A Terahertz Telescope Observation Task Scheduling Method Based on Genetic Algorithm

Abstract

The present invention discloses a method for scheduling observation tasks of a terahertz telescope based on a genetic algorithm. Combining the specific constraints of the observation task scheduling of a space terahertz telescope, innovations are made in chromosome coding, population individual construction, and fitness function definition. It mainly includes seven functional parts: discretization of observation scheduling time, construction of scheduling chromosomes, calculation of fitness functions, construction of parents, genetic operations of crossover and mutation, screening of offspring, and obtaining the finally optimized scheduling individuals. By quantifying the constraints for observing terahertz target sources, population individuals corresponding to legal scheduling schemes are constructed. The observation completion degree of all target sources is used as the fitness function value of individuals and is quantified. The Roulette Wheel selection method is used to screen the parents for offspring inheritance, effectively reducing the number of generations and population size required for the scheduling optimization problem.

Description

Technical Field

[0001] The present invention belongs to the technical field of astronomical observation, and particularly relates to a method for scheduling observation tasks of a terahertz telescope based on a genetic algorithm. Background Technique

[0002] The terahertz frequency band range refers to 0.1 - 10 THz. There are rich molecular rotational spectral lines and atomic fine structure spectral lines in this frequency band. Terahertz band astronomical observation is an effective means for studying astrophysics such as galaxy and star formation. However, limited by the influence of water vapor in the atmosphere, some spectral line information is absorbed by the atmosphere. Therefore, space is an inevitable location for some terahertz band observations.

[0003] The China Space Survey Telescope (CSST) is a flagship project of Chinese space astronomy and is expected to be launched around 2024. CSST is the largest space telescope independently developed in China, with outstanding features such as a large field of view, high image quality, and wide frequency band. Its detection sensitivity and spatial resolution are comparable to those of the internationally famous Hubble Space Telescope (HST), but the observation field of view and the scale of data acquisition will far exceed those of the Hubble Space Telescope. Multiple terminal detection modules are carried on CSST, which is highly competitive and is expected to make major breakthroughs in research directions such as cosmology, galaxies and active galactic nuclei, the Milky Way and stars, astrometry, exoplanets, and solar system celestial bodies. Among them, a high-sensitivity terahertz detection module (hereinafter referred to as the terahertz module for short) is carried, and its main scientific goals include the formation and evolution of molecular clouds, the neutral carbon survey of nearby galaxies, and the terahertz spectral line survey in the 410 - 510 GHz frequency band.

[0004] Compared with ground-based telescopes, space terahertz telescopes will face many restrictive conditions when conducting scientific mission observations: 1. The change in the operating orbit of the space telescope will cause the observed target to be blocked by the Earth, resulting in the unobservability of the observed target; 2. The change in the operating orbit of the space telescope will cause the angle between the receiving antenna of the telescope and the sun and the moon to change continuously. When the angles between the receiving antenna of the telescope and the sun and the moon are too small, the detector module of the telescope will be affected by strong interference signals from the sun and the moon, resulting in the inability of the detector module to observe normally; 3. The change in the operating orbit of the space station causes the sub-satellite point of the terahertz module to move on the Earth's surface. When the sub-satellite point falls in the South Atlantic Anomaly (SAA), all the equipment on the entire space station will be affected by strong geomagnetic interference. To protect the equipment on the space station, all the observation equipment on the space station will suspend observations in this case; 4. The energy provided by the solar panels on the space station is variable and the total amount is limited, so it is impossible to observe all the carried observation modules simultaneously; 5. The total amount of heat dissipation on the space station is limited, and the operating temperature of the control moment gyroscope has strict constraints on the number of maneuvers allowed within one orbit time, which will limit the total change in the antenna pointing angle during the observation of the telescope within the on-orbit time.

[0005] In addition to the observation constraints, the weights of the scientific significance of the observed targets are also different. Therefore, for space observation tasks with highly fragmented on-orbit continuous observable time, it is particularly necessary to achieve the maximum scientific output within the limited observable time. In view of this, conducting research on the optimization and scheduling method of the observation tasks for space terahertz telescopes to complete as many observations of the target sources to be measured as possible under various restrictive conditions has very important practical significance for improving the observation efficiency of the telescope and promoting the scientific output of the space telescope.

[0006] Internationally, the famous Hubble Space Telescope has long listed the autonomous observation task scheduling method as an important supporting technology for its space observations. The research and development team of the Hubble Telescope developed a graphical observation task scheduling software called Spike for the actual constraints such as the orbital operation status of the Hubble Space Telescope, the observation conditions of the payloads, and the scientific priority of the observation tasks, and provided a variety of optimization scheduling algorithms to calculate the optimized scheduling results. Ground operation and control personnel can very easily configure the constraint condition parameters of the observation tasks through the Spike software and present the calculated scheduling results in a friendly manner. After the Spike software achieved success on the Hubble Telescope, it was also used by the later SIRTF (Space Infrared Telescope Facility) and AXAF (Advanced X-ray Astrophysics Facility) for their observation task scheduling. Spike is a non-open-source program and is embargoed against our country.

[0007] With the steady progress of the CSST research and development task, the space telescope will pose a huge demand for the optimized scheduling technology of on-orbit observation tasks. Referring to NASA's Hubble Space Station Telescope, CSST needs to establish a systematic scheduling architecture to support its diverse payloads for observations. In addition, it should also have a variety of scheduling optimization algorithms that can meet the actual scheduling of the sky survey observation tasks under different observation strategies. Summary of the Invention

[0008] In view of the deficiencies in the prior art, the present invention provides a method for scheduling observation tasks of a terahertz telescope based on a genetic algorithm, and the technical solutions adopted are as follows:

[0009] A method for scheduling observation tasks of a terahertz telescope based on a genetic algorithm includes the following steps:

[0010] S1. Perform genetic encoding on the observation scheduling of the target source i to be observed, discretize the total time T of the observation scheduling of the target source i at time intervals of ΔT, and construct a 0, 1 sequence with a length L i = Ceil(T i / ΔT), where the nth value L i (n) = 0 indicates that the target source i is not observed in the nth time period, and L i (n) = 1 indicates that the target source i is observed in the nth time period, and when L i (n) takes the value of 1, it needs to meet the constraint conditions for the target to be observable; i S2. According to step S1, perform genetic encoding on the observation scheduling of m target sources respectively, and concatenate the obtained m sequences in sequence according to the target source number size to obtain a length

[0011] ​The 0, 1 sequence is used to form a chromosome containing m target source scheduling schemes; each chromosome is used as an individual of a scheduling scheme, and a population with an individual size of M is constructed;

[0012] S3. For each individual of the scheduling scheme in the k-th generation, calculate the total observation task completion degree according to its chromosome, and use the total observation task completion degree as its fitness function value; count and record the maximum value of the fitness function values of the M individuals in the k-th generation, and the chromosome of the individual corresponding to the maximum value;

[0013] S4. For the M individuals in the k-th generation, use the Roulette Wheel selection method, and respectively with probability Randomly select M p Individuals as the parents of the k-th generation, p i Represents the probability of the i-th individual, F j Represents the fitness function value of the j-th individual;

[0014] S5. For the parents of the k-th generation, with a crossover probability P c Randomly select two individuals from them and denote them as I a And I b , among the chromosomes of I a And I b With a length of Randomly select a corresponding crossover site and swap the values at this crossover site to obtain two genetically offspring individuals; judge whether the chromosome values of the genetically offspring individuals meet the constraints of the target observable, if not, re-select the crossover site;

[0015] S6. Repeat step S5 until the number of individuals of the genetically offspring generated reaches M;

[0016] S7. Among the M individuals generated in step S6, with a mutation probability P m Randomly select an individual, randomly select a site from its chromosome and reverse the value at this site to obtain a mutant individual; judge whether the chromosome value of the mutant individual meets the constraints of the target observable, if not, re-select the mutant site; use the M individuals including the mutant individual as the population individuals of the (k + 1)-th generation;

[0017] S8. Repeat steps S3 to S7 until the number of generations of inheritance exceeds N or the maximum value of the individual fitness function values remains unchanged for 10 consecutive generations; select the individual with the largest fitness function value from the individuals with the largest fitness function value in each generation, and decode its chromosome to obtain the optimized observation scheduling result.

[0018] Furthermore, the constraints of the target observable described in S1 include but are not limited to:

[0019] 1) The included angle between the telescope's line of sight and the sun is ≥α1, and the included angle with the moon is ≥α2;

[0020] 2) The included angle between the telescope's line of sight and the bright edge of the earth's atmosphere shall not be less than β1, and the included angle with the dark edge of the earth's atmosphere shall not be less than β2;

[0021] 3) The sub-satellite point of the telescope is not in the South Atlantic Anomaly region;

[0022] 4) The target source is not blocked by the earth.

[0023] Furthermore, the constraint conditions for target observability need to calculate the relative positions of the telescope with respect to the sun, the moon, and the earth, as well as the position of the telescope's sub-satellite point, corresponding to each locus in the chromosome within the scheduled time period [t i , t i +ΔT]. That is, starting from time t i , at 1-minute intervals, the positions of the sun, the moon, and the earth in the J2000 coordinate system at the corresponding times t i , t i +1, t i +2, …, t i +ΔT are obtained by calling the interface of the JPL405 ephemeris; the position of the telescope at the corresponding times is obtained by using the general Lagrange interpolation method through the telescope's orbit.

[0024] Furthermore, the method for determining whether the telescope's sub-satellite point is in the South Atlantic Anomaly region is as follows:

[0025] By successively calculating the offset positions of the satellite-earth geocentric coordinate system in the J2000 coordinate system at each moment within the scheduled time period and superimposing them on the position of the telescope in the J2000 coordinates, a series of sub-satellite point positions of the telescope at the corresponding times can be obtained, and the standard convex hull algorithm is used to determine whether the telescope's sub-satellite point is within the South Atlantic Anomaly region.

[0026] Furthermore, to determine the included angles between the telescope's line of sight and the bright and dark edges of the earth's atmosphere, it is first necessary to determine whether the telescope is in the sunlit area, the shaded area, or the transition area based on the included angle between the telescope and the sun. If the included angle between the vector connecting the telescope to the earth's center and the vector connecting the sun to the earth's center is less than 70 degrees, the telescope is in the sunlit area; if the included angle is greater than 110 degrees, the telescope is in the shaded area; if the included angle is greater than 70 degrees and less than 110 degrees, the telescope is in the transition area;

[0027] Calculate the included angle between the vector connecting the target source to the earth's center and the vector connecting the telescope to the earth's center. If the telescope is in the sunlit area, when this included angle is not greater than 90° - β1, it is considered to meet the bright edge included angle in constraint condition 2); if the telescope is in the shaded area, when this included angle is not greater than 90° - β2, it is considered to meet the dark edge included angle in constraint condition 2);

[0028] For the transition region, first determine whether the target source is located in the sunlit area. The determination method is as follows: If the angle between the vector connecting the target source and the center of the earth and the vector connecting the sun and the center of the earth is greater than 90 degrees, the target source is located in the sunlit area; if the angle is less than 90 degrees, the target source is located in the shaded area. For the target source located in the sunlit area, calculate the angle between the vector connecting the target source and the center of the earth and the vector connecting the telescope and the center of the earth. If the angle is not greater than 90° - β1, it is considered to meet the bright edge angle in constraint condition 2). For the target source located in the shaded area, calculate the angle between the vector connecting the target source and the center of the earth and the vector connecting the telescope and the center of the earth. If the angle is not greater than 90° - β2, it is considered to meet the dark edge angle in constraint condition 2).

[0029] Furthermore, approximate the angle between the telescope's optical axis and the sun as the angle between the vector formed by the telescope and the target source coordinates and the vector formed by the telescope and the sun, and approximate the angle between the telescope's optical axis and the moon as the angle between the vector formed by the telescope and the target source coordinates and the vector formed by the telescope and the moon. Use the cosine theorem with the position vectors of the sun and the moon relative to the earth to obtain these two angles.

[0030] Whether the target source is blocked by the earth depends on whether the angle between the vector connecting the telescope and the center of the earth and the vector connecting the target source and the center of the earth is greater than the critical angle. If it is greater than the critical angle, the target source is blocked by the earth; if it is not greater than the critical angle, the target source is not blocked by the earth. Considering the interference of the atmospheric boundary on the target source, the critical angle is set to 80°.

[0031] Furthermore, the calculation method of the total observation task completion degree described in S3 is as follows:

[0032] Assume that the number of target sources to be observed is N s , obtain the coding value of each locus in the chromosome and the total observation scheduling time of the target source corresponding to this locus, calculate the completion degree of each locus in the chromosome and accumulate it. The obtained result is the total observation task completion degree. The calculation formula for the completion degree of each locus in the chromosome is: The total observation task completion degree is x ki represents the coding value of the i-th locus in the chromosome of target source k, and T k is the total observation scheduling time of target source k, with the unit of hour.

[0033] Furthermore, the total observation scheduling time T of each target source i is the same. The time interval ΔT takes values from 5 to 15 minutes; the number of individuals M in each generation takes the value of 100, the number of genetic generations N takes the value of 500, and the number of parental individuals M in each generation p takes the value of 60; the crossover probability P c takes values in the range of 0.7 to 0.8, and the mutation probability Pm The value range of

[0034] Further, in S7, reversing the value of a certain locus on the chromosome means performing an exclusive OR operation between the original value of the chromosome locus and 1.

[0035] Further, the value ranges of both α1 and α2 are 50° - 60°, the value range of β1 is 60° - 70°, and the value range of β2 is 30° - 40°.

[0036] The beneficial effects of the present invention are:

[0037] The method proposed by the present invention realizes the optimized scheduling of the observation tasks of the space terahertz telescope based on the genetic algorithm. By quantifying the constraint conditions for observing terahertz target sources, population individuals corresponding to legal scheduling schemes are constructed. The observation completion degree of all target sources is used as the fitness function value of the individual and is quantified. The RouletteWheel selection method is used to screen the parents for offspring inheritance, effectively reducing the number of generations and the population size required for the scheduling optimization problem. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 is the main process block diagram for the optimized scheduling of the observation tasks of the space terahertz telescope based on the genetic algorithm;

[0039] Figure 2 is the schematic diagram of the 0, 1 sequence in the present invention;

[0040] Figure 3 is the flow chart of the method of the present invention;

[0041] Figure 4 is the graph showing the relationship between the fitness function of the optimal individual in each generation and the number of generations of inheritance;

[0042] Figure 5 is the simulation result of the optimized scheduling of the observation tasks in the embodiment. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0043] With the steady progress of the CSST research and development tasks, the space terahertz telescope will pose a huge demand for the optimized scheduling technology of on-orbit observation tasks, which requires solving two practical problems: (1) a general architecture for observing task scheduling to realize the observation task scheduling of multiple sky survey payloads including the terahertz module; (2) a general and effective scheduling optimization algorithm to realize the optimized scheduling of observation tasks under multiple constraint conditions.

[0044] The above two practical requirement problems will restrict the actual scheduling efficiency of the CSST observation tasks and have an adverse impact on the scientific task output of the space station sky survey. For this reason, based on the actual requirements of observation task scheduling, the present invention designs a general observation task scheduling architecture, which can realize the unified scheduling of various sky survey payload modules including the terahertz module. Secondly, the present invention establishes a general model for observation task scheduling under multiple constraints based on the genetic algorithm, and optimizes the scheduling algorithm in terms of chromosome coding and fitness function definition for solving problems.

[0045] As an artificial intelligence method, the genetic algorithm transforms the actual problem into a chromosome coding structure, designs a fitness function, sets the population size, generation number, crossover probability, and mutation probability. After obtaining the optimal population through several generations of inheritance and then restoring the actual problem, the optimized result of the actual problem can be obtained. As an important and effective operation research optimization method, the genetic algorithm is feasible in the application of the observation task scheduling of space terahertz telescopes and has considerable performance optimization space.

[0046] Now, the present invention will be further described in detail with reference to the accompanying drawings.

[0047] As shown in the Figure 1 accompanying drawings, the method of the present invention includes 7 functional parts: discretization of observation scheduling time, construction of scheduling chromosomes, calculation of fitness functions, construction of parents, crossover mutation inheritance, screening of offspring, and obtaining the finally optimized scheduling individuals.

[0048] Among them, the discretization of observation scheduling time is to realize the chromosome modeling of the scheduling problem. As Figure 2 shown, the total time T of the observation scheduling of the target source i that the space telescope needs to observe is i discretized at time intervals ΔT, and a 0, 1 sequence with length L i = Ceil(T i / ΔT) is constructed to realize the genetic coding of the scheduling problem of the target source i. Among them, L i (n) = 0 means that the target source i is not observed in the nth time period, and L i (n) = 1 means that the target source i is observed in the nth time period. When L i (n) takes the value of 1, it needs to satisfy the constraint conditions for the target to be observable. For m target sources, each is genetically encoded by the method described in step S2, and thus m sequences are obtained. The m 0, 1 sequences are concatenated in sequence according to the size of the observation target numbers to obtain a length of The 0,1 sequence is used to form a chromosome containing m target source scheduling schemes, and each chromosome encoded with m target source scheduling schemes constitutes an individual of the scheduling scheme. It is necessary to construct chromosomes containing m target source scheduling schemes with the number of individuals being M. Each locus must satisfy the observable constraint conditions of the corresponding target source within the time period, that is, the observable constraint conditions that need to be satisfied for each locus in the chromosome to take the value of 1 include but are not limited to the following aspects:

[0049] 1) The angle between the telescope's optical axis and the sun ≥ α1, and the angle between the telescope's optical axis and the moon ≥ α2;

[0050] 2) The angle between the telescope's optical axis and the bright edge of the earth's atmosphere shall not be less than β1 and the angle between the telescope's optical axis and the dark edge of the earth's atmosphere shall not be less than β2;

[0051] 3. The sub-satellite point of the telescope is not in the South Atlantic Anomaly (SAA);

[0052] 4. The target source is not blocked by the earth.

[0053] Among them, calculating and judging the above constraint conditions involves the calculation of the relative position quantities of the satellite, the earth, the sun, and the moon. Most of the relevant calculation methods are mature general algorithms. For the convenience of description, the positions of the four celestial bodies are all discussed in the J2000 coordinate system. Judging the above constraint conditions requires calculating the relative positions of the telescope with the sun, the moon, and the earth and the position of the telescope's sub-satellite point within the scheduling time period [t i ,t i +ΔT] corresponding to each locus in the chromosome (with 1 minute as the minimum interval).

[0054] Specifically, starting from the moment t i , at 1-minute intervals, a series of positions of the sun, the moon, and the earth in the J2000 coordinate system at the corresponding moments t i ,t i +1,t i +2,…,t i +ΔT are obtained by calling the interface of the JPL405 ephemeris; a series of telescope positions at the above corresponding moments are obtained by using the general Lagrangian interpolation method through the telescope's orbit; by calculating the offset positions of the earth's geocentric coordinate system in the J2000 coordinate system at the above moments in sequence and superimposing them on the coordinate positions of the telescope, a series of sub-satellite point positions of the telescope at the corresponding moments can be obtained.

[0055] The SAA region can be approximated as a convex polygon. With the help of standard convex hull algorithms, it is possible to quickly determine whether the sub-satellite point of the telescope is within the SAA region. The angle between the telescope's line of sight and the sun can be approximated as the angle between the vector formed by the telescope and the target source coordinates and the vector formed by the telescope and the sun. The angle between the telescope's line of sight and the moon can be approximated as the angle between the vector formed by the telescope and the target source coordinates and the vector formed by the telescope and the moon. Both of these angles can be obtained using the cosine theorem based on the position vectors of the sun and the moon relative to the earth.

[0056] To determine the angles between the telescope's line of sight and the bright and dark edges of the earth's atmosphere, it is first necessary to determine whether the telescope is in the sunlit area, the shaded area, or the transition area based on the angle between the telescope and the sun. The criterion is that if the angle between the vector connecting the telescope to the earth's center and the vector connecting the sun to the earth's center is less than 70 degrees, the telescope is in the sunlit area; if the angle is greater than 110 degrees, the telescope is in the shaded area; if the angle is greater than 70 degrees and less than 110 degrees, the telescope is in the transition area. Calculate the angle between the vector connecting the target source to the earth's center and the vector connecting the telescope to the earth's center. If the telescope is in the sunlit area, when this angle is not greater than 90° - β1, it is considered to satisfy the bright edge angle in Constraint 2. If the telescope is in the shaded area, when this angle is not greater than 90° - β2, it is considered to satisfy the dark edge angle in Constraint 2. For the transition area, first determine whether the target source is in the sunlit area. The method is: if the angle between the vector connecting the target source to the earth's center and the vector connecting the sun to the earth's center is greater than 90 degrees, the target source is in the sunlit area; if the angle is less than 90 degrees, the target source is in the shaded area. For a target source in the sunlit area, calculate the angle between the vector connecting the target source to the earth's center and the vector connecting the telescope to the earth's center. If the angle is not greater than 90° - β1, it is considered to satisfy the bright edge angle in Constraint 2. For a target source in the shaded area, calculate the angle between the vector connecting the target source to the earth's center and the vector connecting the telescope to the earth's center. If the angle is not greater than 90° - β2, it is considered to satisfy the dark edge angle in Constraint 2.

[0057] Whether the target is blocked by the earth depends on whether the angle between the vector connecting the telescope to the earth's center and the vector connecting the target source to the earth's center is greater than the critical angle. If it is greater than the critical angle, the target is blocked by the earth; if it is not greater than the critical angle, the target is not blocked by the earth. Considering the interference of the atmospheric boundary on the target source, the critical angle for this judgment is set to 80 degrees.

[0058] In the above judgment process, only when all the constraint conditions are satisfied at all discrete times within the scheduling period, L i (n) can be taken as 1. At this time, L i can also be taken as 0, indicating that no observation is performed within the scheduling period [t i , t i +ΔT]. Otherwise, if any of the constraint conditions cannot be satisfied at a certain moment, the target is unobservable, and Li (n) can only take 0.

[0059] For each individual, it is necessary to calculate the value of its corresponding fitness function. For the observation scheduling problem involved in the present invention, it is necessary to calculate the total observation task completion degree according to the scheduling scheme encoded in its chromosome. The specific steps are as follows: Assume that the number of target sources to be observed is N s , starting from the 1st locus of the chromosome, obtain the encoded value x of each locus one by one ki (0 or 1), query the total required observation time T k (unit: hour) of the observation target source corresponding to this locus, and obtain the completion degree of this locus of the chromosome as: By accumulating one by one, the completion degree of the scheduling scheme encoded in the chromosome can be obtained: For the M populations in each generation, there are M fitness function values. Take the maximum value and its corresponding chromosome and retain them.

[0060] If the number of genetic generations exceeds N or the optimal individual fitness function value newly calculated remains constant for 10 consecutive generations, then terminate the inheritance, and decode the optimal scheduling chromosome to obtain the optimal observation scheduling scheme.

[0061] If continuing the inheritance, the Roulette Wheel selection method needs to be adopted, that is, for the M individuals in the current generation, respectively with probability Randomly select M p individuals from the M individuals as the parents of the kth generation, and with a crossover probability P c Randomly select 2 individuals from them, denoted as I a and I b , in the chromosomes of I a and I b with a length of randomly select a corresponding crossover locus and swap the values of this crossover locus to obtain two genetic offspring individuals, and judge whether the chromosome values of the two genetic offspring are reasonable. If not, it is necessary to reselect the crossover locus until the chromosome values of the cross-inherited offspring meet the requirements. Repeat this step until the number of generated offspring individuals reaches M. Among the M reasonable individuals generated, with a mutation probability P m select an individual, randomly select a locus from its chromosome with a length of L, and invert the value of this locus (if the original value is 1, it becomes 0, if the original value is 0, it becomes 1, that is, perform an exclusive OR operation with 1) to obtain the mutated individual. Judge whether the chromosome value of the mutated individual is reasonable. If not, it is necessary to reselect the mutation locus until the chromosome value of the mutated individual is reasonable. Thus, the population individuals of the (k + 1)th generation are obtained, replacing the parents of the kth generation population, and the iteration number is incremented by one, as Figure 3As shown, repeat the above genetic iteration. Eventually, the optimal observation arrangement scheme can be obtained by decoding the optimal arrangement chromosome.

[0062] Figure 4 The relationship between the maximum fitness function value of each generation and the number of genetic generations during 500 generations of genetic iteration for 87 target sources is given. Figure 5 The timing allocation diagram of the optimal arrangement scheme obtained after 500 generations of genetic iteration for 87 target sources is given.

[0063] The above is only the preferred embodiment of the present invention, and the protection scope of the present invention is not limited to the above embodiments. All technical solutions falling within the idea of the present invention belong to the protection scope of the present invention. It should be noted that for those of ordinary skill in the art, several improvements and refinements made without departing from the principle of the present invention should be regarded as within the protection scope of the present invention.

Claims

1. A terahertz telescope observation task scheduling method based on a genetic algorithm, characterized in that It includes the following steps: S1. Genetically encode the observation schedule of the target source i to be observed, and set the total time T of the observation schedule of the target source i i Discretize it at time interval ΔT to construct a 0, 1 sequence with length L i = Ceil(T i / ΔT), where the nth value L i (n) = 0 indicates that the target source i is not observed in the nth time period, and L i (n) = 1 indicates that the target source i is observed in the nth time period. When L i (n) takes the value of 1, it needs to satisfy the constraint conditions for the target to be observable; S2. Perform genetic encoding on the observation arrangements of m target sources respectively according to step S1, and concatenate the obtained m sequences in sequence according to the number size of the target sources to obtain a 0, 1 sequence with a length to form a chromosome containing the arrangement schemes of m target sources; take each chromosome as an individual of an arrangement scheme, and construct a population with an individual scale of M; S3. For each individual of the scheduling scheme in the k-th generation, calculate the total observation task completion degree according to its chromosome, and take the total observation task completion degree as the value of its fitness function; Statistically record the maximum value of the fitness function values of the M individuals in the k-th generation, and the chromosome of the individual corresponding to the maximum value; S4. For the M individuals of the k-th generation, use the Roulette Wheel selection method, and respectively with probability randomly select M p individuals from them as the parents of the k-th generation, p i represents the probability of the i-th individual, and F j represents the fitness function value of the j-th individual; S5. For the parents of the k-th generation, with a crossover probability P c Randomly select two individuals from them and denote them as I a and I b . Among the chromosomes of I a and I b with a length of , randomly select a corresponding crossover site and swap the values at this crossover site to obtain two genetically offspring individuals; determine whether the chromosome values of the genetically offspring individuals satisfy the constraints of the target observable. If not, reselect the crossover site; S6. Repeat step S5 until the number of individuals of the generated genetic offspring reaches M; S7. Among the M individuals generated in step S6, with a mutation probability P m randomly select an individual, randomly select a locus from its chromosome and reverse the value of this locus to obtain a mutant individual; determine whether the chromosome value of the mutant individual satisfies the constraint conditions of the target observable, if not, reselect the mutation locus; use the M individuals including the mutant individual as the population individuals of the (k + 1)-th generation. S8. Repeat steps S3 to S7 until the number of genetic generations exceeds N or the maximum value of the individual fitness function values remains unchanged for 10 consecutive generations; Select the individual with the largest fitness function value from the individuals with the largest fitness function value in each generation, and decode its chromosome to obtain the optimized observation scheduling result; The target observable constraint conditions described in S1 include but are not limited to: 1) The angle between the telescope's line of sight and the sun ≥ α1, and the angle with the moon ≥ α2; 2) The angle between the telescope's line of sight and the bright edge of the Earth's atmosphere shall not be less than β1 and the angle with the dark edge of the Earth's atmosphere shall not be less than β2; 3) The sub-satellite point of the telescope is not in the South Atlantic Anomaly; 4) The target source is not blocked by the Earth; The constraint conditions for target observability need to calculate the relative positions of the telescope with respect to the sun, moon, and earth, as well as the position of the telescope's sub-satellite point, within the scheduling time period [t i , t i +ΔT] corresponding to each site in the chromosome. That is, starting from time t i , at 1-minute intervals, the positions of the sun, moon, and earth at the corresponding times t i , t i +1, t i +2, …, t i +ΔT in the J2000 coordinate system are obtained by calling the interface of the ephemeris JPL405; the position of the telescope at the corresponding times is obtained by using the general Lagrangian interpolation method through the telescope's orbit; The calculation method of the total observation task completion degree described in S3 is: Assume that the number of target sources to be observed is N s , obtain the coding value of each locus in the chromosome and the total observation scheduling time of the target source corresponding to this locus, calculate the completion degree of each locus in the chromosome and accumulate it, and the obtained result is the completion degree of the total observation task; The calculation formula for the completion degree of each locus of the chromosome is as follows: The total observed task completion degree is x ki represents the coding value of the i-th locus in the chromosome of the target source k, and T k is the total observation scheduling time of the target source k, with the unit of hour.

2. The method for scheduling observation tasks of a terahertz telescope based on a genetic algorithm according to claim 1, wherein The method for judging whether the sub-satellite point of the telescope is in the South Atlantic Anomaly is: By sequentially calculating the offset positions of the satellite's Earth geocentric coordinate system at each moment in the scheduling time period in the J2000 coordinate system, and superimposing them on the position of the telescope in the J2000 coordinate, a series of sub-satellite point positions of the telescope at the corresponding moments can be obtained, and the standard convex hull algorithm is used to judge whether the sub-satellite point of the telescope is in the South Atlantic Anomaly area.

3. The terahertz telescope observation task scheduling method based on genetic algorithm according to claim 1, wherein The angles between the telescope's line of sight and the bright edge and the dark edge of the Earth's atmosphere need to first judge whether the telescope is in the sunlit area, the shaded area, or the transition area according to the angle between the telescope and the sun. If the angle between the vector connecting the telescope and the Earth's center and the vector connecting the sun and the Earth's center is less than 70 degrees, the telescope is in the sunlit area. If the angle is greater than 110 degrees, the telescope is in the shaded area. If the angle is greater than 70 degrees and less than 110 degrees, the telescope is in the transition area; Calculate the angle between the vector connecting the target source and the Earth's center and the vector connecting the telescope and the Earth's center. If the telescope is in the sunlit area, when this angle is not greater than 90° - β1, it is considered to meet the bright edge angle in constraint condition 2); If the telescope is in the shaded area, when this angle is not greater than 90° - β2, it is considered to meet the dark edge angle in constraint condition 2); For the transition area, first judge whether the target source is in the sunlit area. The judgment method is: If the angle between the vector connecting the target source and the Earth's center and the vector connecting the sun and the Earth's center is greater than 90 degrees, the target source is in the sunlit area. If the angle is less than 90 degrees, the target source is in the shaded area; For the target source in the sunlit area, calculate the angle between the vector connecting the target source and the Earth's center and the vector connecting the telescope and the Earth's center. If the angle is not greater than 90° - β1, it is considered to meet the bright edge angle in constraint condition 2); For the target source in the shaded area, calculate the angle between the vector connecting the target source and the Earth's center and the vector connecting the telescope and the Earth's center. If the angle is not greater than 90° - β2, it is considered to meet the dark edge angle in constraint condition 2).

4. The terahertz telescope observation task scheduling method based on genetic algorithm according to claim 1, characterized in that Approximate the angle between the telescope's line of sight and the sun as the angle between the vector formed by the telescope and the target source coordinates and the vector formed by the telescope and the sun. Approximate the angle between the telescope's line of sight and the moon as the angle between the vector formed by the telescope and the target source coordinates and the vector formed by the telescope and the moon. Use the cosine theorem with the position vectors of the sun and the moon relative to the earth to obtain these two angles. Whether the target source is blocked by the earth depends on whether the angle between the vector connecting the telescope and the earth's center and the vector connecting the target source and the earth's center is greater than the critical angle. If it is greater than the critical angle, the target source is blocked by the earth; if it is not greater than the critical angle, the target source is not blocked by the earth. Considering the interference of the atmospheric boundary on the target source, the critical angle is set to 80°.

5. The terahertz telescope observation task scheduling method based on genetic algorithm according to claim 1, wherein The total time T of the observation arrangement for each target source i is the same. The time interval ΔT ranges from 5 to 15 min; the number of individuals M in each generation is 100, the number of genetic generations N is 500, and the number of parental individuals M p in each generation is 60; the crossover probability P c ranges from 0.7 to 0.8, and the mutation probability P m ranges from 0.01 to 0.

05.

6. The method for scheduling observation tasks of a terahertz telescope based on a genetic algorithm according to claim 1, wherein In S7, reversing the value at a certain locus of the chromosome means performing an exclusive OR operation between the original value of the chromosome locus and 1.

7. A terahertz telescope observation task scheduling method based on genetic algorithm according to claim 1, characterized in that The value ranges of α1 and α2 are both 50° to 60°, the value range of β1 is 60° to 70°, and the value range of β2 is 30° to 40°.

Images