Distributed multi-satellite mission planning method considering maximum benefit and load balancing

Abstract

The present invention provides a distributed multi-satellite mission planning method and system considering maximum benefit and load balancing, and relates to the technical field of multi-satellite mission planning. In the present invention, a satellite resource set and a task set to be observed are obtained; a multi-satellite mission planning model that combines the preset maximum total observation benefit target and the star cluster load balancing target is used to obtain multiple feasible solutions for multi-satellite mission planning; and a distributed co-evolution particle swarm algorithm is used to obtain the final distributed multi-satellite mission collaborative observation planning scheme. Considering the distributed satellite load balancing, a multi-satellite mission planning model is constructed; a distributed co-evolution particle swarm algorithm is designed, and the particles are co-evolved and then randomly matched to resolve conflicts, so as to obtain complete new particles, thereby improving population diversity. At the same time, a pairwise competition mechanism is introduced to increase the source of particle social learning, and avoid the situation of "premature" convergence, thereby obtaining a better distributed multi-satellite mission collaborative observation planning scheme and improving the efficiency of satellite resource utilization.

Description

Technical Field

[0001] The present invention relates to the technical field of multi-satellite mission planning, and in particular to a distributed multi-satellite mission planning method, system, storage medium and electronic device that considers maximum benefit and load balancing. Background Art

[0002] Imaging reconnaissance satellites operate in space on a certain orbit, forming an observation area on the ground with the sub-satellite point trajectory as the axis, which is determined by the field of view angle, roll angle, pitch angle, etc. The satellite has a strict time window for observing targets in the area and requires a certain amount of observation resources.

[0003] The distributed imaging satellite collaborative observation scheduling problem can be described as follows: when any satellite in a group of satellites receives mission information, it transmits the mission information to other satellites in the group through intersatellite links. The group jointly conducts mission planning and formulates a reasonable planning scheme to maximize the mission benefits while ensuring the efficiency of mission planning. Genetic algorithms, particle swarm algorithms and other heuristic algorithms are generally used to solve the satellite emergency mission scheduling problem.

[0004] The particle swarm optimization algorithm simulates the predation behavior of a flock of birds. It is initialized as a group of random particles (random solutions). The particles follow the optimal particles in the solution space to search, and then find the optimal solution through iteration. The particle swarm optimization algorithm is simple, easy to implement, and does not have many parameters to adjust, so it has been widely used in function optimization. However, since the particle swarm optimization algorithm "flies" towards the optimal solution based on the search experience of all particles and its own, the convergence speed slows down significantly in the later stage of evolution. At the same time, when the algorithm converges to a certain accuracy, it is very easy to fall into local extreme points; in addition, particles sometimes appear to oscillate near the optimal solution. The algorithm cannot continue to optimize, so the accuracy that the algorithm can achieve is poor, and thus the efficiency of satellite resource utilization cannot be fully improved. Summary of the invention

[0005] 1. Technical issues to be solved

[0006] In view of the shortcomings of the prior art, the present invention provides a distributed multi-satellite mission planning method, system, storage medium and electronic device that takes into account maximum benefit and load balancing, solving the technical problem of not being able to fully improve the efficiency of satellite resource utilization.

[0007] (II) Technical solution

[0008] To achieve the above objectives, the present invention is implemented through the following technical solutions:

[0009] A distributed multi-satellite mission planning method considering maximum benefit and load balancing includes:

[0010] S1. Obtain a set of satellite resources and a set of tasks to be observed;

[0011] S2. Acquire multiple feasible solutions for multi-satellite mission planning according to the satellite resource set and the task set, in combination with a preset multi-satellite mission planning model that maximizes the total observation benefit target and the constellation load balancing target;

[0012] S3. According to the plurality of feasible solutions, a distributed co-evolutionary particle swarm algorithm is used to obtain a final distributed multi-satellite mission collaborative observation planning scheme.

[0013] Preferably, the multi-satellite mission planning model in S2 includes:

[0014] The objective function of maximizing the total observation benefit objective and the constellation load balancing objective is:

[0015]

[0016] Among them, λ1 and λ2 represent the weights of the total observation benefit target and the constellation load balancing target respectively;

[0017] Satellite resource collection, N S is the number of satellite resources;

[0018] The set of tasks to be observed, N T is the number of tasks;

[0019] For task t i In satellite S j The number of visible time windows on the

[0020] is the decision variable,

[0021]

[0022] p i For task t i Observational benefits of

[0023] CT j Satellite S j Observe the number of scheduled tasks;

[0024] LD is the load balancing evaluation value of the constellation, that is, the standard deviation of the number of tasks completed by each satellite in the planning scheme:

[0025]

[0026] Preferably, the multi-satellite mission planning model in S3 further includes constraints:

[0027]

[0028]

[0029]

[0030] ots ij +dur i =ote ij (7)

[0031] Among them, formula (4) means that a task is observed at most once;

[0032] Formula (5) indicates that the task observation needs to meet the time window requirement, that is, the execution time window of the task must be within the visible time window; For task t i In satellite S j The ath visible time window on the previous page; For task t i In satellite S j The start time of the last visible time window; For task t i In satellite S j The end time of the last visible time window;

[0033] Formula (6) represents the time interval constraint of continuous observation tasks; tr i,i+1 Satellite S j Satellite sensor observation mission i and task t i+1 The time between posture transitions;

[0034] Formula (7) indicates that the observation end time of a task is equal to the sum of the observation start time and the task observation duration; dur i Satellite S j Satellite sensor mission t i Observation duration; ots ij 、ote ij Represents tasks t i In satellite S j The actual start and end time of the observation.

[0035] Preferably, the S3 specifically includes:

[0036] S31, set t=0;

[0037] Each initial solution is put into the initial particle population as a particle. The total number of particles is TK, TK = K·N S ; Randomly initialize the position and velocity of each particle in the entire population;

[0038] S32, each particle p in the initial particle population tkThe gene matrix is ​​split into strips according to the satellite number, and the satellite S j The corresponding particle gene vector dpt k (j), forming a set of particle gene vectors TDP(j) = {dp1(j),…,dp tk (j),…,dp TK (j)}; the gene matrix includes a sequence matrix and a weight matrix;

[0039] S33. Calculate the particle gene vector dp tk (j) The fitness values ​​of all sub-particles;

[0040]

[0041] in, is the t-th generation particle dp tk (j) fitness value; w tk,j,m is the element in the weight matrix, i.e., the particle p tk (j) Satellite S in the corresponding scheme j The weight of the mth observation task;

[0042] And from the sub-particle gene vector set TDP(j), according to the elite selection strategy, K sub-particles are selected to form the initial sub-population DP(j) = {dp1(j),…,dp k (j),…,dp K (k)}, a total of N S Initial sub-populations: DP(1),…,DP(j),…,DP(N S );

[0043] S34. Define the initial positions and fitness values ​​of the sub-particles in the initial sub-population DP(j) as For the particle with the largest fitness value in the initial sub-population DP(j), its position and evaluation value are defined as gBest t (j) and

[0044] For the initial sub-population DP(j), the preset movement formula is used to perform sub-particle co-evolution to update the speed and position of the sub-particles;

[0045] S35, select the sub-particles dp in the current minute group in turn k (j) Randomly select a sub-particle from each of the other current sub-populations to form a quasiparticle qp k , and then perform matching conflict resolution to obtain complete particles. Each satellite S j Get the subpopulation P(j), j=1,2,…,N S , the number of particles in the subpopulation P(j) is K;

[0046] S36. Calculate the particle p in each subpopulation P(j) k The location and fitness value of Thus, the position and fitness value of the particle with the largest fitness value in each sub-population P(j) are determined. AI t (gBest j );

[0047] Identify AI t (gBest j ) the smallest first subpopulation P(j), and AI t (gBest j ) the second sub-population P(j) with the largest fitness value, and exchange the particle with the smallest fitness value in the first sub-population P(j) with a randomly selected particle in the second sub-population P(j);

[0048] S37, using the movement formula to update the speed and position of all particles in the entire sub-population P(j);

[0049] S38, calculate each particle p in the current sub-population P(j) k The current fitness value of renew and

[0050]

[0051] S39, when the number of iterations reaches the preset maximum number of evolutions Nmax, or the population convergence coefficient reaches the preset threshold ε0, the iteration is stopped, and the corresponding AI (gBest j ), determine the global optimal position of the current initial particle population TP and its fitness value gBest t 、AI gBest , decode the corresponding particles to obtain the final distributed multi-satellite mission collaborative observation planning scheme; otherwise, merge the current sub-populations P(j) to obtain a particle population and use it as the initial particle population TP in the next iteration process, set t=t+1, and return to S32.

[0052] Preferably, in the distributed collaborative evolution particle swarm algorithm:

[0053] Based on the decimal encoding method of the task number, the initial particle population P = {p1, p2, ..., p tk ,…,p TK}, the total number of particles in the population is TK = K·N S , where p tk represents the tkth particle;

[0054] Particle p k It is composed of a gene matrix, which is divided into a sequence matrix X k and the weight matrix W k , X k The position of the gene in represents the execution order of the task; k With X k Correspondingly, W k The task weights in X k The task number in ;

[0055] Then the particle dp k (j) is a gene vector corresponding to the particle gene matrix. The gene vector is divided into a sequence vector X k (j) and the weight vector W k (j), X k The position of the genes in (j) represents the execution order of the tasks; W k (j) and X k (j) Correspondingly, W k The task weight in (j) corresponds to X k The task number in (j);

[0056] At the tth generation of evolution, particle p k The serial number matrix for:

[0057]

[0058] Particle p k The weight matrix for:

[0059]

[0060] Then the particle dp k The sequence vector of (j) is:

[0061]

[0062] particle dp k The weight vector of :

[0063]

[0064] in, and Respectively represent: the particle dp at the tth generation of evolution k (j) In the corresponding scheme, satellite S j The sequence number and weight of the mth observation task;

[0065] Represents particle pk The historical optimal position achieved during the evolution from the initial position to the tth generation, The corresponding weight matrix is have:

[0066]

[0067] but Indicates the particle dp k (j) The historical optimal position achieved during the evolution from the initial position to the tth generation, The corresponding weight vector is have:

[0068]

[0069] in, express In the corresponding scheme, satellite S j The weight of the mth observation task;

[0070] gBest t Indicates the optimal position reached by the particle population from the initial position to the tth generation, gBest t The corresponding weight matrix is ​​gBestW t ,have:

[0071]

[0072] gBest t (j) represents the optimal position reached by the particle population from the initial position to the tth generation, gBest t (j) The corresponding weight vector is gBestW t (j) with:

[0073]

[0074] in, gBest t (j) Satellite S in the corresponding scheme j The weight of the mth observation task.

[0075] Preferably, the sub-particle co-evolution in S34 updates the speed and position of the sub-particle, specifically including:

[0076] Move the particle gene vector dp according to the movement formula k (j), at the t+1th generation of evolution, the particle dp k The movement formula is

[0077]

[0078]

[0079] in, Respectively represent the evolutionary time-division particle dp of the tth and t+1th generation k (j) In the corresponding scheme, satellite S j The weight of the i-th observation task is used to characterize the particle position;

[0080] dp respectively k (j) Movement speed during the evolution of the tth and t+1th generations;

[0081] ω is the inertia coefficient, which indicates the influence of the velocity of the sub-particles of the tth generation on the moving velocity of the sub-particles of the t+1th generation;

[0082] c1 is the sub-particle self-evolution acceleration, c2 is the sub-particle competition learning acceleration, and c3 is the sub-particle global movement acceleration;

[0083] r1, r2, r3 are random numbers uniformly distributed in [0, 1];

[0084] Indicates the particle dp k (j) According to the pairwise competition mechanism, the particles dp k′ (j),dp k′ (j)≠dp k (j) Learning: In pairwise competition, if dp k (j) Losing the game, i.e. AI k (j) <AI k′ (j), then in the next step, update the formula c2≠0, dp k (j) By submitting to dp k′ (j) Learn to update its position, otherwise c2 = 0;

[0085] Indicates the particle dp k (j) Global social learning.

[0086] Preferably, the conflict resolution negotiation rules in S35 include:

[0087] When a task can be observed by different satellites, satellites with smaller loads are given priority to perform the observation task; when multiple tasks conflict, observation tasks with greater benefits are given priority.

[0088] A distributed multi-satellite mission planning system considering maximum benefit and load balancing, comprising:

[0089] An acquisition module is used to acquire a set of satellite resources and a set of tasks to be observed;

[0090] A selection module is used to obtain multiple feasible solutions for multi-satellite mission planning according to the satellite resource set and the task set, in combination with a preset multi-satellite mission planning model with a maximum total observation benefit target and a constellation load balancing target;

[0091] The planning module is used to obtain the final distributed multi-satellite mission collaborative observation planning scheme based on the multiple feasible solutions by using the distributed co-evolutionary particle swarm algorithm.

[0092] A storage medium stores a computer program for distributed multi-satellite mission planning considering maximum benefit and load balancing, wherein the computer program enables a computer to execute the distributed multi-satellite mission planning method as described above.

[0093] An electronic device, comprising:

[0094] one or more processors;

[0095] Memory; and

[0096] One or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, the programs including instructions for executing the distributed multi-satellite mission planning method as described above.

[0097] (III) Beneficial effects

[0098] The present invention provides a distributed multi-satellite mission planning method, system, storage medium and electronic device that considers maximum benefit and load balancing. Compared with the prior art, it has the following beneficial effects:

[0099] In the present invention, firstly, a satellite resource set and a task set to be observed are obtained; then, according to the satellite resource set and the task set, a multi-satellite task planning model with a preset maximum total observation benefit target and a star cluster load balancing target is combined to obtain multiple feasible solutions for multi-satellite task planning; finally, according to the multiple feasible solutions, a distributed co-evolutionary particle swarm algorithm is used to obtain the final distributed multi-satellite task collaborative observation planning scheme. Considering the distributed satellite load balancing, a distributed co-evolutionary multi-satellite task planning model is constructed, and a distributed co-evolutionary particle swarm algorithm is designed. The particles are co-evolved and then randomly matched to resolve conflicts, and complete new particles are obtained, which improves the population diversity. At the same time, a pairwise competition mechanism is introduced to increase the source of particle social learning, avoid the situation of "premature" convergence, thereby obtaining a better distributed multi-satellite task collaborative observation planning scheme and improving the efficiency of satellite resource utilization. BRIEF DESCRIPTION OF THE DRAWINGS

[0100] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0101] Figure 1 A flow chart of a distributed multi-satellite mission planning method considering maximum benefit and load balancing provided by an embodiment of the present invention;

[0102] Figure 2 A schematic diagram of obtaining a particle gene vector provided by an embodiment of the present invention;

[0103] Figure 3 A schematic diagram of obtaining a subpopulation provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0104] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention are clearly and completely described. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0105] The embodiments of the present application solve the technical problem of not being able to fully improve the efficiency of satellite resource utilization by providing a distributed multi-satellite mission planning method, system, storage medium and electronic device that takes maximum benefit and load balancing into consideration.

[0106] The technical solution in the embodiment of the present application is to solve the above technical problems, and the overall idea is as follows:

[0107] In the embodiment of the present invention, a satellite resource set and a task set to be observed are first obtained; then, based on the satellite resource set and the task set, a multi-satellite task planning model with a preset maximum total observation benefit target and a constellation load balancing target is combined to obtain multiple feasible solutions for multi-satellite task planning; finally, based on the multiple feasible solutions, a distributed co-evolutionary particle swarm algorithm is used to obtain the final distributed multi-satellite task collaborative observation planning scheme. Considering the distributed satellite load balancing, a distributed co-evolutionary multi-satellite task planning model is constructed, and a distributed co-evolutionary particle swarm algorithm is designed. The particles are co-evolved and then randomly matched to resolve conflicts, and complete new particles are obtained, which improves the population diversity. At the same time, a pairwise competition mechanism is introduced to increase the source of particle social learning and avoid the situation of "premature" convergence, thereby obtaining a better distributed multi-satellite task collaborative observation planning scheme and improving the efficiency of satellite resource utilization.

[0108] In order to better understand the above technical solution, the above technical solution will be described in detail below in conjunction with the accompanying drawings and specific implementation methods.

[0109] Example:

[0110] like Figure 1 As shown, an embodiment of the present invention provides a distributed multi-satellite mission planning method considering maximum benefit and load balancing, including:

[0111] S1. Obtain a set of satellite resources and a set of tasks to be observed;

[0112] S2. Acquire multiple feasible solutions for multi-satellite mission planning according to the satellite resource set and the task set, in combination with a preset multi-satellite mission planning model that maximizes the total observation benefit target and the constellation load balancing target;

[0113] S3. According to the plurality of feasible solutions, a distributed co-evolutionary particle swarm algorithm is used to obtain a final distributed multi-satellite mission collaborative observation planning scheme.

[0114] The embodiment of the present invention takes distributed satellite load balancing into consideration, constructs a distributed co-evolutionary multi-satellite mission planning model, designs a distributed co-evolutionary particle swarm algorithm, co-evolves particles and then resolves conflicts by random matching to obtain complete new particles, thereby improving population diversity. At the same time, a pairwise competition mechanism is introduced to increase the source of particle social learning and avoid the occurrence of "premature" convergence, thereby obtaining a better distributed multi-satellite mission collaborative observation planning scheme and improving the efficiency of satellite resource utilization.

[0115] The following will introduce each step of the above technical solution in detail with specific content:

[0116] In step S1, a satellite resource set and a task set to be observed are obtained.

[0117] definition represents the satellite resource set, N S is the number of satellite resources; The set of tasks to be observed, N T is the number of tasks.

[0118] In step S2, according to the satellite resource set and the task set, a plurality of feasible solutions of the multi-satellite mission planning are obtained in combination with a preset multi-satellite mission planning model of maximizing the total observation benefit objective and the constellation load balancing objective.

[0119] The above multi-satellite mission planning model includes:

[0120] The objective function of maximizing the total observation benefit objective and the constellation load balancing objective is:

[0121]

[0122] Among them, λ1 and λ2 represent the weights of the total observation benefit target and the constellation load balancing target respectively;

[0123] Satellite resource collection, N S is the number of satellite resources;

[0124] The set of tasks to be observed, N T is the number of tasks;

[0125] For task t i In satellite S j The number of visible time windows on the

[0126] is the decision variable,

[0127]

[0128] p i For task t i Observational benefits of

[0129] CT j Satellite S j Observe the number of scheduled tasks;

[0130] LD is the load balancing evaluation value of the constellation, that is, the standard deviation of the number of tasks completed by each satellite in the planning scheme:

[0131]

[0132] And the constraints:

[0133]

[0134]

[0135]

[0136] ots ij +dur i =ote ij (7)

[0137] Among them, formula (4) means that a task is observed at most once;

[0138] Formula (5) indicates that the task observation needs to meet the time window requirement, that is, the execution time window of the task must be within the visible time window; For task t i In satellite S jThe ath visible time window on the previous page; For task t i In satellite S j The start time of the last visible time window; For task t i In satellite S j The end time of the last visible time window;

[0139] Formula (6) represents the time interval constraint of continuous observation tasks; tr i,i+1 Satellite S j Satellite sensor observation mission i and task t i+1 The time between posture transitions;

[0140] Formula (7) indicates that the observation end time of a task is equal to the sum of the observation start time and the task observation duration; dur i Satellite S j Satellite sensor mission t i Observation duration; ots ij 、ote ij Represents tasks t i In satellite S j The actual start and end time of the observation.

[0141] In step S3, based on the plurality of feasible solutions, a distributed collaborative evolutionary particle swarm algorithm is used to obtain a final distributed multi-satellite mission collaborative observation planning scheme.

[0142] First of all, it should be noted that in the distributed co-evolutionary particle swarm algorithm:

[0143] Based on the decimal encoding method of the task number, the initial particle population P = {p1, p2, ..., p tk ,…,p TK}, the total number of particles in the population is TK = K·N S , where p tk represents the tkth particle;

[0144] Particle p k It is composed of a gene matrix, which is divided into a sequence matrix X k and the weight matrix W k , X k The position of the gene in represents the execution order of the task; k With X k Correspondingly, W k The task weights in X k The task number in ;

[0145] like Figure 2The method shown in the figure (in the figure, p1 and p2 represent the sequence matrix and weight matrix corresponding to a particle respectively; dp1(j) and dp2(j) represent the number matrix and weight matrix corresponding to a particle respectively. j After the division, the corresponding sequence vector and weight vector) are used to divide the particle p tk Divide the rows (satellite resources) into differential oartucke (dp), then divide the particles dp k (j) is a gene vector corresponding to the particle gene matrix. The gene vector is divided into a sequence vector X k (j) and the weight vector W k (j), X k The position of the genes in (j) represents the execution order of the tasks; W k (j) and X k (j) Correspondingly, W k The task weight in (j) corresponds to X k The task number in (j);

[0146] At the tth generation of evolution, particle p k The serial number matrix for:

[0147]

[0148] Particle p k The weight matrix for:

[0149]

[0150] Then the particle dp k The sequence vector of (j) is:

[0151]

[0152] particle dp k The weight vector of :

[0153]

[0154] in, and Respectively represent: the particle dp at the tth generation of evolution k (j) In the corresponding scheme, satellite S j The sequence number and weight of the mth observation task;

[0155] Represents particle p k The historical optimal position achieved during the evolution from the initial position to the tth generation, The corresponding weight matrix is have:

[0156]

[0157] but Indicates the particle dp k (j) The historical optimal position achieved during the evolution from the initial position to the tth generation, The corresponding weight vector is have:

[0158]

[0159] in, express In the corresponding scheme, satellite S j The weight of the mth observation task;

[0160] gBest t Indicates the optimal position reached by the particle population from the initial position to the tth generation, gBest t The corresponding weight matrix is ​​gBestW t ,have:

[0161]

[0162] gBest t (j) represents the optimal position reached by the particle population from the initial position to the tth generation, gBest t (j) The corresponding weight vector is gBestW t (j) with:

[0163]

[0164] in, gBest t (j) Satellite S in the corresponding scheme j The weight of the mth observation task.

[0165] The S3 specifically includes:

[0166] S31, set t=0;

[0167] Each initial solution is put into the initial particle population as a particle. The total number of particles is TK, TK = K·N S ; Randomly initialize the position and velocity of each particle in the entire population;

[0168] S32, each particle p in the initial particle population tk The gene matrix is ​​split into strips according to the satellite number, and the satellite S j The corresponding particle gene vector dptk (j), forming a set of particle gene vectors TDP(j) = {dp1(j),…,dp tk (j),…,dp TK (j)}; the gene matrix includes a sequence matrix and a weight matrix;

[0169] S33. Calculate the particle gene vector dp tk (j) The fitness values ​​of all sub-particles;

[0170]

[0171] in, is the t-th generation particle dp tk (j) fitness value; w tk,j,m is the element in the weight matrix, i.e., the particle p tk (j) Satellite S in the corresponding scheme j The weight of the mth observation task;

[0172] And from the sub-particle gene vector set TDP(j), according to the elite selection strategy, K sub-particles are selected to form the initial sub-population DP(j) = {dp1(j),…,dp k (j),…,dp k (j)}, a total of N S Initial sub-populations: DP(1),…,DP(j),…,DP(N S );

[0173] S34. Define the initial positions and fitness values ​​of the sub-particles in the initial sub-population DP(j) as For the particle with the largest fitness value in the initial sub-population DP(j), its position and evaluation value are defined as gBest t (j) and

[0174] For the initial sub-population DP(j), the preset movement formula is used to perform sub-particle co-evolution to update the speed and position of the sub-particles;

[0175] The sub-particle coordinated evolution in S34 updates the speed and position of the sub-particles, specifically including:

[0176] Move the particle gene vector dp according to the movement formula k (j), at the t+1th generation of evolution, the particle dp k The movement formula is

[0177]

[0178]

[0179] in, Respectively represent the evolutionary time-division particle dp of the tth and t+1th generation k (j) In the corresponding scheme, satellite S j The weight of the i-th observation task is used to characterize the particle position;

[0180] dp respectively k (j) Movement speed during the evolution of the tth and t+1th generations;

[0181] ω is the inertia coefficient, which indicates the influence of the velocity of the sub-particles of the tth generation on the moving velocity of the sub-particles of the t+1th generation;

[0182] c1 is the sub-particle self-evolution acceleration, c2 is the sub-particle competition learning acceleration, and c3 is the sub-particle global movement acceleration;

[0183] r1, r2, r3 are random numbers uniformly distributed in [0, 1];

[0184] Indicates the particle dp k (j) According to the pairwise competition mechanism, the particles dp k′ (j),dp k′ (j)≠dp k (j) Learning: In pairwise competition, if dp k (j) Losing the game, i.e. AI k (j) <AI k′ (j), then in the next step, update the formula c2≠0, dp k (j) By submitting to dp k′ (j) learning to update its position, otherwise c2 = 0; the pairwise competition mechanism increases the source of particle social learning, which can greatly avoid the occurrence of "premature" convergence and improve the population diversity.

[0185] Indicates the particle dp k (j) Global social learning.

[0186] S35, such as Figure 3 As shown, select the sub-particles dp in the current minute group in turn k (j) Randomly select a sub-particle from each of the other current sub-populations to form a quasiparticle qp k , and then perform matching conflict resolution to obtain complete particles. Each satellite S j Get the subpopulation P(j), j=1,2,…,N S , the number of particles in the subpopulation P(j) is K;

[0187] The conflict resolution negotiation rules in S35 include:

[0188] When a task can be observed by different satellites, satellites with smaller loads are given priority to perform the observation task; when multiple tasks conflict, observation tasks with greater benefits are given priority.

[0189] S36. Calculate the particle p in each subpopulation P(j) k The location and fitness value of Thus, the position and fitness value of the particle with the largest fitness value in each sub-population P(j) are determined. AI t (gBest j );

[0190] Identify AI t (gBest j ) the smallest first subpopulation P(j), and AI t (gBest j ) the second sub-population P(j) with the largest fitness value, and exchange the particle with the smallest fitness value in the first sub-population P(j) with a randomly selected particle in the second sub-population P(j);

[0191] S37, using the movement formula to update the speed and position of all particles in the entire sub-population P(j);

[0192] In this step, the above formulas (17) and (18) are also used to update the speed and position of all particles in the subpopulation P(j), which will not be repeated here.

[0193] S38, calculate each particle p in the current sub-population P(j) k The current fitness value of renew and

[0194]

[0195] S39, adopt double stopping criteria:

[0196] When the number of iterations reaches the preset maximum number of evolutions Nmax, or the population convergence coefficient reaches the preset threshold ε0, the iteration is stopped and the corresponding sub-populations are compared. AI (gBest j ), determine the global optimal position of the current initial particle population TP and its fitness value gBest t 、AI gBest, decode the corresponding particles to obtain the final distributed multi-satellite mission collaborative observation planning scheme; otherwise, merge the current sub-populations P(j) to obtain a particle population and use it as the initial particle population TP in the next iteration process, set t=t+1, and return to S32.

[0197] The above population convergence coefficient is expressed as ε=|Fmax-F|, where Fmax is the maximum value of the population fitness; F is the average value of the population fitness.

[0198] Therefore, the embodiment of the present invention is applied to distributed co-evolutionary multi-satellite scheduling, taking into account the goal of maximizing the total observation benefit and the goal of balancing the load of the star cluster, and optimizing the particle swarm algorithm; designing the particle gene matrix and the sub-particle gene vector (the particle gene matrix is ​​split into strips to obtain the sub-particle gene matrix), constructing a distributed co-evolutionary particle swarm algorithm, combining the genetic algorithm mutation operation, avoiding the algorithm from falling into the local optimum, thereby obtaining a high-quality solution to the problem in a relatively short time, and completing the distributed multi-satellite mission planning.

[0199] The embodiment of the present invention provides a distributed multi-satellite mission planning system considering maximum benefit and load balancing, including:

[0200] An acquisition module is used to acquire a set of satellite resources and a set of tasks to be observed;

[0201] A selection module is used to obtain multiple feasible solutions for multi-satellite mission planning according to the satellite resource set and the task set, in combination with a preset multi-satellite mission planning model with a maximum total observation benefit target and a constellation load balancing target;

[0202] The planning module is used to obtain the final distributed multi-satellite mission collaborative observation planning scheme based on the multiple feasible solutions by using the distributed co-evolutionary particle swarm algorithm.

[0203] An embodiment of the present invention provides a storage medium storing a computer program for distributed multi-satellite mission planning considering maximum benefit and load balancing, wherein the computer program enables a computer to execute the distributed multi-satellite mission planning method as described above.

[0204] An embodiment of the present invention further provides an electronic device, including:

[0205] one or more processors;

[0206] Memory; and

[0207] One or more programs, wherein the one or more programs are stored in the memory and are configured to be executed by the one or more processors, the programs including programs for executing the distributed multi-satellite mission planning method as described above.

[0208] It can be understood that the distributed multi-satellite mission planning system, storage medium and electronic device considering maximum benefit and load balancing provided in the embodiments of the present invention correspond to the distributed multi-satellite mission planning method considering maximum benefit and load balancing provided in the embodiments of the present invention. The explanations, examples and beneficial effects of the relevant contents can refer to the corresponding parts in the distributed multi-satellite mission planning method, which will not be repeated here.

[0209] In summary, compared with the prior art, the present invention has the following beneficial effects:

[0210] The embodiment of the present invention takes distributed satellite load balancing into consideration, constructs a distributed co-evolutionary multi-satellite mission planning model, designs a distributed co-evolutionary particle swarm algorithm, co-evolves particles and then resolves conflicts by random matching to obtain complete new particles, thereby improving population diversity. At the same time, a pairwise competition mechanism is introduced to increase the source of particle social learning and avoid the occurrence of "premature" convergence, thereby obtaining a better distributed multi-satellite mission collaborative observation planning scheme and improving the efficiency of satellite resource utilization.

[0211] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0212] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A distributed multi-satellite mission planning method considering maximum benefit and load balancing, characterized in that: include: S1. Obtain a set of satellite resources and a set of tasks to be observed; S2. Acquire multiple feasible solutions for multi-satellite mission planning according to the satellite resource set and the task set, in combination with a preset multi-satellite mission planning model that maximizes the total observation benefit target and the constellation load balancing target; S3. According to the plurality of feasible solutions, a distributed co-evolutionary particle swarm algorithm is used to obtain a final distributed multi-satellite mission collaborative observation planning scheme; The multi-satellite mission planning model in S2 includes: The objective function of maximizing the total observation benefit objective and the constellation load balancing objective is: Among them, λ1 and λ2 represent the weights of the total observation benefit target and the constellation load balancing target respectively; Satellite resource collection, N S is the number of satellite resources; The set of tasks to be observed, N T is the number of tasks; For task t i In satellite S j The number of visible time windows on the is the decision variable, p i For task t i Observational benefits of CT j Satellite S j Observe the number of scheduled tasks; LD is the load balancing evaluation value of the constellation, that is, the standard deviation of the number of tasks completed by each satellite in the planning scheme: The multi-satellite mission planning model in S3 also includes constraints: oh ij +dur i =ote ij (7) Among them, formula (4) means that a task is observed at most once; Formula (5) indicates that the task observation needs to meet the time window requirement, that is, the execution time window of the task must be within the visible time window; For task t i In satellite S j The ath visible time window on the previous page; For task t i In satellite S j The start time of the last visible time window; For task t i In satellite S j The end time of the last visible time window; Formula (6) represents the time interval constraint of continuous observation tasks; tr i,i+1 Satellite S j Satellite sensor observation mission i and task t i+1 The time between posture transitions; Formula (7) indicates that the observation end time of a task is equal to the sum of the observation start time and the task observation duration; dur i Satellite S j Satellite sensor mission t i Observation duration; ots ij 、ote ij Represents tasks t i In satellite S j The start and end time of the actual observation on; The S3 specifically includes: S31, set t=0; Each initial solution is put into the initial particle population as a particle. The total number of particles is TK, TK = K·N S ; Randomly initialize the position and velocity of each particle in the entire population; S32, each particle p in the initial particle population tk The gene matrix is ​​split into strips according to the satellite number, and the satellite S j The corresponding particle gene vector dp tk (j), forming a set of particle gene vectors TDP(j) = {dp1(j),…,dp tk (j),…,dp TK (j)}; the gene matrix includes a sequence matrix and a weight matrix; S33. Calculate the particle gene vector dp tk (j) The fitness values ​​of all sub-particles; in, is the t-th generation particle dp tk (j) fitness value; w tk,j,m is the element in the weight matrix, i.e., the particle p tk (j) Satellite S in the corresponding scheme j The weight of the mth observation task; And from the sub-particle gene vector set TDP(j), according to the elite selection strategy, K sub-particles are selected to form the initial sub-population DP(j) = {dp1(j),…,dp k (j),…,dp K (j)}, a total of N S Initial sub-populations: DP(1),…,DP(j),…,DP(N S ); S34. Define the initial positions and fitness values ​​of the sub-particles in the initial sub-population DP(j) as For the particle with the largest fitness value in the initial sub-population DP(j), its position and evaluation value are defined as gBest t (j) and For the initial sub-population DP(j), the preset movement formula is used to perform sub-particle co-evolution to update the speed and position of the sub-particles; S35, select the sub-particles dp in the current minute group in turn k (j) Randomly select a sub-particle from each of the other current sub-populations to form a quasiparticle qp k , and then perform matching conflict resolution to obtain complete particles. Each satellite S j Get the subpopulation P(j), j=1,2,…,N S , the number of particles in the subpopulation P(j) is K; S36. Calculate the particle p in each subpopulation P(j) k The location and fitness value of Thus, the position and fitness value of the particle with the largest fitness value in each sub-population P(j) are determined. AI t (gBest j ); Identify AI t (gBest j ) the smallest first subpopulation P(j), and AI t (gBest j ) the second sub-population P(j) with the largest fitness value, and exchange the particle with the smallest fitness value in the first sub-population P(j) with a randomly selected particle in the second sub-population P(j); S37, using the movement formula to update the speed and position of all particles in the entire sub-population P(j); S38, calculate each particle p in the current sub-population P(j) k The current fitness value of renew and S39, when the number of iterations reaches the preset maximum number of evolutions Nmax, or the population convergence coefficient reaches the preset threshold ε0, the iteration is stopped, and the corresponding AI (gBest j ), determine the global optimal position of the current initial particle population TP and its fitness value gBest t 、AI gBest , decode the corresponding particles to obtain the final distributed multi-satellite mission collaborative observation planning scheme; otherwise, merge the current sub-populations P(j) to obtain a particle population and use it as the initial particle population TP in the next iteration process, set t=t+1, and return to S32.

2. The distributed multi-satellite mission planning method according to claim 1, characterized in that: In the distributed collaborative evolution particle swarm algorithm: Based on the decimal encoding method of the task number, the initial particle population P = {p1, p2, ..., p tk ,…,p TK }, the total number of particles in the population is TK = K·N S , where p tk represents the tkth particle; Particle p k It is composed of a gene matrix, which is divided into a sequence matrix X k and the weight matrix W k , X k The position of the gene in represents the execution order of the task; k With X k Correspondingly, W k The task weights in X k The task number in ; Then the particle dp k (j) is a gene vector corresponding to the particle gene matrix. The gene vector is divided into a sequence vector X k (j) and the weight vector W k (j), X k The position of the genes in (j) represents the execution order of the tasks; W k (j) and X k (j) Correspondingly, W k The task weight in (j) corresponds to X k The task number in (j); At the tth generation of evolution, particle p k The serial number matrix for: Particle p k The weight matrix for: Then the particle dp k The sequence vector of (j) is: particle dp k The weight vector of : in, and Respectively represent: the particle dp at the tth generation of evolution k (j) In the corresponding scheme, satellite S j The sequence number and weight of the mth observation task; Represents particle p k The historical optimal position achieved during the evolution from the initial position to the tth generation, The corresponding weight matrix is have: but Indicates the particle dp k (j) The historical optimal position achieved during the evolution from the initial position to the tth generation, The corresponding weight vector is have: in, express In the corresponding scheme, satellite S j The weight of the mth observation task; gBest t Indicates the optimal position reached by the particle population from the initial position to the tth generation, gBest t The corresponding weight matrix is ​​gBestW t ,have: gBest t (j) represents the optimal position reached by the particle population from the initial position to the tth generation, gBest t (j) The corresponding weight vector is gBestW t (j) has: in, gBest t (j) Satellite S in the corresponding scheme j The weight of the mth observation task.

3. The distributed multi-satellite mission planning method according to claim 2, characterized in that: The sub-particle coordinated evolution in S34 updates the speed and position of the sub-particles, specifically including: Move the particle gene vector dp according to the movement formula k (j), at the t+1th generation of evolution, the particle dp k The movement formula is in, Respectively represent the evolutionary time-division particle dp of the tth and t+1th generation k (j) In the corresponding scheme, satellite S j The weight of the i-th observation task is used to characterize the particle position; dp respectively k (j) Movement speed during the evolution of the tth and t+1th generations; ω is the inertia coefficient, which indicates the influence of the velocity of the sub-particles of the tth generation on the moving velocity of the sub-particles of the t+1th generation; c1 is the sub-particle self-evolution acceleration, c2 is the sub-particle competition learning acceleration, and c3 is the sub-particle global movement acceleration; r1, r2, r3 are random numbers uniformly distributed in [0, 1]; Indicates the particle dp k (j) According to the pairwise competition mechanism, the particles dp k′ (j),dp k′ (j)≠dp k (j) Learning: In pairwise competition, if dp k (j) Losing the game, i.e. AI k (j) <AI k′ (j), then in the next step, update the formula c2≠0, dp k (j) By submitting to dp k′ (j) Learn to update its position, otherwise c2 = 0; Indicates the particle dp k (j) Global social learning.

4. The distributed multi-satellite mission planning method according to any one of claims 1 to 3, characterized in that: The conflict resolution negotiation rules in S35 include: When a task can be observed by different satellites, satellites with smaller loads are given priority to perform the observation task; when multiple tasks conflict, observation tasks with greater benefits are given priority.

5. A distributed multi-satellite mission planning system considering maximum benefit and load balancing, characterized in that: The method for executing the distributed multi-satellite mission planning method according to claim 1 comprises: An acquisition module is used to acquire a set of satellite resources and a set of tasks to be observed; A selection module is used to obtain multiple feasible solutions for multi-satellite mission planning according to the satellite resource set and the task set, in combination with a preset multi-satellite mission planning model with a maximum total observation benefit target and a constellation load balancing target; The planning module is used to obtain the final distributed multi-satellite mission collaborative observation planning scheme based on the multiple feasible solutions by using the distributed co-evolutionary particle swarm algorithm.

6. A storage medium, characterized in that: It stores a computer program for distributed multi-satellite mission planning considering maximum benefit and load balancing, wherein the computer program enables a computer to execute the distributed multi-satellite mission planning method as described in any one of claims 1 to 4.

7. An electronic device, characterized in that: include: one or more processors; Memory; as well as One or more programs, wherein the one or more programs are stored in the memory and are configured to be executed by the one or more processors, and the programs include programs for executing the distributed multi-satellite mission planning method according to any one of claims 1 to 4.

Images