CN114997611A - Distributed multi-satellite task planning method considering maximum profit and load balance - Google Patents
Distributed multi-satellite task planning method considering maximum profit and load balance Download PDFInfo
- Publication number
- CN114997611A CN114997611A CN202210542792.XA CN202210542792A CN114997611A CN 114997611 A CN114997611 A CN 114997611A CN 202210542792 A CN202210542792 A CN 202210542792A CN 114997611 A CN114997611 A CN 114997611A
- Authority
- CN
- China
- Prior art keywords
- satellite
- task
- particle
- population
- observation
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0631—Resource planning, allocation, distributing or scheduling for enterprises or organisations
- G06Q10/06315—Needs-based resource requirements planning or analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/004—Artificial life, i.e. computing arrangements simulating life
- G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0631—Resource planning, allocation, distributing or scheduling for enterprises or organisations
- G06Q10/06312—Adjustment or analysis of established resource schedule, e.g. resource or task levelling, or dynamic rescheduling
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0631—Resource planning, allocation, distributing or scheduling for enterprises or organisations
- G06Q10/06316—Sequencing of tasks or work
Abstract
The invention provides a distributed multi-satellite task planning method and system considering maximum income and load balance, and relates to the technical field of multi-satellite task planning. In the invention, a satellite resource set and a task set to be observed are obtained; combining a preset multi-satellite task planning model of a maximum total observation income target and a constellation load balancing target to obtain a plurality of feasible solutions of multi-satellite task planning; and obtaining a final distributed multi-satellite task collaborative observation planning scheme by adopting a distributed collaborative evolution particle swarm algorithm. Considering distributed satellite load balance, constructing a multi-satellite task planning model; a distributed coevolution particle swarm algorithm is designed, particle coevolution is carried out, then random matching conflict resolution is carried out, complete new particles are obtained, population diversity is improved, meanwhile, a pairwise competition mechanism is introduced, the social learning source of the particles is increased, the early-maturing convergence condition is avoided, a better distributed multi-satellite task coevolution observation planning scheme is obtained, and the satellite resource utilization efficiency is improved.
Description
Technical Field
The invention relates to the technical field of multi-satellite task planning, in particular to a distributed multi-satellite task planning method, a distributed multi-satellite task planning system, a storage medium and electronic equipment considering maximum income and load balance.
Background
The imaging reconnaissance satellite runs in space with a certain orbit, an observation area which takes a sub-satellite point track as an axis and is determined by a field angle, a yaw angle, a pitch angle and the like is formed on the ground, the satellite has a strict time window for observing an object in the area, and certain observation resources are consumed.
The distributed imaging satellite collaborative observation scheduling problem can be described as: any satellite in a group of constellations receives the task information, and then transmits the task information to other satellites in the constellations through the inter-satellite link, the constellations jointly carry out task planning, and the reasonable planning scheme is formulated to maximize the task benefit and ensure the task planning efficiency. Heuristic algorithms such as a genetic algorithm, a particle swarm algorithm and the like are generally adopted to solve the problem of satellite emergency task scheduling.
The particle swarm optimization algorithm simulates predation behaviors of a bird swarm, is initialized to a group of random particles (random solution), the particles follow the optimal particles in a solution space to search, and then the optimal solution is found through iteration. The particle swarm algorithm is simple and easy to implement, and no parameters need to be adjusted, so that the particle swarm algorithm is widely applied to function optimization. However, because the particle swarm optimization algorithm flies towards the direction of the optimal solution according to the whole particles and the search experience of the particle swarm optimization algorithm, the convergence speed is obviously slowed down in the later stage of evolution, and the particle swarm optimization algorithm is easy to fall into a local extreme point when the algorithm converges to a certain precision; in addition, a phenomenon that particles oscillate near the optimal solution sometimes occurs. The algorithm cannot be optimized continuously, so that the accuracy achieved by the algorithm is poor, and the utilization efficiency of satellite resources cannot be improved sufficiently.
Disclosure of Invention
Technical problem to be solved
Aiming at the defects of the prior art, the invention provides a distributed multi-satellite task planning method, a distributed multi-satellite task planning system, a storage medium and electronic equipment considering maximum income and load balance, and solves the technical problem that the utilization efficiency of satellite resources cannot be fully improved.
(II) technical scheme
In order to achieve the purpose, the invention is realized by the following technical scheme:
a distributed multi-star mission planning method considering maximum revenue and load balancing comprises the following steps:
s1, acquiring a satellite resource set and a task set to be observed;
s2, according to the satellite resource set and the task set, combining a preset multi-satellite task planning model for maximizing a total observation income target and a constellation load balancing target to obtain a plurality of feasible solutions of multi-satellite task planning;
and S3, obtaining a final distributed multi-satellite task collaborative observation planning scheme by adopting a distributed collaborative evolution particle swarm algorithm according to the feasible solutions.
Preferably, the multi-star mission planning model in S2 includes:
the objective function of the total observation income target and the star group load balancing target is maximized:
wherein λ is 1 、λ 2 Respectively representing the weight of a total observation income target and the weight of a constellation load balancing target;
p i as task t i (ii) the observed yield;
CT j as a satellite S j Observing the number of the scheduled tasks;
LD is a constellation load balance evaluation value, i.e. the standard deviation of the number of tasks completed by each satellite in the planning scheme:
preferably, the multi-star mission planning model in S3 further includes a constraint condition:
ots ij +dur i =ote ij (7)
wherein, formula (4) indicates that a task is observed at most once;
formula (5) shows that the task needs to meet the time window requirement when being observed, namely the execution time window of the task is within the visible time window;as task t i At satellite S j The upper a-th visible time window;as task t i At satellite S j The upper a-th visible time window start time;as task t i At satellite S j The upper a-th visible time window end time;
equation (6) represents a continuous observation task time interval constraint; tr i,i+1 As a satellite S j Last satellite sensor observation task t i And task t i+1 Time to change posture;
formula (7) indicates that the observation end time of the task is equal to the sum of the observation start time and the observation duration of the task; dur i As a satellite S j Go up satellite sensor to task t i The duration of observation of (d); ots ij 、ote ij Respectively represent tasks t i At satellite S j The start time and the end time of the actual observation.
Preferably, the S3 specifically includes:
s31, making t equal to 0;
putting each initial solution as a particle into the initial particle population, wherein the total number of the particles is TK, and the TK is K.N S (ii) a Randomly initializing the position and speed of each particle in the whole population;
s32, mixing each particle p in the initial particle population tk The gene matrix is divided into strips according to the serial number of the satellite to obtain the satellite S j Corresponding particlized gene vector dpt k (j) A set of partial particle gene vectors TDP (j) ═ dp is constructed 1 (j),…,dp tk (j),…,dp TK (j) }; the gene matrix comprises a sequence number matrix and a weight matrix;
s33, calculating the gene vector dp of the particle tk (j) The fitness value of all the partial particles;
wherein the content of the first and second substances,is t generation minute particle dp tk (j) A fitness value of; w is a tk,j,m As an element in the weight matrix, i.e. a partial particle p tk (j) Corresponding solution satellite S j The weight of the mth observation task;
selecting K sub-particles from the sub-particle gene vector set TDP (j) according to an elite selection strategy to form an initial sub-population DP (j) { dp (j) } dp 1 (j),…,dp k (j),…,dp K (k) Get N altogether S Each initial sub-population: DP (1), …, DP (j), …, DP (N) S );
S34, defining the initial position and adaptability of the particles in the initial sub-population DP (j) asFor the component particle with the maximum fitness value in the initial classification group DP (j), the position and the evaluation value are respectively defined as gBest t (j) And
aiming at the initial seed group DP (j), a preset movement formula is adopted to execute the co-evolution of the sub-particles so as to update the speed and the position of the sub-particles;
s35, selecting the fractional particles dp in the current minute group in sequence k (j) Randomly selecting one sub-particle from other current sub-populations to form quasi-particle qp k Then matching conflict resolution is carried out to obtain complete particles, each satellite S j Obtaining the sub-population P (j), j is 1,2, …, N S The number of particles of the sub-population P (j) is K;
s36, calculating each particle p in each sub-population P (j) k Position of and fitness value thereofThereby determining the position of the particle with the maximum fitness value in each sub-population P (j) and the fitness value thereofAI t (gBest j );
Determining AI t (gBest j ) Minimum first sub-population P (j), and AI t (gBest j ) A second largest sub-population P (j), wherein the particles with the minimum fitness value in the first sub-population P (j) are exchanged with one particle randomly selected from the second sub-population P (j);
s37, updating the speed and the position of all the particles in the whole sub-population P (j) by adopting the moving formula;
s38, calculating each particle p in the current sub-population P (j) k Current fitness value ofUpdatingAnd
s39, stopping iteration when the iteration times reach the preset maximum evolution times Nmax or the population convergence coefficient reaches the preset threshold value epsilon 0, and comparing the corresponding sub-populationsAI(gBest j ) Determining the global optimum position and the fitness value gBest of the current initial particle population TP t 、AI gBest Decoding the corresponding particles to obtain a final distributed multi-satellite task collaborative observation planning scheme; otherwise, merging each current subspeciesGroup p (j) obtains a particle group as an initial particle group TP in the next iteration, let t be t +1, and return to S32.
Preferably, in the distributed coevolution particle swarm algorithm:
decimal coding mode based on task sequence number, and initial particle population P is { P ═ P 1 ,p 2 ,…,p tk ,…,p TK The total number of particles in the population is TK-K.N S Wherein p is tk Denotes the tk-th particle;
particle p k Is composed of gene matrix divided into sequence number matrix X k And weight matrix W k ,X k The position of the middle gene represents the execution sequence of the task; w k And X k Corresponding to, W k Task weight in (1) corresponds to X k The task number in (1);
then particle dp is divided k (j) For a gene vector corresponding to the particle gene matrix, the gene vector is divided into sequence number vectors X k (j) And weight vector W k (j),X k (j) The position of the middle gene represents the execution sequence of the task; w is a group of k (j) And X k (j) Corresponding to, W k (j) Task weight in (1) corresponds to X k (j) The task number in (1);
at the time of evolution of the t-th generation, particle p k Sequence number matrix ofComprises the following steps:
then particle dp is divided k (j) The sequence number vector of (1) is:
particle dp k The weight vector of (2):
wherein the content of the first and second substances,andrespectively represent: particle dp at the time of evolution of the t generation k (j) In a corresponding scheme, satellite S j The serial number and the weight of the last mth observation task;
represents a particle p k The historical optimal position reached in the process of evolution from the initial position to the t-th generation,the corresponding weight matrix isComprises the following steps:
thenRepresenting fractional particles dp k (j) The historical optimal position reached in the process of evolving from the initial position to the t-th generation,the corresponding weight vector isComprises the following steps:
wherein the content of the first and second substances,to representCorresponding solution satellite S j The weight of the mth observation task;
gBest t represents the optimal position, gBest, reached during the evolution of the population of particles from the initial position to the t-th generation t The corresponding weight matrix is gBestW t The method comprises the following steps:
then gBest t (j) Represents the optimal position, gBest, reached by the evolution of the particle-divided population from the initial position to the tth generation t (j) The corresponding weight vector is gBestW t (j) The method comprises the following steps:
wherein the content of the first and second substances,represents gBest t (j) Corresponding solution satellite S j And the weight of the mth observation task.
Preferably, the performing of the partical coevolution in S34 updates the speed and the position of the partical specifically includes:
moving the particle-divided gene vector dp according to the moving formula k (j) At the t +1 th generation, the particle dp is divided k Is the moving formula of
Wherein the content of the first and second substances,respectively representing t generation and t +1 generation evolution time-division particles dp k (j) In a corresponding scheme, satellite S j The weight of the ith observation task is used for representing the position of the particle;
omega is an inertia coefficient and represents the influence degree of the sub-particle speed of the t generation on the moving speed of the sub-particles of the t +1 generation;
c 1 for the acceleration of self evolution of the component particles, c 2 Learning acceleration for particle-based competition, c 3 Is the global movement acceleration of the partial particles;
r 1 ,r 2 ,r 3 is [0,1 ]]Random numbers uniformly distributed therein;
representing fractional particles dp k (j) According to a paired competition mechanism, dividing particles dp k′ (j),dp k′ (j)≠dp k (j) And (4) learning: in pairwise competition if dp k (j) Losing games, i.e. AI k (j)<AI k′ (j) Then c in the formula is updated in the next step 2 ≠0,dp k (j) Through to dp k′ (j) Learn to update its location, otherwise c 2 =0;
Preferably, the conflict resolution negotiation rule in S35 includes:
when the task can be observed by different satellites, the satellite with small load is preferentially selected to execute the observation task; and when the plurality of tasks conflict, the observation tasks with large profit are scheduled preferentially.
A distributed multi-star mission planning system that considers maximum revenue and load balancing, comprising:
the acquisition module is used for acquiring a satellite resource set and a task set to be observed;
the selection module is used for acquiring a plurality of feasible solutions of multi-satellite task planning by combining a preset multi-satellite task planning model for maximizing a total observation income target and a constellation load balancing target according to the satellite resource set and the task set;
and the planning module is used for acquiring a final distributed multi-satellite task collaborative observation planning scheme by adopting a distributed collaborative evolution particle swarm algorithm according to the feasible solutions.
A storage medium storing a computer program for distributed multi-star mission planning taking into account maximum revenue and load balancing, wherein the computer program causes a computer to perform the distributed multi-star mission planning method as described above.
An electronic device, comprising:
one or more processors;
a memory; and
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 comprising instructions for performing the distributed multi-star mission planning method as described above. .
(III) advantageous effects
The invention provides a distributed multi-satellite task planning method, a distributed multi-satellite task planning system, a storage medium and electronic equipment considering maximum income and load balance. Compared with the prior art, the method has the following beneficial effects:
firstly, acquiring a satellite resource set and a task set to be observed; then, according to the satellite resource set and the task set, combining a preset multi-satellite task planning model for maximizing a total observation income target and a constellation load balancing target to obtain a plurality of feasible solutions of multi-satellite task planning; and finally, according to the feasible solutions, obtaining a final distributed multi-satellite task collaborative observation planning scheme by adopting a distributed collaborative evolution particle swarm algorithm. Considering distributed satellite load balance, constructing a distributed co-evolution multi-satellite task planning model, designing a distributed co-evolution particle swarm algorithm, carrying out particle co-evolution and random matching conflict resolution to obtain complete new particles, improving population diversity, introducing a pairing competition mechanism, increasing the social learning source of the particles, avoiding the condition of premature convergence, obtaining a better distributed multi-satellite task co-observation planning scheme, and improving the utilization efficiency of satellite resources.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts.
Fig. 1 is a schematic flowchart of a distributed multi-star mission planning method considering maximum revenue and load balancing according to an embodiment of the present invention;
FIG. 2 is a schematic diagram of obtaining a particle-based gene vector according to an embodiment of the present invention;
fig. 3 is a schematic diagram of acquiring a seed population according to an embodiment of the present invention.
Detailed Description
In order to make the objects, 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, and it is obvious that the described embodiments are a part of the embodiments of the present invention, but not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The embodiment of the application provides a distributed multi-satellite task planning method, a distributed multi-satellite task planning system, a storage medium and electronic equipment considering maximum profit and load balance, and solves the technical problem that the utilization efficiency of satellite resources cannot be fully improved.
In order to solve the technical problems, the general idea of the embodiment of the application is as follows:
in the embodiment of the invention, a satellite resource set and a task set to be observed are obtained firstly; then, according to the satellite resource set and the task set, combining a preset multi-satellite task planning model for maximizing a total observation income target and a constellation load balancing target to obtain a plurality of feasible solutions of multi-satellite task planning; and finally, according to a plurality of feasible solutions, obtaining a final distributed multi-satellite task collaborative observation planning scheme by adopting a distributed collaborative evolution particle swarm algorithm. Considering distributed satellite load balance, constructing a distributed coevolution multi-satellite task planning model, designing a distributed coevolution particle swarm algorithm, performing particle coevolution and then performing random matching conflict resolution to obtain complete new particles, improving population diversity, introducing a pairwise competition mechanism, increasing the social learning source of the particles, avoiding the condition of premature convergence, thus obtaining a better distributed multi-satellite task collaborative observation planning scheme and improving the utilization efficiency of satellite resources.
In order to better understand the technical scheme, the technical scheme is described in detail in the following with reference to the attached drawings of the specification and specific embodiments.
The embodiment is as follows:
as shown in fig. 1, an embodiment of the present invention provides a distributed multi-star mission planning method considering maximum revenue and load balancing, including:
s1, acquiring a satellite resource set and a task set to be observed;
s2, according to the satellite resource set and the task set, combining a preset multi-satellite task planning model for maximizing a total observation income target and a constellation load balancing target to obtain a plurality of feasible solutions of multi-satellite task planning;
and S3, obtaining a final distributed multi-satellite task collaborative observation planning scheme by adopting a distributed collaborative evolution particle swarm algorithm according to the feasible solutions.
According to the embodiment of the invention, distributed satellite load balance is considered, a distributed co-evolution multi-satellite task planning model is constructed, a distributed co-evolution particle swarm algorithm is designed, particle co-evolution is carried out, then random matching conflict resolution is carried out, complete new particles are obtained, the population diversity is improved, meanwhile, a pairing competition mechanism is introduced, the social learning source of the particles is increased, the condition of premature convergence is avoided, a better distributed multi-satellite task co-observation planning scheme is obtained, and the utilization efficiency of satellite resources is improved.
The following will describe each step of the above technical solution in detail with reference to the specific content:
in step S1, a set of satellite resources and a set of tasks to be observed are acquired.
Definition ofRepresenting a set of satellite resources, N S The amount of satellite resources;set of tasks to be observed, N T Is the number of tasks.
In step S2, according to the satellite resource set and the task set, a preset multi-satellite task planning model that maximizes a total observation revenue target and a constellation load balancing target is combined to obtain a plurality of feasible solutions for multi-satellite task planning.
The multi-satellite mission planning model comprises:
maximizing the objective function of the total observation income objective and the constellation load balancing objective:
wherein λ is 1 、λ 2 Respectively representing the weight of a total observation income target and the weight of a constellation load balancing target;
p i as task t i The observation yield of (1);
CT j as a satellite S j Observing the number of the scheduled tasks;
LD is a constellation load balance evaluation value, i.e. the standard deviation of the number of tasks completed by each satellite in the planning scheme:
and the constraint condition is as follows:
ots ij +dur i =ote ij (7)
wherein, formula (4) indicates that a task is observed at most once;
formula (5) shows that the task needs to meet the time window requirement when being observed, namely the execution time window of the task is within the visible time window;as task t i At satellite S j The upper a-th visible time window;as task t i At satellite S j The upper a-th visible time window start time;as task t i At satellite S j The upper a-th visible time window end time;
equation (6) represents the continuous observation task time interval constraint; tr i,i+1 As a satellite S j Last satellite sensor observation task t i And task t i+1 Time to change posture;
formula (7) indicates that the observation end time of the task is equal to the sum of the observation start time and the observation duration of the task; dur i As a satellite S j Go up satellite sensor to task t i The duration of observation of (d); ots ij 、ote ij Respectively represent tasks t i At satellite S j The start time and the end time of the actual observation.
In step S3, according to the multiple feasible solutions, a final distributed multi-star task collaborative observation planning scheme is obtained by using a distributed collaborative evolution particle swarm algorithm.
It should be noted that, in the distributed coevolution particle swarm algorithm:
decimal coding mode based on task sequence number, and initial particle population P is { P ═ P 1 ,p 2 ,…,p tk ,…,p TK The total number of particles in the population is TK ═ K.N S Wherein p is tk Denotes the tk-th particle;
particle p k Is composed of gene matrix divided into sequence number matrix X k And weight matrix W k ,X k The position of the middle gene represents the execution sequence of the task; w k And X k Corresponding to, W k Task weight in (1) corresponds to X k The task number in (1);
the method shown in FIG. 2 (p in the figure) 1 、p 2 Respectively representing a serial number matrix and a weight matrix corresponding to a particle; dp 1 (j)、dp 2 (j) Respectively representing a component of the satellite S j After division, corresponding sequence number vector and weight vector), the particle p is divided into tk Divided into differant (dp) by row (satellite resource), then dp is k (j) For a gene vector corresponding to the particle gene matrix, the gene vector is divided into sequence number vectors X k (j) And weight vector W k (j),X k (j) The position of the middle gene represents the execution sequence of the task; w k (j) And X k (j) Corresponding to, W k (j) Task weight in (1) corresponds to X k (j) The task number in (1);
at the time of evolution of the t-th generation, particle p k Sequence number matrix ofComprises the following steps:
then the particle dp is divided k (j) The sequence number vector of (1) is:
particle dp k The weight vector of (2):
wherein, the first and the second end of the pipe are connected with each other,andrespectively represent: particle dp at the time of evolution of the t generation k (j) In a corresponding scheme, satellite S j The serial number and the weight of the last mth observation task;
represents a particle p k The historical optimal position reached in the process of evolving from the initial position to the t-th generation,the corresponding weight matrix isComprises the following steps:
thenRepresenting fractional particles dp k (j) The historical optimal position reached in the process of evolving from the initial position to the t-th generation,the corresponding weight vector isComprises the following steps:
wherein the content of the first and second substances,representCorresponding solution satellite S j The weight of the mth observation task;
gBest t represents the optimal position, gBest, reached during the evolution of the population of particles from the initial position to the t-th generation t The corresponding weight matrix is gBestW t The method comprises the following steps:
then gBest t (j) Represents the optimal position, gBest, reached during the evolution of the population of subparticles from the initial position to the t-th generation t (j) The corresponding weight vector is gBestW t (j) The method comprises the following steps:
wherein the content of the first and second substances,represents gBest t (j) Corresponding solution satellite S j And the weight of the mth observation task.
The S3 specifically includes:
s31, making t equal to 0;
putting each initial solution as a particle into the initial particle population, wherein the total number of the particles is TK, and the TK is K.N S (ii) a Randomly initializing the position and speed of each particle in the whole population;
s32, mixing each particle p in the initial particle population tk The gene matrix is divided into strips according to the satellite serial number to obtain the satellite S j Corresponding particlized gene vector dp tk (j) A set of partial particle gene vectors TDP (j) ═ dp is constructed 1 (j),…,dp tk (j),…,dp TK (j) }; the gene matrix comprises a sequence number matrix and a weight matrix;
s33, calculating the gene vector dp of the partical tk (j) The fitness value of all the partial particles;
wherein the content of the first and second substances,is t generation minute particle dp tk (j) A fitness value of; w is a tk,j,m As an element in the weight matrix, i.e. a partial particle p tk (j) Corresponding solution satellite S j The weight of the mth observation task;
selecting K sub-particles from the sub-particle gene vector set TDP (j) according to an elite selection strategy to form an initial sub-population DP (j) { dp (j) } dp 1 (j),…,dp k (j),…,dp k (j) Get N altogether S Each initial sub-population: DP (1), …, DP (j), …, DP (N) S );
S34, defining the initial position and adaptability of the particles in the initial sub-population DP (j) asFor the component particle with the maximum fitness value in the initial classification group DP (j), the position and the evaluation value are respectively defined as gBest t (j) And
aiming at the initial seed group DP (j), a preset movement formula is adopted to execute the particle coevolution to update the speed and the position of the particles;
in S34, performing partical coevolution to update the speed and position of the partical, specifically including:
moving the particle-divided gene vector dp according to the moving formula k (j) At the t +1 th generation, the particle dp is divided k Is the moving formula of
Wherein the content of the first and second substances,respectively represents the t generation evolution and the t +1 generation evolution time-division particle dp k (j) In a corresponding scheme, satellite S j The weight of the ith observation task is used for representing the position of the particle;
omega is an inertia coefficient and represents the influence degree of the sub-particle speed of the t generation on the moving speed of the sub-particles of the t +1 generation;
c 1 for the acceleration of self-evolution of the molecule, c 2 For the competition of particles for learning acceleration, c 3 Is the global moving acceleration of the particle;
r 1 ,r 2 ,r 3 is [0,1 ]]Random numbers uniformly distributed therein;
representing fractional particles dp k (j) According to a paired competition mechanism, dividing particles dp k′ (j),dp k′ (j)≠dp k (j) And (4) learning: in pairwise competition if dp k (j) Losing games, i.e. AI k (j)<AI k′ (j) Then c in the formula is updated next step 2 ≠0,dp k (j) By directing toward dp k′ (j) Learn to update its location, otherwise c 2 0; the paired competition mechanism increases the source of particle social learning, can greatly avoid the condition of premature convergence, and improves the population diversity.
S35, as shown in FIG. 3, the particle dp in the current minute group is selected in turn k (j) Randomly selecting one sub-particle from other current sub-populations to form quasi-particle qp k Then carrying out matching conflict resolution to obtain complete particles, each satellite S j Obtaining a sub-population P (j), j is 1,2, …, N S The number of particles of the sub-population P (j) is K;
the conflict resolution negotiation rule in S35 includes:
when the task can be observed by different satellites, the satellite with small load is preferentially selected to execute the observation task; and when the plurality of tasks conflict, the observation tasks with large profit are scheduled preferentially.
S36, calculating each particle p in each sub-population P (j) k Position of and fitness value thereofThereby determining the position of the particle with the maximum fitness value in each sub-population P (j) and the fitness value thereofAI t (gBest j );
Determining AI t (gBest j ) Minimum first sub-population P (j), and AI t (gBest j ) A second largest sub-population P (j), wherein the particles with the minimum fitness value in the first sub-population P (j) are exchanged with one particle randomly selected from the second sub-population P (j);
s37, updating the speed and the position of all particles in the whole sub-population P (j) by adopting the moving formula;
in this step, the velocity and position of all particles in the sub-population p (j) are updated by the above equations (17) and (18), which are not described herein again.
S38, calculating each particle p in the current sub-population P (j) k Current fitness value ofUpdatingAnd
s39, adopting a double stopping criterion:
stopping iteration when the iteration times reach a preset maximum evolution time Nmax or the population convergence coefficient reaches a preset threshold value epsilon 0, and comparing the iteration times with the preset maximum evolution time NmaxCorresponding to a current sub-populationAI(gBest j ) Determining the global optimum position and the fitness value gBest of the current initial particle population TP t 、AI gBest Decoding the corresponding particles to obtain a final distributed multi-satellite task collaborative observation planning scheme; otherwise, combining the current sub-populations p (j) to obtain a particle population, and using the particle population as the initial particle population TP in the next iteration process, making t equal to t +1, and returning to S32.
The population convergence coefficient is represented as ∈ ═ Fmax-F |, where Fmax is the maximum value of population fitness; and F is the average value of population fitness.
Therefore, the embodiment of the invention is applied to distributed coevolution multi-satellite scheduling, takes the aim of maximizing the total observation yield target and the star group load balancing target as the aim, and optimizes the particle swarm algorithm; designing a particle gene matrix and a particle-divided gene vector (splitting the particle gene matrix into strips to obtain the particle-divided gene matrix), constructing a distributed coevolution particle swarm algorithm, combining the genetic algorithm variation operation, and avoiding the algorithm from falling into local optimization, thereby obtaining a solution with high problem quality in a short time and completing distributed multi-satellite task planning.
The embodiment of the invention provides a distributed multi-satellite task planning system considering maximum profit and load balance, which comprises:
the acquisition module is used for acquiring a satellite resource set and a task set to be observed;
the selection module is used for acquiring a plurality of feasible solutions of multi-satellite task planning by combining a preset multi-satellite task planning model for maximizing a total observation income target and a constellation load balancing target according to the satellite resource set and the task set;
and the planning module is used for acquiring a final distributed multi-satellite task collaborative observation planning scheme by adopting a distributed collaborative evolution particle swarm algorithm according to the feasible solutions.
An embodiment of the present invention provides a storage medium storing a computer program for distributed multi-star mission planning considering maximum revenue and load balancing, wherein the computer program causes a computer to execute the distributed multi-star mission planning method as described above.
An embodiment of the present invention further provides an electronic device, including:
one or more processors;
a memory; and
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 comprising instructions for performing the distributed multi-star mission planning method as described above.
It can be understood that the distributed multi-satellite mission planning system, the storage medium, and the electronic device considering the maximum profit and the load balancing provided in the embodiment of the present invention correspond to the distributed multi-satellite mission planning method considering the maximum profit and the load balancing provided in the embodiment of the present invention, and the explanation, the example, the beneficial effects, and other parts of the relevant contents may refer to the corresponding parts in the distributed multi-satellite mission planning method, and are not described herein again.
In summary, compared with the prior art, the method has the following beneficial effects:
according to the embodiment of the invention, distributed satellite load balancing is considered, a distributed co-evolution multi-satellite task planning model is constructed, a distributed co-evolution particle swarm algorithm is designed, particle co-evolution is carried out, then random matching conflict resolution is carried out, complete new particles are obtained, the population diversity is improved, meanwhile, a pairwise competition mechanism is introduced, the source of particle social learning is increased, the condition of premature convergence is avoided, a better distributed multi-satellite task co-observation planning scheme is obtained, and the utilization efficiency of satellite resources is improved.
It should be noted that, in this document, relational terms such as first and second, and the like are used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other identical elements in a process, method, article, or apparatus that comprises the element.
The above examples are only intended to illustrate the technical solution of the present invention, but not to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.
Claims (10)
1. A distributed multi-star mission planning method considering maximum revenue and load balancing is characterized by comprising the following steps:
s1, acquiring a satellite resource set and a task set to be observed;
s2, according to the satellite resource set and the task set, combining a preset multi-satellite task planning model for maximizing a total observation income target and a constellation load balancing target to obtain a plurality of feasible solutions of multi-satellite task planning;
and S3, obtaining a final distributed multi-satellite task collaborative observation planning scheme by adopting a distributed collaborative evolution particle swarm algorithm according to the feasible solutions.
2. The distributed multi-star mission planning method of claim 1, wherein the multi-star mission planning model in S2 comprises:
maximizing the objective function of the total observation income objective and the constellation load balancing objective:
wherein λ is 1 、λ 2 Respectively representing the weight of a total observation income target and the weight of a constellation load balancing target;
p i as task t i The observation yield of (1);
CT j as a satellite S j Observing the number of the scheduled tasks;
LD is a constellation load balance evaluation value, i.e. the standard deviation of the number of tasks completed by each satellite in the planning scheme:
3. the distributed multi-star mission planning method of claim 2, wherein said multi-star mission planning model in S3 further comprises constraints:
ots ij +dur i =ote ij (7)
wherein, formula (4) indicates that a task is observed at most once;
formula (5) indicates that the time window requirement needs to be met when the task is observed, namely the execution time window of the task is within the visible time window;as task t i At satellite S j The upper a-th visible time window;as task t i At satellite S j The upper a-th visible time window start time;as task t i At satellite S j The upper a-th visible time window end time;
equation (6) represents the continuous observation task time interval constraint; tr is i,i+1 As a satellite S j Last satellite sensor observation task t i And task t i+1 Time to change posture;
formula (7) indicates that the observation end time of the task is equal to the sum of the observation start time and the observation duration of the task; dur i As a satellite S j Go up satellite sensor to task t i The duration of observation of (d); ots ij 、ote ij Respectively represent tasks t i At satellite S j The start time and the end time of the actual observation.
4. The distributed multi-star mission planning method according to claim 2 or 3, wherein the S3 specifically includes:
s31, making t equal to 0;
putting each initial solution as a particle into the initial particle population, wherein the total number of the particles is TK, and the TK is K.N S (ii) a Randomly initializing the position and speed of each particle in the whole population;
s32, mixing each particle p in the initial particle population tk The gene matrix is divided into strips according to the satellite serial number to obtain the satellite S j Corresponding particlized gene vector dp tk (j) A set of partial particle gene vectors TDP (j) ═ dp is constructed 1 (j),…,dp tk (j),…,dp TK (j) }; the gene matrix comprises a sequence number matrix and a weight matrix;
s33, calculating the gene vector dp of the particle tk (j) Fitness values of all the partial particles;
wherein the content of the first and second substances,is the t generation divided particle dp tk (j) A fitness value of; w is a tk,j,m As elements in a weight matrix, i.e. partial particles p tk (j) Corresponding solution satellite S j The weight of the mth observation task;
and selecting K scores from the score particle gene vector set TDP (j) according to an elite selection strategyInitial particle composition fraction dp (j) ═ dp 1 (j),…,dp k (j),…,dp K (k) Get N altogether S Each initial sub-population: DP (1), …, DP (j), …, DP (N) S );
S34, defining the initial position and adaptability of the particles in the initial sub-population DP (j) asFor the component particle with the maximum fitness value in the initial classification group DP (j), the position and the evaluation value are respectively defined as gBest t (j) Andaiming at the initial seed group DP (j), a preset movement formula is adopted to execute the particle coevolution to update the speed and the position of the particles;
s35, selecting the fractional particles dp in the current minute group in sequence k (j) Randomly selecting one sub-particle from other current sub-populations to form quasi-particle qp k Then carrying out matching conflict resolution to obtain complete particles, each satellite S j Obtaining a sub-population P (j), j is 1,2, …, N S The number of particles of the sub-population P (j) is K;
s36, calculating each particle p in each sub-population P (j) k Position of and fitness value thereofThereby determining the position of the particle with the maximum fitness value in each sub-population P (j) and the fitness value thereofAI t (gBest j );
Determining AI t (gBest j ) Minimum first sub-population P (j), and AI t (gBest j ) A second largest sub-population P (j), wherein the particles with the minimum fitness value in the first sub-population P (j) and one particle randomly selected from the second sub-population P (j) are exchanged;
s37, updating the speed and the position of all particles in the whole sub-population P (j) by adopting the moving formula;
s38, calculating each particle p in the current sub-population P (j) k Current fitness value ofUpdatingAnd
s39, stopping iteration when the iteration times reach the preset maximum evolution times Nmax or the population convergence coefficient reaches the preset threshold value epsilon 0, and comparing the corresponding sub-populationsAI(gBest j ) Determining the global optimum position and the fitness value gBest of the current initial particle population TP t 、AI gBest Decoding the corresponding particles to obtain a final distributed multi-satellite task collaborative observation planning scheme; otherwise, combining the current sub-populations p (j) to obtain a particle population, and using the particle population as the initial particle population TP in the next iteration process, making t equal to t +1, and returning to S32.
5. The distributed multi-star mission planning method of claim 4, wherein in said distributed coevolution particle swarm algorithm:
decimal coding mode based on task sequence number, and initial particle population P is { P ═ P 1 ,p 2 ,…,p tk ,…,p TK The total number of particles in the population is TK-K.N S Wherein p is tk Represents the kth numberParticles;
particle p k Is composed of gene matrix divided into sequence number matrix X k And weight matrix W k ,X k The position of the middle gene represents the execution sequence of the task; w k And X k Corresponding to, W k Task weight in (1) corresponds to X k The task number in (1);
then particle dp is divided k (j) For a gene vector corresponding to the particle gene matrix, the gene vector is divided into sequence number vectors X k (j) And weight vector W k (j),X k (j) The position of the middle gene represents the execution sequence of the task; w k (j) And X k (j) Corresponding to, W k (j) Task weight in (1) corresponds to X k (j) The task number in (1);
at the time of evolution of the t-th generation, particle p k Sequence number matrix ofComprises the following steps:
then the particle dp is divided k (j) The sequence number vector of (1) is:
particle dp k The weight vector of (2):
wherein the content of the first and second substances,andrespectively represent: particle dp at the time of evolution of the t generation k (j) In a corresponding scheme, satellite S j The serial number and the weight of the last mth observation task;
represents a particle p k The historical optimal position reached in the process of evolution from the initial position to the t-th generation,the corresponding weight matrix isComprises the following steps:
thenRepresenting fractional particles dp k (j) The historical optimal position reached in the process of evolving from the initial position to the t-th generation,the corresponding weight vector isComprises the following steps:
wherein, the first and the second end of the pipe are connected with each other,to representCorresponding solution satellite S j The weight of the mth observation task;
gBest t represents the optimal position, gBest, reached during the evolution of the population of particles from the initial position to the t-th generation t The corresponding weight matrix is gBestW t The method comprises the following steps:
then gBest t (j) Represents the optimal position, gBest, reached by the evolution of the particle-divided population from the initial position to the tth generation t (j) The corresponding weight vector is gBestW t (j) The method comprises the following steps:
6. The distributed multi-star mission planning method of claim 5,
in S34, performing partical coevolution to update the speed and position of the partical, specifically including:
moving the particle-divided gene vector dp according to the movement formula k (j) At the t +1 th generation, the particle dp is divided k Is the moving formula of
Wherein the content of the first and second substances,respectively representing t generation and t +1 generation evolution time-division particles dp k (j) In a corresponding scheme, satellite S j The weight of the ith observation task is used for representing the position of the particle;
omega is an inertia coefficient and represents the influence degree of the sub-particle speed of the t generation on the moving speed of the sub-particles of the t +1 generation;
c 1 for the acceleration of self evolution of the component particles, c 2 For the competition of particles for learning acceleration, c 3 Is the global movement acceleration of the partial particles;
r 1 ,r 2 ,r 3 is [0,1 ]]Random numbers uniformly distributed therein;
representing fractional particles dp k (j) According to a paired competition mechanism, dividing particles dp k′ (j),dp k′ (j)≠dp k (j) And (4) learning: in pairwise competition if dp k (j) Losing games, i.e. AI k (j)<AI k′ (j) Then c in the formula is updated in the next step 2 ≠0,dp k (j) By directing toward dp k′ (j) Learn to update its location, otherwise c 2 =0;
7. The distributed multi-star mission planning method according to any one of claims 3 to 6, wherein the conflict resolution negotiation rules in S35 include:
when the task can be observed by different satellites, the satellite with small load is preferentially selected to execute the observation task; and when the plurality of tasks conflict, the observation tasks with large profit are scheduled preferentially.
8. A distributed multi-star mission planning system that considers maximum revenue and load balancing, comprising:
the acquisition module is used for acquiring a satellite resource set and a task set to be observed;
the selection module is used for acquiring a plurality of feasible solutions of multi-satellite task planning by combining a preset multi-satellite task planning model for maximizing a total observation income target and a constellation load balancing target according to the satellite resource set and the task set;
and the planning module is used for acquiring a final distributed multi-satellite task collaborative observation planning scheme by adopting a distributed collaborative evolution particle swarm algorithm according to the feasible solutions.
9. A storage medium storing a computer program for distributed multi-star mission planning with maximum revenue and load balancing taken into account, wherein the computer program causes a computer to perform the distributed multi-star mission planning method according to any one of claims 1 to 7.
10. An electronic device, comprising:
one or more processors;
a memory; and
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 comprising instructions for performing the distributed multi-star mission planning method of any of claims 1-7.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210542792.XA CN114997611A (en) | 2022-05-18 | 2022-05-18 | Distributed multi-satellite task planning method considering maximum profit and load balance |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210542792.XA CN114997611A (en) | 2022-05-18 | 2022-05-18 | Distributed multi-satellite task planning method considering maximum profit and load balance |
Publications (1)
Publication Number | Publication Date |
---|---|
CN114997611A true CN114997611A (en) | 2022-09-02 |
Family
ID=83027425
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202210542792.XA Pending CN114997611A (en) | 2022-05-18 | 2022-05-18 | Distributed multi-satellite task planning method considering maximum profit and load balance |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN114997611A (en) |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN115686874A (en) * | 2023-01-03 | 2023-02-03 | 中南大学 | Dynamic inter-satellite multi-satellite cooperative computing method, system, equipment and medium |
CN116109036A (en) * | 2023-01-12 | 2023-05-12 | 中国科学院软件研究所 | Regional gridding multi-star observation planning method, device and equipment |
CN116450366A (en) * | 2023-06-19 | 2023-07-18 | 中国人民解放军战略支援部队航天工程大学 | Multi-objective optimal scheduling method, equipment and medium for satellite measurement, operation and control resources |
CN116915311A (en) * | 2023-07-14 | 2023-10-20 | 中国人民解放军战略支援部队航天工程大学 | Satellite task scheduling method based on co-evolution algorithm |
CN116915311B (en) * | 2023-07-14 | 2024-04-30 | 中国人民解放军战略支援部队航天工程大学 | Satellite task scheduling method based on co-evolution algorithm |
-
2022
- 2022-05-18 CN CN202210542792.XA patent/CN114997611A/en active Pending
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN115686874A (en) * | 2023-01-03 | 2023-02-03 | 中南大学 | Dynamic inter-satellite multi-satellite cooperative computing method, system, equipment and medium |
CN116109036A (en) * | 2023-01-12 | 2023-05-12 | 中国科学院软件研究所 | Regional gridding multi-star observation planning method, device and equipment |
CN116109036B (en) * | 2023-01-12 | 2023-11-07 | 中国科学院软件研究所 | Regional gridding multi-star observation planning method, device and equipment |
CN116450366A (en) * | 2023-06-19 | 2023-07-18 | 中国人民解放军战略支援部队航天工程大学 | Multi-objective optimal scheduling method, equipment and medium for satellite measurement, operation and control resources |
CN116450366B (en) * | 2023-06-19 | 2023-08-18 | 中国人民解放军战略支援部队航天工程大学 | Multi-objective optimal scheduling method, equipment and medium for satellite measurement, operation and control resources |
CN116915311A (en) * | 2023-07-14 | 2023-10-20 | 中国人民解放军战略支援部队航天工程大学 | Satellite task scheduling method based on co-evolution algorithm |
CN116915311B (en) * | 2023-07-14 | 2024-04-30 | 中国人民解放军战略支援部队航天工程大学 | Satellite task scheduling method based on co-evolution algorithm |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN114997611A (en) | Distributed multi-satellite task planning method considering maximum profit and load balance | |
He et al. | A generic Markov decision process model and reinforcement learning method for scheduling agile earth observation satellites | |
EP3406531B1 (en) | Innovative satellite scheduling method based on genetic algorithms and simulated annealing and related mission planner | |
Duan et al. | Rl $^ 2$: Fast reinforcement learning via slow reinforcement learning | |
Du et al. | A data-driven parallel scheduling approach for multiple agile earth observation satellites | |
CN110412869B (en) | Multi-satellite cooperative space dynamic target real-time tracking method | |
CN111913788B (en) | Task scheduling method and system for imaging satellite | |
CN111913787B (en) | Imaging satellite scheduling method and system based on genetic algorithm | |
CN110400002B (en) | Multi-satellite imaging task planning method | |
CN112198870B (en) | Unmanned aerial vehicle autonomous guiding maneuver decision method based on DDQN | |
CN107679748B (en) | Star-ground combined operation method for autonomous planning of constellation observation task | |
Wu et al. | Ensemble of metaheuristic and exact algorithm based on the divide-and-conquer framework for multisatellite observation scheduling | |
CN114185362A (en) | Unmanned aerial vehicle cluster task dynamic allocation method based on suburb information entropy | |
Lin et al. | Nonlinear system control using self-evolving neural fuzzy inference networks with reinforcement evolutionary learning | |
CN113051815A (en) | Agile imaging satellite task planning method based on independent pointer network | |
CN116090342B (en) | Distributed task allocation method for large-scale unmanned aerial vehicle based on alliance forming game | |
Madani et al. | Cooperative coevolutionary multi-guide particle swarm optimization algorithm for large-scale multi-objective optimization problems | |
CN114399161A (en) | Multi-unmanned aerial vehicle cooperative task allocation method based on discrete mapping differential evolution algorithm | |
Peng et al. | Continual match based training in Pommerman: Technical report | |
Xu et al. | An improved league championship algorithm with free search and its application on production scheduling | |
CN115374933A (en) | Intelligent planning and decision-making method for landing behavior of multi-node detector | |
Zhou et al. | Multi-satellite scheduling problem with marginal decreasing imaging duration: An improved adaptive ant colony algorithm | |
CN115329595B (en) | Unmanned aerial vehicle cluster task planning method and system based on knowledge and experience | |
CN114997610A (en) | Multi-satellite task planning method and system based on distributed cooperative particle swarm optimization | |
Zhou et al. | A novel mission planning method for UAVs’ course of action |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination |