloring scheme, and each column corresponds to one of the M vertices in the graph model G, that is, color_set(x,y) represents the color of the yth wave position in the xth scheme, where 1≤x≤N, 1≤y≤M; S12: Calculate the fitness of each coloring scheme in the population and save the best individual: According to the coloring of the main constellation wavemap model in the known overlapping coverage area between constellations, calculate the fitness value corresponding to each sub-constellation wavemap model coloring scheme in color_set through the fitness calculation formula. The fitness value corresponding to the x-th coloring scheme is , 1≤x≤N, and then use the elite mechanism to save the best individuals; S13: Use the roulette wheel selection operator to select the coloring scheme as the parent of the next generation; S14: using the intersection crossover operator, based on the crossover rate, to realize the generation of the offspring coloring scheme after the crossover operation; S15: Based on the local search of the four-color principle, the crossover result is improved to obtain a new offspring coloring scheme; S16: Based on the local search with reduced interference, the crossover result is further improved to obtain a new offspring coloring scheme; S17: using a mutation operator to generate a coloring scheme for offspring after the mutation operation according to the mutation rate; S18: Merge the offspring populations and select the best ones to form a new generation population based on the fitness calculation results; S19: Determine whether the current number of iterations has reached the maximum number of iterations. If not, replace the color_set with the new generation population and jump to step S12. Otherwise, if the maximum number of iterations has been reached, find the coloring scheme with the smallest fitness value from the new generation population, which is the optimal coloring scheme and output it as the result.

2. The method according to claim 1, characterized in that The fitness calculation formula is as follows: ,in, i A coloring scheme for the sub-constellation wave pattern model within the overlapping coverage area between constellations, For the plan i The corresponding fitness value is is the penalty factor for the violation of the four-color principle by the internal wave position of the sub-constellation satellite, which is set to 10000; The number of times the four-color principle is violated in a coloring scheme of a constellation wave diagram model; is the penalty factor for the same-wavelength-position-and-color interference between constellations, which is set to 100; is the number of times the same color interference exists in the same wave position between constellations; is the penalty factor for the presence of same-color interference in adjacent wave positions between constellations, set to 1; is the number of times the same-color interference exists in adjacent wave positions between constellations; The penalty factor for the failure of the secondary constellation to provide service is set to 0.0001; It is the sum of spectrum requirements corresponding to each closed service wave position in the secondary constellation.

3. The method according to claim 1, characterized in that The step S13 comprises: The probability that the xth coloring scheme is selected is , the cumulative probability of the xth coloring solution , the process of selecting a coloring scheme is: generate a uniformly distributed random number r1 in the interval [0,1], compare it with the cumulative probability of each scheme in turn, and select scheme t so that True; repeat the above selection process, select N coloring schemes from color_set as the next generation parents, and store them row by row into the two-dimensional array selected_set.

4. The method according to claim 3, characterized in that The step S14 comprises: Traverse each coloring scheme in the two-dimensional array selected_set, determine whether each coloring scheme needs to be crossed as a parent, if not, directly copy the parent to the next generation, if necessary, randomly select another coloring scheme from the two-dimensional array selected_set as another parent, and generate a next generation coloring scheme according to a crossover process; in this way, finally obtain N next generation coloring schemes and store them in the two-dimensional array crossover_set; The step of determining whether each coloring scheme as a parent needs to perform a crossover operation comprises: generating a uniformly distributed random number r2 in the interval [0,1], and when the crossover rate δ>r2, the parent coloring scheme must perform a crossover; The steps of one crossover include: assuming a and b are two parent coloring schemes participating in one crossover, and recording the vertex number partition blocks of the graph model with color c (c=1,2,3,4) in a and b as and (c=1,2,3,4), initialization (c=1,2,3,4), respectively (c=1,2,3,4), find the partition block from b , so that and The intersection is maximum, then ,from and Delete Repeat the above process for c=1,2,3,4 until (c=1,2,3,4) are all generated, and then the remaining vertex numbers that have not yet been included in the partition block are randomly inserted in sequence (c=1,2,3,4), finally divide the blocks according to the color (c=1,2,3,4), and obtain a next-generation coloring solution after crossover.

5. The method according to claim 4, characterized in that The step S15 comprises: For each coloring scheme in the two-dimensional array crossover_set, search_times cycles are performed in turn. Suppose a coloring scheme is crossover_set(m,:) (m=1,2,…,N). In each cycle, the number of times the graph model vertices violate the four-color principle is counted for the coloring scheme, and the number of times each vertex is the endpoint of an edge that violates the four-color principle is calculated. The vertex with the largest number of violations is selected, and the color is swapped with another randomly selected vertex to generate a new coloring scheme. If the fitness value of the new scheme is less than the original scheme, the new scheme is used to replace the original scheme, otherwise, the original scheme is kept unchanged; then the next cycle operation is performed until the cycle reaches search_times times, and a coloring scheme can be obtained and stored in search_set1(m,:); after all the coloring schemes in crossover_set have completed the local search, the two-dimensional array search_set1 is finally obtained.

6. The method according to claim 5, characterized in that The step S16 comprises: About the plan (n=1,2,…,N), consider the case where each vertex causes interference between constellations. Assume that the coefficient of non-reusable same color is 10 and the coefficient of non-reusable neighboring region is 1. Then the cumulative value of inter-constellation interference caused by vertex d (d=1,2,…,M) is , where E is the same when the primary and secondary constellations have the same color at vertex d. s is 1, otherwise E s is 0; E nh is the number of points with the same color as the vertex d of the secondary constellation and the adjacent points of the vertex d of the primary constellation; About the plan i n For each vertex, assuming there are vertices with non-zero interference cumulative values, Two independent search paths are used to synchronize the solution. i n After the search process is completed, the coloring scheme obtained according to search path 1 is recorded as , the coloring solution obtained according to search path 2 is recorded as ,Compare and The fitness value of the smaller one is stored in search_set2(n,:). When all coloring schemes in search_set1 complete the above local search, the two-dimensional array search_set2 is finally obtained. The method for searching path 1 is as follows: , sort by non-zero interference cumulative value from large to small, and get the corresponding vertex sorting , sort the vertices in this order (j=1,2,…, ) to operate Take 1 to 4 in turn, calculate the fitness values ​​of these four coloring schemes respectively, and Update to the coloring scheme corresponding to the minimum fitness value, and then Repeat the same procedure until All operations on vertices are completed, and the coloring scheme is obtained ; The method of searching path 2 is: , the non-zero interference cumulative values ​​are not sorted by size, and the original order of the vertices corresponding to the non-zero interference cumulative values ​​is , sort the vertices in this order (j=1,2,…, ) to operate Take 1 to 4 in turn, calculate the fitness values ​​of these four coloring schemes respectively, and Update to the coloring scheme corresponding to the minimum fitness value, and then Repeat the same procedure until All operations on vertices are completed, and the coloring scheme is obtained .

7. The method according to claim 6, characterized in that The step S17 comprises: For a coloring scheme in search_set2 (p=1,2,…,N), determine whether the coloring scheme needs to be mutated. If mutation is required, randomly select Randomly change the color of a vertex in the image to generate a new coloring scheme , let mutation_set = , otherwise the coloring scheme Will be directly stored in mutation_set ; When all coloring schemes in search_set2 complete the above operations, the two-dimensional array mutation_set is finally obtained; The method for determining whether a coloring scheme needs to be mutated includes: generating a uniformly distributed random number r3 in the interval [0,1], and when the mutation rate When , the coloring scheme must be mutated; otherwise, no mutation is required.

8. The method according to claim 7, characterized in that The step S12 adopts the elite mechanism to save the best individual, including: retaining the best individual in color_set coloring schemes are saved as elite individuals in the two-dimensional array elite_set, where each row corresponds to a coloring scheme. a proportion reserved for the elite; The step S18 comprises: calculating the fitness value of each coloring scheme in mutation_set, arranging the coloring schemes from small to large according to the fitness value, and replacing the worst one with elite_set. A coloring scheme is used to obtain the new generation population next_set.

9. A device for allocating beam resources in overlapping areas between low-orbit constellations, characterized in that: include: Initialize the population module, which is used to construct a spectrum resource allocation graph model of the beams between low-orbit constellations according to the known parameters of the low-orbit constellations. The number of wave positions in the overlapping area between constellations is recorded as M. For the sub-constellation wave position graph model G in the overlapping coverage area between constellations, randomly generate an initial coloring scheme set color_set with a population size of N that conforms to the four-color principle, where color_set is a two-dimensional array, each row corresponds to a coloring scheme, and each column corresponds to one of the M vertices in the graph model G, that is, color_set(x,y) represents the color of the yth wave position in the xth scheme, where 1≤x≤N, 1≤y≤M; The module for calculating the fitness of the coloring scheme and saving the best individual is used to calculate the fitness of each coloring scheme in the population and save the best individual: according to the coloring of the main constellation wave pattern model in the known overlapping coverage area between constellations, the fitness value corresponding to each coloring scheme in color_set is calculated by the fitness calculation formula. The fitness value corresponding to the x-th coloring scheme is , 1≤x≤N, the best individuals are saved by using the elite mechanism; A roulette wheel selection module, for selecting a coloring scheme as a parent of the next generation using a roulette wheel selection operator; A crossover operation module, used for realizing the generation of a coloring scheme of offspring after the crossover operation according to the crossover rate by using the intersection crossover operator; The first local search module is used for local search based on the conformity of the four-color principle to improve the crossover result and obtain a new offspring coloring scheme; A second local search module is used to further improve the crossover result based on the local search with reduced interference to obtain a new offspring coloring scheme; A mutation operation module is used to use a mutation operator to realize the generation of a coloring scheme of offspring after the mutation operation according to the mutation rate; The population merging module is used to merge the offspring populations and select the best ones to form a new generation of populations based on the fitness calculation results; The iteration termination judgment and result output module is used to judge whether the current number of iterations has reached the maximum number of iterations. If it has not reached the maximum number of iterations, the color_set is replaced with the new generation population, and the module jumps to the fitness calculation and best individual storage module of the coloring scheme. On the contrary, if it reaches the maximum number of iterations, the coloring scheme with the smallest fitness value is found from the new generation population, which is the optimal coloring scheme and is output as the result.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-executable instructions, and the computer-executable instructions are used to enable a computer to execute the method according to any one of claims 1 to 8.

Images