A method and device for fast task scheduling under satellite network failures

Abstract

The present invention provides a method and apparatus for rapid task scheduling under satellite network failures, which relates to the technical field of satellite communication. The method is applied to a global observation satellite in a satellite computing network. The global observation satellite obtains global information in the network, and when it determines that there are faulty computing satellites, it obtains the task information of all unfinished tasks on the target faulty computing satellite; uses a genetic algorithm to generate an initial population, where each individual in the population represents a joint offloading strategy; iteratively updates the initial population according to the global information, task information, genetic algorithm, and particle swarm algorithm until a preset number of iterations is reached, so as to use the joint offloading strategy corresponding to the globally optimal individual as the task scheduling strategy for the target faulty computing satellite. The execution entity of this method is the global observation satellite that applies the genetic algorithm and the particle swarm algorithm. Compared with the reinforcement learning algorithm, it has lower memory requirements for satellites on the premise of ensuring rapid task scheduling under satellite network failures.

Description

Technical Field

[0001] The present invention relates to the technical field of satellite communication, and in particular, to a method and device for rapid task scheduling under satellite network failures. Background Art

[0002] Satellite computing networks usually consist of satellites distributed in different orbits. These satellites are equipped with computing devices and communication modules, enabling data processing and information interaction in space. However, satellite networks are at risk of node failures due to environmental conditions, operating systems, limited resources, and other factors. The impact of satellite network failures is extremely severe. Therefore, it is crucial to respond promptly to satellite network failures and ensure network robustness.

[0003] In the prior art, to address satellite node failures, intelligent algorithms such as reinforcement learning are deployed on satellite nodes. However, this deployment method requires satellite nodes to store a large amount of experience data of the interaction between the agent and the environment. As the learning process progresses, the data volume continues to increase, which is a heavy burden on satellites with tight memory resources. Summary of the Invention

[0004] The purpose of the present invention is to provide a method and device for rapid task scheduling under satellite network failures, so as to reduce the memory requirements for satellites on the premise of ensuring rapid task scheduling under satellite network failures.

[0005] In a first aspect, the present invention provides a method for rapid task scheduling under satellite network failures, which is applied to a global observation satellite in a satellite computing network and includes: obtaining global information in the satellite computing network; wherein, the satellite computing network includes: a global observation satellite, a cloud platform, multiple computing satellites, and multiple ground terminals; the global information includes: the status information of the cloud platform, the status information of the computing satellites, the status information of the ground terminals, inter-satellite link information, and satellite-ground link information; when it is determined that there are faulty computing satellites in the satellite computing network, obtaining the task information of all unfinished tasks on the target faulty computing satellite; wherein, the target faulty computing satellite represents any one of all faulty computing satellites; based on the task information of all unfinished tasks, initializing a population using a genetic algorithm to obtain an initial population; wherein, each individual in the initial population represents a joint offloading strategy, and the joint offloading strategy represents a set of task offloading strategies for all unfinished tasks; iteratively updating the initial population based on the global information, task information, genetic algorithm, and particle swarm algorithm until a preset number of iterations is reached, so as to use the joint offloading strategy corresponding to the globally optimal individual as the task scheduling strategy for the target faulty computing satellite; wherein, during the update process of each generation of the population, the particle swarm algorithm is used to optimize the initial offspring population generated based on the genetic algorithm, so as to construct the next generation of the population using the optimized offspring population and the individuals with the top descending order of fitness function values in the initial offspring population; the fitness function value of an individual is negatively correlated with the task processing delay of the joint offloading strategy corresponding to it.

[0006] Optionally, generating an initial offspring population based on the genetic algorithm includes: Step S201, selecting two parent individuals from the current population using the roulette wheel algorithm; Step S202, performing a crossover operation on the two parent individuals to obtain two crossed parent individuals; Step S203, mutating the two crossed parent individuals respectively according to the adaptive mutation probability to obtain two initial offspring individuals corresponding to the two parent individuals; repeating Steps S201 to S203 until the number of initial offspring individuals is greater than or equal to the number of individuals in the initial population, and taking the set of all initial offspring individuals as the initial offspring population.

[0007] Optionally, performing a crossover operation on the two parent individuals includes: randomly generating the starting crossover point and the ending crossover point of the gene segment to be crossed; exchanging the gene segments of the two parent individuals between the starting crossover point and the ending crossover point to obtain two crossed parent individuals; wherein, the gene segment represents a continuous plurality of gene positions, and each gene position represents a task offloading strategy for an unfinished task in the joint offloading strategy.

[0008] Optionally, mutating the two crossed parent individuals respectively according to the adaptive mutation probability to obtain two initial offspring individuals corresponding to the two parent individuals includes: using the formula Calculate the probability of mutation for the parental individuals in the current population to obtain the adaptive mutation probability; where, represents the preset initial mutation probability, represents the preset final mutation probability, represents the current iteration number of the population, represents the preset iteration number; for the target parental individual, a random number between 0 and 1 is randomly generated; where, the target parental individual represents any one of the two crossed parental individuals; if the random number is less than the adaptive mutation probability, at least one gene locus of the target parental individual is mutated to obtain the initial offspring individual corresponding to the target parental individual; if the random number is greater than or equal to the adaptive mutation probability, the target parental individual is used as its corresponding initial offspring individual.

[0009] Optionally, use the particle swarm optimization algorithm to optimize the initial offspring population generated based on the genetic algorithm, including: mapping each initial offspring individual in the initial offspring population to a particle in the particle swarm optimization algorithm to obtain the first particle swarm; updating the positions and velocities of all particles in the first particle swarm to obtain the second particle swarm; mapping each particle in the second particle swarm to an individual in the genetic algorithm to obtain the optimized offspring population.

[0010] Optionally, each task offloading strategy for the unfinished tasks includes one of the following: offloading to the cloud platform, offloading to the healthy computing satellite, offloading to the ground terminal, offloading to the cloud platform through the healthy computing satellite; mapping each initial offspring individual in the initial offspring population to a particle in the particle swarm optimization algorithm includes: based on the total number of unfinished tasks on the target faulty computing satellite and the number of optional task offloading strategies, matching a corresponding probability distribution interval for each joint offloading strategy; determining the probability distribution interval corresponding to the joint offloading strategy corresponding to the target initial offspring individual to obtain the target probability distribution interval; where, the target initial offspring individual represents any individual in the initial offspring population; taking the median value of the target probability distribution interval as the position of the particle corresponding to the target initial offspring individual; the velocity of the particle is obtained through random initialization.

[0011] Optionally, use the formula to calculate the fitness function value of the individual; where, represents the transmission delay of the i-th unfinished task in the joint offloading strategy, represents the computing delay of the i-th unfinished task in the joint offloading strategy, represents the total number of unfinished tasks on the target faulty computing satellite.

[0012] Second aspect, the present invention provides a task fast scheduling device under satellite network failures, which is applied to the global observation satellite in the satellite computing network, and includes: a first acquisition module, configured to acquire the global information in the satellite computing network; wherein, the satellite computing network includes: a global observation satellite, a cloud platform, a plurality of computing satellites and a plurality of ground terminals; the global information includes: the status information of the cloud platform, the status information of the computing satellites, the status information of the ground terminals, the inter-satellite link information and the satellite-ground link information; a second acquisition module, configured to acquire the task information of all unfinished tasks on the target failed computing satellite when it is determined that there is a failed computing satellite in the satellite computing network; wherein, the target failed computing satellite represents any one of all the failed computing satellites; an initialization module, configured to initialize a population by using a genetic algorithm based on the task information of all unfinished tasks to obtain an initial population; wherein, each individual in the initial population represents a joint offloading strategy, and the joint offloading strategy represents a set of task offloading strategies for all unfinished tasks; an update and determination module, configured to iteratively update the initial population based on the global information, the task information, the genetic algorithm and the particle swarm algorithm until a preset number of iterations is reached, so as to use the joint offloading strategy corresponding to the globally optimal individual as the task scheduling strategy for the target failed computing satellite; wherein, during the update process of each generation of population, the particle swarm algorithm is used to optimize the initial offspring population generated based on the genetic algorithm, so as to construct the next generation of population by using the optimized offspring population and the individuals with the top descending order of fitness function values in the initial offspring population; the fitness function value of an individual is negatively correlated with the task processing delay of the joint offloading strategy corresponding to it.

[0013] Third aspect, the present invention provides an electronic device, including a memory and a processor, where a computer program executable on the processor is stored on the memory, and when the processor executes the computer program, it implements the task fast scheduling method under satellite network failures according to any one of the foregoing embodiments.

[0014] Fourth aspect, the present invention provides a computer-readable storage medium, where computer instructions are stored on the computer-readable storage medium, and when the computer instructions are executed by a processor, they implement the task fast scheduling method under satellite network failures according to any one of the foregoing embodiments.

[0015] The present invention provides a method for fast task scheduling under satellite network failures. The method proposes a satellite computing network, including: a global observation satellite, a cloud platform, multiple computing satellites, and multiple ground terminals. The task fast scheduling method is applied to the global observation satellite in the satellite computing network. The global observation satellite obtains the global information in the satellite computing network. When it determines that there are faulty computing satellites in the network, it obtains the task information of all unfinished tasks on the target faulty computing satellite, and uses a genetic algorithm to generate an initial population. Each individual in the population represents a joint offloading strategy. Then, based on the global information, task information, genetic algorithm, and particle swarm algorithm, the initial population is iteratively updated until a preset number of iterations is reached, so as to use the joint offloading strategy corresponding to the global optimal individual as the task scheduling strategy for the target faulty computing satellite. The execution entity of this method is the global observation satellite, and the genetic algorithm and particle swarm algorithm are used to solve the task scheduling problem. Compared with the reinforcement learning algorithm, it has lower memory requirements for satellites on the premise of ensuring fast task scheduling under satellite network failures, and is more suitable for the situation where satellite resources are limited. Moreover, the combined application of the genetic algorithm and particle swarm algorithm can effectively improve the stability of algorithm convergence. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following will briefly introduce the drawings required for use in the description of the specific embodiments or the prior art. Obviously, the drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0017] Figure 1 It is a flowchart of a method for fast task scheduling under satellite network failures provided by an embodiment of the present invention;

[0018] Figure 2 It is an architecture diagram of satellite-ground network task scheduling in a satellite computing network provided by an embodiment of the present invention;

[0019] Figure 3 It is a system architecture diagram of a satellite network fault detection and recovery mechanism implemented in the method of an embodiment of the present invention;

[0020] Figure 4 It is a performance comparison result diagram of the improved adaptive probability genetic algorithm and the traditional genetic algorithm provided by an embodiment of the present invention;

[0021] Figure 5 It is a functional module diagram of a task fast scheduling device under satellite network failures provided by an embodiment of the present invention;

[0022] Figure 6 It is a schematic diagram of an electronic device provided by an embodiment of the present invention. Detailed implementation manners

[0023] To make the objectives, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. Generally, the components of the embodiments of the present invention described and illustrated in the accompanying drawings here can be arranged and designed in various different configurations.

[0024] Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed present invention, but merely represents selected embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0025] The following will describe in detail some implementation manners of the present invention with reference to the accompanying drawings. Without conflict, the following embodiments and the features in the embodiments can be combined with each other.

[0026] Embodiment 1

[0027] Computing offloading is a technology that transfers computing tasks from local devices to other devices or servers with stronger computing capabilities for processing. Computing offloading after satellite network failures is an effective strategy to cope with satellite network failures, aiming to ensure the continuity of services and service quality by transferring computing tasks from satellite network devices with failures to other available computing resources.

[0028] For the problem of rapid task scheduling in the case of satellite network failures, the embodiments of the present invention provide a method for rapid task scheduling in the case of satellite network failures, which is applied to a global observation satellite in a satellite computing network, such as Figure 1 shown, and the method specifically includes the following steps:

[0029] Step S102, obtain the global information in the satellite computing network.

[0030] Figure 2 For a satellite-ground network task scheduling architecture diagram in a satellite computing network provided by the embodiments of the present invention, such as Figure 2 shown, the satellite computing network includes: a global observation satellite, a cloud platform, a plurality of computing satellites and a plurality of ground terminals; the global observation satellite is used to observe the global information in the network, where the global information includes: the status information of the cloud platform, the status information of the computing satellites, the status information of the ground terminals, the inter-satellite link information and the satellite-ground link information.

[0031] In the embodiments of the present invention, the status information of the cloud platform includes: the computing resources of the cloud platform (i.e., the computing frequency); the state information of the satellite to be calculated includes: idle computing resources and the working state of the node. The state information of satellite j can be expressed as: , where represents the idle computing resources of satellite node j, represents the working state of satellite node j, = 0 indicates that satellite node j is faulty, i.e., satellite node j is a faulty computing satellite; represents that satellite node j is working properly, i.e., satellite node j is a healthy computing satellite. The state information of the ground terminal includes: the computing resources of the ground terminal . The inter-satellite link information represents the data transmission rate of the links between satellite nodes in the satellite computing network, and the satellite-ground link information represents the data transmission rate of the links between satellite nodes and the ground terminal in the satellite computing network.

[0032] Figure 3 is the system architecture diagram of the satellite network fault detection and recovery mechanism implemented in the method of the embodiment of the present invention. As Figure 3 shown, the global observation satellite includes a global observation module and a scheduling decision module. The computing satellite includes a fault prediction module and a task computing module. The global observation satellite will continuously detect the state of satellite nodes periodically, collect and update the global network information. At the same time, the computing satellite can perform fault prediction and actively send the fault information to the global observation satellite. That is, the embodiment of the present invention actually applies a hybrid mechanism of global observation and active reporting to provide guarantee for rapid satellite fault judgment and rapid task scheduling.

[0033] When it is determined that there is a faulty computing satellite, the global observation satellite will make a task scheduling decision on its unfinished tasks and send the scheduling instruction to the computing satellite. The computing satellite can unload the tasks to different locations for processing according to the scheduling instruction, so as to ensure the reliable execution of the tasks. Compared with the method of making scheduling decisions through the ground control center, implementing the method in the embodiment of the present invention by the global observation satellite can reduce the instruction transmission delay and improve the response speed. The specific implementation scheme of the scheduling decision module will be introduced in detail below.

[0034] Step S104, when it is determined that there is a faulty computing satellite in the satellite computing network, obtain the task information of all unfinished tasks on the target faulty computing satellite.

[0035] Among them, the target faulty computing satellite represents any one of all faulty computing satellites.

[0036] Specifically, if the global observation satellite determines that there is a faulty computing satellite in the network, it is necessary to further obtain the task information of all unfinished tasks on each faulty computing satellite. Among them, the i-th unfinished task on the target faulty computing satellite The task information can be expressed as: , represents the data size, represents the number of CPU cycles required for execution.

[0037] Step S106: Based on the task information of all unfinished tasks, use the genetic algorithm to initialize the population and obtain the initial population.

[0038] To solve the problem of rapid scheduling of unfinished computing tasks of faulty computing satellites in satellite network failures, an embodiment of the present invention proposes a genetic algorithm with an improved adaptive probability based on the particle swarm algorithm to jointly optimize the multi-task offloading decision. First, the global observation satellite uses the genetic algorithm to generate the initial population. Among them, each individual in the initial population represents a joint offloading strategy, and the joint offloading strategy represents the set of task offloading strategies for all unfinished tasks. That is, the individual represents the joint offloading strategy of multiple tasks. In the embodiment of the present invention, the task offloading strategies include the following four types: offloading to the cloud platform, offloading to a healthy computing satellite, offloading to a ground terminal, and offloading to the cloud platform through a healthy computing satellite.

[0039] For example, if there are M unfinished tasks on the target faulty computing satellite, and each unfinished task has 4 task offloading strategies, then the target faulty computing satellite has feasible joint offloading strategies. Therefore, the Y individuals in the initial population are randomly selected from feasible joint offloading strategies. And the purpose of the embodiment of the present invention is to explore the optimal joint offloading strategy from feasible joint offloading strategies.

[0040] Step S108: Iteratively update the initial population based on the global information, task information, genetic algorithm, and particle swarm algorithm until the preset number of iterations is reached, and use the joint offloading strategy corresponding to the global optimal individual as the task scheduling strategy for the target faulty computing satellite.

[0041] After obtaining the initial population, if only the genetic algorithm is used, then according to its global search ability in the solution space, by iteratively updating the population, the global optimal individual can be determined. However, an embodiment of the present invention improves the genetic algorithm based on the particle swarm algorithm. The particle swarm algorithm has an efficient local exploration ability, which can improve the overall performance of the algorithm, thereby further enhancing the robustness and adaptability of the method proposed in the embodiment of the present invention and accelerating the convergence speed.

[0042] Specifically, during the update process of each generation of population, the particle swarm algorithm is used to optimize the initial offspring population generated by the genetic algorithm, so as to construct the next generation of population by using the optimized offspring population and the individuals ranked at the top in the descending order of the fitness function values in the initial offspring population. That is to say, after the current population is updated by the genetic algorithm to obtain the initial offspring population, it is not directly used as the next generation of population. Instead, the particle swarm algorithm needs to be further used to optimize the offspring population to obtain the optimized offspring population. Then, the optimized offspring population and the initial offspring population are merged, and the fitness function values of each individual in the merged population are calculated. Furthermore, relatively excellent individuals are selected from the merged population according to the fitness function values to construct the next generation of population.

[0043] In the embodiment of the present invention, the fitness function value of an individual is negatively correlated with the task processing delay of the corresponding joint offloading strategy. That is, the greater the task processing delay of the joint offloading strategy, the smaller the fitness function value of the corresponding individual; on the contrary, the smaller the task processing delay of the joint offloading strategy, the greater the fitness function value of the corresponding individual. Therefore, if there are X offspring individuals in the merged population, then the fitness function values of these X offspring individuals are sorted in descending order, and the first Y offspring individuals are selected as the individuals in the next generation of population, where Y represents the number of individuals in the initial population.

[0044] The embodiment of the present invention provides a task fast scheduling method under satellite network failures. The method proposes a satellite computing network, including: a global observation satellite, a cloud platform, multiple computing satellites and multiple ground terminals. And the task fast scheduling method is applied to the global observation satellite in the satellite computing network. The global observation satellite obtains the global information in the satellite computing network, and when it is determined that there is a faulty computing satellite in the network, it obtains the task information of all unfinished tasks on the target faulty computing satellite, and generates an initial population by using the genetic algorithm. Each individual in the population represents a joint offloading strategy. Then, the initial population is iteratively updated according to the global information, task information, genetic algorithm and particle swarm algorithm until a preset number of iterations is reached, so as to use the joint offloading strategy corresponding to the global optimal individual as the task scheduling strategy for the target faulty computing satellite. The execution subject of this method is the global observation satellite, and the genetic algorithm and the particle swarm algorithm are used to solve the task scheduling problem. Compared with the reinforcement learning algorithm, it has lower memory requirements for satellites on the premise of ensuring fast task scheduling under satellite network failures, and is more suitable for the situation where satellite resources are limited. Moreover, the combined application of the genetic algorithm and the particle swarm algorithm can effectively improve the stability of algorithm convergence.

[0045] In an optional implementation manner, generating the initial offspring population based on the genetic algorithm specifically includes the following steps:

[0046] Step S201, select two parent individuals from the current population using the roulette wheel algorithm.

[0047] Specifically, first calculate the total fitness of the current population , and set the fitness function value of the y-th parent individual in the current population to be represented as . Then the total fitness of the current population is: , where Y represents the population size, that is, the total number of individuals in the population.

[0048] Next, generate a random number pick within the range of , and represents the cumulative fitness of the first k parent individuals in the current population. If the random number pick satisfies the following formula: , then select the k-th parent individual. Referring to the above method, generate two random numbers and determine the cumulative fitness ranges they are in to select two parent individuals from the current population.

[0049] Step S202, perform a crossover operation on the two parent individuals to obtain two crossover parent individuals.

[0050] To increase the diversity of the population and promote the global search ability, after selecting two parent individuals, perform a crossover operation on the above two parent individuals to obtain two crossover parent individuals. The embodiments of the present invention do not specifically limit the method of the crossover operation, and users can select according to actual needs. For example, single-point crossover, multi-point crossover, uniform crossover, etc.

[0051] Step S203, perform mutation on the two crossover parent individuals respectively according to the adaptive mutation probability to obtain two initial offspring individuals corresponding to the two parent individuals.

[0052] To improve the species diversity and further enhance the global search ability of the algorithm, after obtaining the two crossover parent individuals, the embodiments of the present invention perform mutation on the two crossover parent individuals respectively based on the adaptive mutation probability to obtain two initial offspring individuals. Among them, the adaptive mutation probability changes with the population iteration times. The more the population iteration times, the smaller the adaptive mutation probability, thereby accelerating the convergence speed of the algorithm.

[0053] Repeat steps S201 to S203 until the number of initial offspring individuals is greater than or equal to the number of individuals in the initial population, and use the set of all initial offspring individuals as the initial offspring population.

[0054] Performing one round of steps S201 to S203 can generate two initial offspring individuals. To obtain all initial offspring individuals, the above steps need to be repeated. If the number Y of individuals in the initial population is even, then the number of repetitions is Y / 2; if the number Y of individuals in the initial population is odd, then the number of repetitions is (Y + 1) / 2. Correspondingly, the number of individuals in all offspring populations can be maintained by randomly deleting an initial offspring individual, or by other means to maintain the population size, or the population size can be changed to Y + 1. This change in population size does not affect the performance of the algorithm.

[0055] In an optional implementation manner, for the above step S202, the crossover operation on two parent individuals specifically includes the following steps:

[0056] Step S2021, randomly generate the starting crossover point and the ending crossover point of the gene segment to be crossed.

[0057] Step S2022, exchange the gene segments of the two parent individuals between the starting crossover point and the ending crossover point to obtain two crossed parent individuals; wherein, the gene segment represents a continuous plurality of gene positions, and each gene position represents a task offloading strategy for an unfinished task in the joint offloading strategy.

[0058] Based on the above description, it can be seen that the embodiment of the present invention adopts a multi-point crossover method to perform a crossover operation on two parent individuals. By randomly selecting two crossover points and segmentally exchanging the gene segments of the parent individuals, it can better maintain gene diversity and avoid premature convergence to a local optimum.

[0059] For ease of understanding, the following is an example. If there are a total of M gene positions in the parent individual, and the randomly generated starting crossover point and ending crossover point are 3 and 7 respectively, then performing multi-point crossover on the two parent individuals is equivalent to exchanging the continuous plurality of gene positions: 3, 4, 5, 6, 7 corresponding gene segments of the two parent individuals.

[0060] In an optional implementation manner, for the above step S203, mutate the two crossed parent individuals respectively according to the adaptive mutation probability to obtain two initial offspring individuals corresponding to the two parent individuals, specifically including the following steps:

[0061] Step S2031, use the formula Calculate the mutation probability of the parent individuals in the current population to obtain the adaptive mutation probability; wherein, represents the preset initial mutation probability, represents the preset final mutation probability, represents the current iteration number of the population, represents the preset iteration number.

[0062] Step S2032: Generate a random number between 0 and 1 for the target parent individual, where the target parent individual represents any one of the two crossed parent individuals.

[0063] Step S2033: If the random number is less than the adaptive mutation probability, mutate at least one gene locus of the target parent individual to obtain the initial offspring individual corresponding to the target parent individual.

[0064] Step S2034: If the random number is greater than or equal to the adaptive mutation probability, use the target parent individual as its corresponding initial offspring individual.

[0065] Specifically, the embodiment of the present invention adopts an adaptive mutation probability function. From the expression of the adaptive mutation probability, it can be seen that the mutation probability can be dynamically adjusted according to the current iteration number of the population. This adaptive mutation probability mechanism can mutate individuals with a relatively high probability in the initial stage of the algorithm to improve species diversity, and reduce the individual mutation probability in the later stage, thereby accelerating the convergence speed.

[0066] In an optional embodiment, the particle swarm algorithm is used to optimize the initial offspring population generated based on the genetic algorithm, which specifically includes the following steps:

[0067] Step S301: Map each initial offspring individual in the initial offspring population to a particle in the particle swarm algorithm to obtain the first particle swarm.

[0068] Step S302: Update the positions and velocities of all particles in the first particle swarm to obtain the second particle swarm.

[0069] Step S303: Map each particle in the second particle swarm to an individual in the genetic algorithm to obtain the optimized offspring population.

[0070] Specifically, the embodiment of the present invention introduces the particle swarm algorithm to conduct a fine exploration of the initial offspring population. Through the movement and information sharing of particles in the solution space, it can quickly converge to the vicinity of the local optimal solution. Specifically, for each initial offspring population, to conduct a local exploration in the way of a particle swarm, first, each initial offspring individual in the population needs to be mapped to a particle in the particle swarm algorithm. The embodiment of the present invention does not specifically limit the mapping method, and users can choose according to actual needs.

[0071] After obtaining the first particle swarm, use the velocity update formula and position update formula of particles in the existing particle swarm algorithm to update the positions and velocities of each particle to obtain the second particle swarm. Finally, use the method "opposite" to mapping the initial offspring population to the first particle swarm to map all particles in the second particle swarm to the genetic algorithm to obtain the optimized offspring population.

[0072] In an alternative implementation, the task offloading strategy for each unfinished task includes one of the following: offloading to the cloud platform, offloading to the health computing satellite, offloading to the ground terminal, offloading to the cloud platform through the health computing satellite; the above step S301 of mapping each initial offspring individual in the initial offspring population to a particle in the particle swarm algorithm specifically includes the following steps:

[0073] Step S3011: Based on the total number of unfinished tasks on the target fault computing satellite and the optional number of task offloading strategies, match a corresponding probability distribution interval for each joint offloading strategy.

[0074] Step S3012: Determine the probability distribution interval corresponding to the joint offloading strategy of the target initial offspring individual to obtain the target probability distribution interval; where the target initial offspring individual represents any individual in the initial offspring population.

[0075] Step S3013: Take the median value of the target probability distribution interval as the position of the particle corresponding to the target initial offspring individual; the velocity of the particle is obtained through random initialization.

[0076] For ease of understanding, referring to the example in the above text, assume that there are a total of M unfinished tasks on the target fault computing satellite, and each unfinished task has 4 task offloading strategies, then the target fault computing satellite has feasible joint offloading strategies. Based on this, it can be determined that probability distribution intervals: , ,… . Then match and bind each joint offloading strategy with each probability distribution interval.

[0077] If the probability distribution interval corresponding to the joint offloading strategy of the target initial offspring individual is , then the median value is the position of the particle corresponding to the target initial offspring individual. And the velocity of the particle can be represented as a vector, and the components of the vector respectively represent the adjustment direction and amplitude of the corresponding task offloading strategy.

[0078] Correspondingly, after obtaining the second particle swarm, according to the position values of the particles in the second particle swarm, the corresponding probability distribution intervals can be matched, and then according to the matching relationship between the joint offloading strategy and the probability distribution interval determined above, it can be mapped into the genetic algorithm to obtain the corresponding optimized offspring individuals.

[0079] In an alternative implementation, use the formula to calculate the fitness function value of the individual; where Denote the transmission delay of the $i$-th unfinished task in the joint offloading strategy. Denote the computing delay of the $i$-th unfinished task in the joint offloading strategy. Denote the total number of unfinished tasks on the target faulty computing satellite.

[0080] In the embodiments of the present invention, the optimization objective of the joint offloading strategy is to minimize the processing delay (transmission delay + computing delay) of unfinished tasks. Therefore, the fitness function value is negatively correlated with the task processing delay of the joint offloading strategy. Given the complex satellite network state, the processing times of tasks may vary greatly. In the embodiments of the present invention, in order to suppress the influence of large values on the fitness function value, logarithmic transformation processing is adopted for the processing delays of all unfinished tasks. Through logarithmic transformation, the values can be compressed, making the fitness function values more balanced, so as to more fairly evaluate various task allocation schemes.

[0081] Next, the transmission delay and computing delay of each task offloading strategy for the $i$-th unfinished task are specifically introduced.

[0082] 1. Offload to a healthy computing satellite: , , where Denote the data transmission rate of the inter-satellite link between satellite $j$ and satellite . Denote the computing frequency allocated to task $i$ by satellite .

[0083] 2. Offload to a ground terminal: , , where Denote the data transmission rate of the satellite-ground link between satellite $j$ and ground terminal $i$. Denote the computing frequency allocated to task $i$ by the ground terminal.

[0084] 3. Offload to a cloud platform: , , where Denote the data transmission rate of the link between satellite $j$ and the cloud platform. Denote the computing frequency allocated to task $i$ by the cloud platform.

[0085] 4. Offload to the cloud platform through a healthy computing satellite: , , where Denote the data transmission rate of the link between satellite and the cloud platform. That is, this strategy considers two transmission delays, namely the transmission delay from satellite $j$ to satellite and the transmission delay from satellite to the cloud platform.

[0086] In summary, the embodiment of the present invention discloses a task fast scheduling mechanism for satellite network faults, providing an effective solution for the reliable processing of tasks in the satellite node fault scenario. In the present invention, the global observation satellite is used to obtain the state information of each satellite node and the multi-dimensional information of the network. The satellite will also actively send the detected fault information, which helps the global observation satellite to timely grasp the fault situation and avoid scheduling delay caused by untimely information acquisition. This hybrid mechanism of global observation and active reporting can speed up the fault response speed, reduce the impact of faults on task execution, and provide guarantee for the fast scheduling of network tasks. A high-speed and reliable communication link is established between the global observation satellite and each satellite node, which can transmit network fault information and offloading instructions to the nodes in the network. Based on rich data experience, the global satellite can quickly make task scheduling decisions, reasonably allocate network resources, and ensure that tasks are quickly and effectively transmitted to other nodes for execution.

[0087] The embodiment of the present invention adopts an improved adaptive genetic algorithm, which can effectively jointly optimize multi-task offloading decisions. The genetic algorithm can perform global search in the search space. It starts searching from a population, which enables it to explore multiple regions simultaneously, increasing the chance of finding the global optimal solution and also speeding up the search speed. It is very suitable for the fast switching scenario of satellite fault network tasks. The improved adaptive genetic algorithm introduces a multi-point crossover mechanism, which helps to maintain the diversity of the population. Through the adaptive mutation probability mechanism, a higher mutation probability is maintained in the initial stage of algorithm exploration to avoid prematurely reaching the local optimum, and the mutation probability is reduced in the later stage of exploration to speed up the algorithm convergence speed, thus effectively improving the performance of the algorithm. At the same time, a particle swarm optimization improvement mechanism is introduced to increase the local exploration ability.

[0088] Figure 4 It is a graph showing the performance comparison results of the improved adaptive probability genetic algorithm and the traditional genetic algorithm in the embodiment of the present invention. By Figure 4 It can be seen that the algorithm adopted by the present invention has more stable convergence and higher fitness compared with the traditional genetic algorithm, and thus has better performance.

[0089] Embodiment 2

[0090] The embodiment of the present invention also provides a task fast scheduling device for satellite network faults. This device is applied to the global observation satellite in the satellite computing network and is mainly used to execute the task fast scheduling method for satellite network faults provided in the above Embodiment 1. The following is a specific introduction to the task fast scheduling device for satellite network faults provided in the embodiment of the present invention.

[0091] Figure 5 It is a functional module diagram of a task fast scheduling device for satellite network faults provided in the embodiment of the present invention. AsFigure 5 As shown in Figure 5 , the device mainly includes: a first acquisition module 10, a second acquisition module 20, an initialization module 30, and an update and determination module 40, where:

[0092] The first acquisition module 10 is configured to acquire global information in the satellite computing network; wherein, the satellite computing network includes: global observation satellites, a cloud platform, multiple computing satellites, and multiple ground terminals; the global information includes: the status information of the cloud platform, the status information of the computing satellites, the status information of the ground terminals, inter-satellite link information, and satellite-ground link information.

[0093] The second acquisition module 20 is configured to acquire the task information of all unfinished tasks on the target failed computing satellite when it is determined that there is a failed computing satellite in the satellite computing network; wherein, the target failed computing satellite represents any one of all the failed computing satellites.

[0094] The initialization module 30 is configured to initialize a population using a genetic algorithm based on the task information of all unfinished tasks to obtain an initial population; wherein, each individual in the initial population represents a joint offloading strategy, and the joint offloading strategy represents a set of task offloading strategies for all unfinished tasks.

[0095] The update and determination module 40 is configured to iteratively update the initial population based on the global information, task information, genetic algorithm, and particle swarm algorithm until a preset number of iterations is reached, so as to use the joint offloading strategy corresponding to the globally optimal individual as the task scheduling strategy for the target failed computing satellite; wherein, during the update process of each generation of the population, the particle swarm algorithm is used to optimize the initial offspring population generated based on the genetic algorithm, so as to construct the next generation of the population using the optimized offspring population and the individuals with the top descending order of fitness function values in the initial offspring population; the fitness function value of an individual is negatively correlated with the task processing delay of the corresponding joint offloading strategy.

[0096] An embodiment of the present invention provides a task fast scheduling device under satellite network failures. The device proposes a satellite computing network, including: a global observation satellite, a cloud platform, multiple computing satellites, and multiple ground terminals. And the device is applied to the global observation satellite in the satellite computing network. The global observation satellite obtains the global information in the satellite computing network. When it is determined that there are faulty computing satellites in the network, it obtains the task information of all unfinished tasks on the target faulty computing satellite, and generates an initial population using a genetic algorithm. Each individual in the population represents a joint offloading strategy. Then, based on the global information, task information, genetic algorithm, and particle swarm algorithm, the initial population is iteratively updated until a preset number of iterations is reached, so as to use the joint offloading strategy corresponding to the global optimal individual as the task scheduling strategy for the target faulty computing satellite. The execution entity of the device is the global observation satellite, and the genetic algorithm and particle swarm algorithm are used to solve the task scheduling problem. Compared with the reinforcement learning algorithm, it has lower memory requirements for satellites on the premise of ensuring fast task scheduling under satellite network failures, and is more suitable for the situation where satellite resources are limited. Moreover, the combined application of the genetic algorithm and the particle swarm algorithm can effectively improve the stability of algorithm convergence.

[0097] Optionally, the device further includes: a generation module, configured to generate an initial offspring population based on a genetic algorithm. The generation module includes:

[0098] A selection unit, configured to select two parent individuals from the current population using a roulette wheel algorithm.

[0099] A crossover unit, configured to perform a crossover operation on the two parent individuals to obtain two crossover parent individuals.

[0100] A mutation unit, configured to perform mutations on the two crossover parent individuals respectively according to an adaptive mutation probability to obtain two initial offspring individuals corresponding to the two parent individuals.

[0101] A repeated call unit, configured to repeatedly call the selection unit, the crossover unit, and the mutation unit until the number of initial offspring individuals is greater than or equal to the number of individuals in the initial population, and use the set of all initial offspring individuals as the initial offspring population.

[0102] Optionally, the crossover unit is specifically configured to:

[0103] Randomly generate a start crossover point and an end crossover point of the gene segment to be crossed.

[0104] Exchange the gene segments of the two parent individuals between the start crossover point and the end crossover point to obtain two crossover parent individuals; where the gene segment represents a continuous plurality of gene positions, and each gene position represents a task offloading strategy for an unfinished task in the joint offloading strategy.

[0105] Optionally, the mutation unit is specifically configured to:

[0106] Use the arithmetic formula To calculate the mutation probability of the parental individuals in the current population, and obtain the adaptive mutation probability; where Represents the preset initial mutation probability, Represents the preset final mutation probability, Represents the current iteration number of the population, Represents the preset iteration number.

[0107] For the target parental individual, a random number between 0 and 1 is randomly generated; where the target parental individual represents any one of the two crossed parental individuals.

[0108] If the random number is less than the adaptive mutation probability, at least one gene locus of the target parental individual is mutated to obtain the initial offspring individual corresponding to the target parental individual.

[0109] If the random number is greater than or equal to the adaptive mutation probability, the target parental individual is used as its corresponding initial offspring individual.

[0110] Optionally, the device further includes: an optimization module, configured to optimize the initial offspring population generated based on the genetic algorithm by using the particle swarm algorithm. The optimization module includes:

[0111] The first mapping unit is configured to map each initial offspring individual in the initial offspring population to a particle in the particle swarm algorithm to obtain the first particle swarm.

[0112] The update unit is configured to update the positions and velocities of all particles in the first particle swarm to obtain the second particle swarm.

[0113] The second mapping unit is configured to map each particle in the second particle swarm to an individual in the genetic algorithm to obtain the optimized offspring population.

[0114] Optionally, the task offloading strategy for each unfinished task includes one of the following: offloading to the cloud platform, offloading to the healthy computing satellite, offloading to the ground terminal, and offloading to the cloud platform through the healthy computing satellite; the first mapping unit is specifically configured to:

[0115] Based on the total number of unfinished tasks on the target failure computing satellite and the optional number of task offloading strategies, match a corresponding probability distribution interval for each joint offloading strategy.

[0116] Determine the probability distribution interval corresponding to the joint offloading strategy corresponding to the target initial offspring individual to obtain the target probability distribution interval; where the target initial offspring individual represents any individual in the initial offspring population.

[0117] Take the median of the target probability distribution interval as the position of the particle corresponding to the target initial offspring individual; the velocity of the particle is obtained by random initialization.

[0118] Optionally, use the formula to calculate the fitness function value of the individual; where represents the transmission delay of the i-th unfinished task in the joint offloading strategy, represents the computing delay of the i-th unfinished task in the joint offloading strategy, represents the total number of unfinished tasks on the target fault calculation satellite.

[0119] Embodiment III

[0120] Refer to Figure 6 In this embodiment of the present invention, an electronic device is provided. The electronic device includes: a processor 60, a memory 61, a bus 62, and a communication interface 63. The processor 60, the communication interface 63, and the memory 61 are connected through the bus 62; the processor 60 is configured to execute an executable module stored in the memory 61, such as a computer program.

[0121] Among them, the memory 61 may include a high-speed random access memory (RAM), and may also include a non-volatile memory, such as at least one disk memory. Through at least one communication interface 63 (which may be wired or wireless), a communication connection between this system network element and at least one other network element can be realized, and the Internet, wide area network, local area network, metropolitan area network, etc. can be used.

[0122] The bus 62 may be an ISA bus, a PCI bus, an EISA bus, etc. The bus can be divided into an address bus, a data bus, a control bus, etc. For the sake of convenience of representation, Figure 6 only a bidirectional arrow is used in

[0123] but it does not mean that there is only one bus or one type of bus.

[0124] The processor 60 may be an integrated circuit chip with the ability to process signals. In the implementation process, each step of the above method can be completed by the integrated logic circuit of the hardware in the processor 60 or the instructions in the form of software. The above-mentioned processor 60 may be a general-purpose processor, including a central processing unit (CPU for short), a network processor (NP for short), etc.; it may also be a digital signal processor (DSP for short), an application specific integrated circuit (ASIC for short), a field-programmable gate array (FPGA for short), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components. It can implement or execute the various methods, steps and logic block diagrams disclosed in the embodiments of the present invention. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc. The steps of the method disclosed in combination with the embodiments of the present invention can be directly embodied as being executed and completed by the hardware decoding processor, or executed and completed by the combination of the hardware and software modules in the decoding processor. The software module may be located in a mature storage medium in the art such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory, or an electrically erasable programmable memory, a register, etc. This storage medium is located in the memory 61, and the processor 60 reads the information in the memory 61 and combines its hardware to complete the steps of the above method.

[0125] A computer program product of a method and device for fast task scheduling under satellite network failures provided by an embodiment of the present invention includes a computer-readable storage medium storing non-volatile program code executable by a processor. The instructions included in the program code can be used to execute the method described in the foregoing method embodiments. For specific implementation, reference can be made to the method embodiments and will not be elaborated here.

[0126] In addition, in each embodiment of the present invention, each functional unit may be integrated in a processing unit, or each unit may exist physically alone, or two or more units may be integrated in one unit.

[0127] When the above-mentioned functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a non-volatile computer-readable storage medium executable by a processor. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to enable a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memories (ROM, Read-Only Memory), random access memories (RAM, Random Access Memory), magnetic disks, or optical discs that can store program codes.

[0128] It should be noted that: similar reference numerals and letters represent similar items in the following drawings. Therefore, once an item is defined in one drawing, it does not need to be further defined and explained in subsequent drawings.

[0129] In the description of the present invention, it should be noted that the orientation or positional relationship indicated by the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer", etc. is based on the orientation or positional relationship shown in the drawings, or the orientation or positional relationship in which the product of this invention is usually placed during use. It is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be construed as a limitation to the present invention. In addition, the terms "first", "second", "third", etc. are only used for descriptive distinction and cannot be understood as indicating or implying relative importance.

[0130] In addition, the terms "horizontal", "vertical", "hanging", etc. do not mean that the components are required to be absolutely horizontal or hanging, but can be slightly inclined. For example, "horizontal" only means that its direction is more horizontal relative to "vertical", and does not mean that this structure must be completely horizontal, but can be slightly inclined.

[0131] In the description of the present invention, it should also be noted that unless otherwise clearly specified and limited, the terms "set", "install", "connect", "couple" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be directly connected or indirectly connected through an intermediate medium, and it can be the communication inside two elements. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific situations.

[0132] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements on some or all of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for rapid task scheduling under satellite network failure, characterized in that: Global observation satellites used in satellite computing networks include: Acquire global information in the satellite computing network; wherein the satellite computing network includes: a global observation satellite, a cloud platform, multiple computing satellites and multiple ground terminals; the global information includes: status information of the cloud platform, status information of the computing satellite, status information of the ground terminal, inter-satellite link information and satellite-to-ground link information; When it is determined that there is a faulty computing satellite in the satellite computing network, task information of all unfinished tasks on the target faulty computing satellite is obtained; wherein the target faulty computing satellite represents any satellite among all the faulty computing satellites; Based on the task information of all the unfinished tasks, a population is initialized using a genetic algorithm to obtain an initial population; wherein each individual in the initial population represents a joint unloading strategy, and the joint unloading strategy represents a set of task unloading strategies of all the unfinished tasks; The initial population is iteratively updated based on the global information, the task information, the genetic algorithm and the particle swarm algorithm until a preset number of iterations is reached, so that the joint unloading strategy corresponding to the global optimal individual is used as the task scheduling strategy of the target fault calculation satellite; wherein, in each generation of population update process, the particle swarm algorithm is used to optimize the initial offspring population generated based on the genetic algorithm, so as to construct the next generation population using the optimized offspring population and individuals with the highest fitness function values ​​in descending order in the initial offspring population; the fitness function value of the individual is negatively correlated with the task processing delay of the corresponding joint unloading strategy; Among them, using the formula Calculate the fitness function value of the individual; wherein, represents the transmission delay of the i-th unfinished task in the joint offloading strategy, represents the computation delay of the i-th unfinished task in the joint offloading strategy, Indicates the total number of uncompleted tasks on the target fault computing satellite.

2. The method for rapid task scheduling under satellite network failure according to claim 1, characterized in that: Generate the initial offspring population based on the genetic algorithm, including: Step S201, using a roulette algorithm to select two parent individuals from the current population; Step S202, performing a crossover operation on the two parent individuals to obtain two crossover parent individuals; Step S203, mutating the two parent individuals after the crossover according to the adaptive mutation probability, to obtain two initial offspring individuals corresponding to the two parent individuals; Repeat steps S201 to S203 until the number of the initial offspring individuals is greater than or equal to the number of individuals in the initial population, and take the set of all the initial offspring individuals as the initial offspring population.

3. The method for rapid task scheduling under satellite network failure according to claim 2, characterized in that: Performing a crossover operation on the two parent individuals includes: Randomly generate the starting intersection point and the ending intersection point of the gene fragment to be crossed; The gene fragments of the two parent individuals located between the starting intersection and the ending intersection are exchanged to obtain two parent individuals after crossing; wherein the gene fragments represent a plurality of continuous gene loci, and each gene locus represents a task unloading strategy for an unfinished task in the joint unloading strategy.

4. The method for rapid task scheduling under satellite network failure according to claim 2, characterized in that: Mutating the two parent individuals after the crossover according to the adaptive mutation probability to obtain two initial offspring individuals corresponding to the two parent individuals, including: Using the formula The probability of mutation of the parent individuals in the current population is calculated to obtain the adaptive mutation probability; wherein, represents the preset initial mutation probability, represents the preset final mutation probability, represents the current iteration number of the population, represents the preset number of iterations; For the target parent individual, a random number between 0 and 1 is randomly generated; wherein the target parent individual represents any one of the two parent individuals after the crossover; If the random number is less than the adaptive mutation probability, mutating at least one gene locus of the target parent individual to obtain an initial offspring individual corresponding to the target parent individual; If the random number is greater than or equal to the adaptive mutation probability, the target parent individual is used as its corresponding initial child individual.

5. The method for rapid task scheduling under satellite network failure according to claim 1, characterized in that: The particle swarm algorithm is used to optimize the initial offspring population generated by the genetic algorithm, including: Mapping each initial offspring individual in the initial offspring population to a particle in the particle swarm algorithm to obtain a first particle swarm; Updating the positions and velocities of all particles in the first particle group to obtain a second particle group; Each particle in the second particle group is mapped to an individual in the genetic algorithm to obtain the optimized offspring population.

6. The method for rapid task scheduling under satellite network failure according to claim 5, characterized in that: The task offloading strategy of each of the unfinished tasks includes one of the following: offloading to the cloud platform, offloading to a healthy computing satellite, offloading to a ground terminal, and offloading to the cloud platform via a healthy computing satellite; Mapping each initial offspring individual in the initial offspring population to a particle in the particle swarm algorithm comprises: Calculate the total number of unfinished tasks on the satellite and the optional number of task offloading strategies based on the target fault, and match a corresponding probability distribution interval for each joint offloading strategy; Determine the probability distribution interval corresponding to the joint unloading strategy corresponding to the target initial offspring individual, and obtain the target probability distribution interval; wherein the target initial offspring individual represents any individual in the initial offspring population; The median of the target probability distribution interval is used as the position of the particle corresponding to the target initial offspring individual; the speed of the particle is obtained by random initialization.

7. A task rapid scheduling device under satellite network failure, characterized in that: Global observation satellites used in satellite computing networks include: A first acquisition module is used to acquire global information in the satellite computing network; wherein the satellite computing network includes: a global observation satellite, a cloud platform, multiple computing satellites and multiple ground terminals; the global information includes: status information of the cloud platform, status information of the computing satellite, status information of the ground terminal, inter-satellite link information and satellite-to-ground link information; A second acquisition module is used to acquire task information of all unfinished tasks on a target faulty computing satellite when it is determined that there is a faulty computing satellite in the satellite computing network; wherein the target faulty computing satellite represents any satellite among all the faulty computing satellites; An initialization module, configured to initialize a population using a genetic algorithm based on the task information of all the unfinished tasks to obtain an initial population; wherein each individual in the initial population represents a joint unloading strategy, and the joint unloading strategy represents a collection of task unloading strategies of all the unfinished tasks; An updating and determining module is used to iteratively update the initial population based on the global information, the task information, the genetic algorithm and the particle swarm algorithm until a preset number of iterations is reached, so as to use the joint unloading strategy corresponding to the global optimal individual as the task scheduling strategy of the target fault calculation satellite; wherein, in each generation of population update process, the particle swarm algorithm is used to optimize the initial offspring population generated based on the genetic algorithm, so as to construct the next generation population using the optimized offspring population and the individuals with the highest fitness function values ​​in descending order in the initial offspring population; the fitness function value of the individual is negatively correlated with the task processing delay of the joint unloading strategy corresponding to it; Among them, using the formula Calculate the fitness function value of the individual; wherein, represents the transmission delay of the i-th unfinished task in the joint offloading strategy, represents the computation delay of the i-th unfinished task in the joint offloading strategy, Indicates the total number of uncompleted tasks on the target fault computing satellite.

8. An electronic device comprising a memory and a processor, wherein the memory stores a computer program that can be run on the processor, wherein: When the processor executes the computer program, the method for rapid task scheduling under satellite network failure described in any one of claims 1 to 6 is implemented.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and when the computer instructions are executed by a processor, the method for rapid task scheduling under a satellite network failure according to any one of claims 1 to 6 is implemented.

Images