Space-based resource intelligent task planning method based on intensity pareto evolutionary algorithm

Abstract

The application discloses a space-based resource intelligent task planning method based on an intensity Pareto evolution algorithm, and comprises the following steps: S1, modeling a space-based resource task planning problem as a multi-objective vehicle path problem model with a time window; S2, establishing a multi-objective optimization model based on the multi-objective vehicle path problem model with the time window; S3, generating an initial population corresponding to the multi-objective optimization model by using a heuristic rule; S4, iteratively evolving and solving the initial population to obtain a final non-dominated solution set; and S5, selecting a final task planning scheme from the final non-dominated solution set, wherein the scheme explicitly indicates space-based resources corresponding to each observation task, a service path of each resource and service time of each task. The application can combine multi-objective collaborative optimization, adopt a hybrid search strategy of global and local collaboration, and efficiently generate a set of Pareto optimal task planning schemes considering cost, resources and benefits under the condition of meeting complex constraint conditions.

Description

Technical Field

[0001] This invention relates to the field of aerospace technology, and more specifically to a space-based resource intelligent mission planning method based on the strength Pareto evolutionary algorithm. Background Technology

[0002] Space-based observation missions are a crucial means of acquiring space information. Their core lies in utilizing space-based platforms such as satellites and detectors to conduct periodic, continuous, or emergency observations of surface or space targets. With the expansion of constellation size and the increasing diversification of mission requirements, the scheduling and planning of space-based observation resources has become a key aspect of space system operation and management. Currently, there are numerous Earth observation satellites in orbit, encompassing various payload types including optical, radar, multispectral, and hyperspectral, with mission requirements covering a wide range of important fields such as meteorological monitoring, disaster early warning, resource surveys, and ocean observation. Scientific and rational mission planning can significantly improve the utilization efficiency of observation resources and mission completion rates, effectively reducing unnecessary attitude maneuvers and energy consumption, thereby helping to extend the service life of spacecraft. This is of vital importance for improving the overall operational efficiency of the constellation and supporting the construction of the national space information system.

[0003] In the context of complex multi-satellite and multi-mission applications, space-based observation mission planning exhibits typical characteristics of multiple objectives, multiple constraints, and strong dynamic coupling. The planning process requires comprehensive consideration of various complex factors, including satellite attitude maneuverability constraints, observation time window visibility, onboard energy and storage resource load, data transmission and communication timing, imaging mission priorities, and switching time between different imaging modes, to ensure mission continuity and system stability. Therefore, researching an intelligent mission planning method capable of achieving optimal resource allocation and maximizing observation efficiency under multi-satellite and multi-mission conditions has significant scientific value and engineering application prospects for improving the overall operational efficiency and emergency mission response capabilities of space-based systems.

[0004] Currently, existing task planning methods can be mainly categorized into the following two types:

[0005] The first category is rule-based scheduling methods. For example, in prior art [1] (He C, Dong Y, Li H, LiewY. Reasoning-Based Scheduling Method for Agile Earth Observation Satellite with Multi-Subsystem Coupling, Remote Sensing, 2023, 15(6):1577.), a reasoning-based agile Earth observation satellite mission scheduling method was proposed. This method achieves mission allocation through a predefined rule base and real-time mission state reasoning. The method has a clear structure and is easy to implement in engineering. However, the performance of such methods is highly dependent on manually pre-set rules and mission priorities. In complex, multi-objective conflict and dynamically changing mission scenarios, it is often difficult to balance the global optimality of the scheme with the adaptability to real-time disturbances.

[0006] The second category is planning methods based on optimization algorithms. For example, in the prior art [2] (Zheng et al., "Amodified genetic algorithm for large-scale and joint satellite mission planning," Egyptian Informatics Journal, 2025.), an improved genetic algorithm for large-scale joint satellite mission planning was proposed. It unifies the modeling of task allocation and observation path optimization, and obtains a better planning scheme through global evolutionary search of the population. Such methods can explore diverse feasible schemes in a large solution space, but they still generally suffer from problems such as high computational cost, slow convergence speed, and significant influence of the initial population on the quality of the solution. Especially when dealing with multi-objective optimization and complex constraint scenarios, the algorithm is prone to getting trapped in local optima. In addition, most traditional optimization algorithms are usually designed only for a single optimization objective (such as minimizing energy consumption or maximizing task coverage), ignoring the continuity requirements, revisit requirements, and robustness of the scheme during execution in actual tasks.

[0007] In real-world space mission execution environments, mission planning not only needs to achieve efficient scheduling under limited resources but also needs to effectively address various uncertainties such as cloud cover, communication link delays, and onboard equipment malfunctions. This requires planning methods to have the ability to dynamically balance mission priorities, system resource load, and mission execution risks under multiple objectives and constraints, thereby generating a multi-objective optimal mission planning scheme that simultaneously considers high feasibility and strong robustness. The shortcomings of existing technologies in these aspects are precisely the core problems that this invention aims to solve. Summary of the Invention

[0008] To overcome the shortcomings of existing technologies, the present invention aims to provide a space-based resource intelligent mission planning method based on the strength Pareto evolutionary algorithm. This method combines multi-objective collaborative optimization and adopts a hybrid search strategy of global and local collaboration to efficiently generate a set of Pareto optimal mission planning schemes that take into account cost, resources, and benefits under complex constraints.

[0009] To achieve the objective of this invention, the following solution is adopted:

[0010] A space-based resource intelligent mission planning method based on the strength Pareto evolutionary algorithm includes the following steps:

[0011] S1. Model the space-based resource mission planning problem as a multi-objective vehicle routing problem model with a time window, and define the set of virtual customer nodes, the set of space-based resources, the planning period, and the attribute parameters of each node and resource;

[0012] S2. Based on the multi-objective vehicle routing problem model with time window, establish a multi-objective optimization model with the optimization objectives of minimizing the total resource switching cost, minimizing the amount of resources used, and maximizing the total benefit of task service, and including resource constraints, path continuity constraints, and time window constraints.

[0013] S3. Set the population size, external file size, maximum number of iterations, and objective function weight vector for the Pareto evolutionary algorithm, and use heuristic rules to generate an initial population corresponding to the multi-objective optimization model.

[0014] S4. Based on the multi-objective optimization model, the initial population is iteratively evolved to solve the problem, including: assigning fitness values, selecting environments and updating files, mating selection, combining mutation and recombination with local search operators, updating the population and repairing paths, until the termination condition is met, and the final non-dominated solution set is obtained.

[0015] S5. Select the final mission planning scheme from the final non-dominated solution set. The scheme specifies the space-based resources corresponding to each observation mission, the service path of each resource, and the service time of each mission.

[0016] Furthermore, in step S1, the multi-objective vehicle routing problem model with time window abstracts space-based observation resources into virtual vehicles and observation tasks into virtual client nodes; the attributes of virtual client nodes include the earliest service start time, the latest service start time, the shortest service time, and service capacity requirements; the attributes of virtual vehicles include the maximum service capacity; and the attributes of the edges between nodes include service switching cost and time.

[0017] Further, in step S2, the optimization objective of the multi-objective optimization model is specifically:

[0018] The expression for minimizing the total switching cost of resources is: ;

[0019] The expression for minimizing resource usage is: ;

[0020] The expression for maximizing the total benefit of task services is: .

[0021] Furthermore, in step S3, the heuristic rule is: when generating the initial solution, the space-based resources should prioritize serving the virtual client node closest to its current location.

[0022] Further, in step S4, the fitness assignment includes: calculating the strength value of each individual based on the objective function value of the multi-objective optimization model. and density estimation And according to the formula Calculate the final fitness, where, For the final fitness, For individuals The quantity related to being dominated by other individuals.

[0023] Further, in step S4, the environment selection and file update include: copying non-dominated individuals from the current population and external files to a temporary file; if the temporary file exceeds the preset capacity, a truncation method based on the k-th nearest neighbor distance is used to remove some solutions; if the capacity is not reached, the best-fit dominated individual is added, and finally the updated temporary file replaces the original file.

[0024] Further, in step S4, the mutation and recombination operation combining local search operators specifically includes: integrating multiple local search operators, including a 2-opt operator for optimizing the access order within a single path, a node swapping operator for exchanging client nodes between different paths, a node insertion operator for inserting a client node from one path into another, a path intersection operator for exchanging two path segments, and a path splitting and reconnecting operator for splitting and reconnecting paths.

[0025] Further, in step S4, the path repair includes: based on the constraints of the multi-objective optimization model, identifying and removing customer nodes that violate the constraints in the current path, connecting the remaining nodes to form a partially feasible path; attempting to re-insert the removed nodes into feasible positions in the existing path according to the time window order; if there are still nodes that cannot be inserted, creating a new path and inserting them in sequence, finally obtaining a fully feasible task planning scheme.

[0026] Furthermore, in step S4, the termination condition is that the number of iterations reaches the preset maximum number of iterations.

[0027] Furthermore, in step S5, the final task planning scheme includes: which specific space-based resource serves each virtual client node, the service path and access order of each resource, the service start and end time of each node, the total service cost, the total number of resources used, and the cumulative service time.

[0028] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0029] 1. This invention models the complex space-based resource mission planning problem as a multi-objective vehicle routing problem with a time window, and constructs a multi-objective optimization model that simultaneously minimizes the total resource switching cost, minimizes the amount of resources used, and maximizes the total service benefit of the mission. This model is then solved using an improved intensive Pareto evolutionary algorithm, effectively coordinating the optimization of multiple conflicting objectives such as service cost, resource load, and observation benefits. This method outputs a set of Pareto non-dominated solutions in a single step, providing decision-makers with multiple planning schemes that balance different preferences (such as cost priority, resource conservation, or benefit maximization), overcoming the limitation of narrow decision space in traditional single-objective optimization methods.

[0030] 2. This invention innovatively integrates various local search operators, including 2-opt, node swapping, and node insertion, into the global search framework of the Pareto evolutionary algorithm to mutate and recombine solutions, and designs a dedicated path repair mechanism to handle solutions that violate constraints. This strategy of combining global exploration with refined local search significantly enhances the algorithm's optimization ability and search efficiency under complex constraints (such as time windows and resource capabilities), enabling rapid generation and repair of high-quality feasible solutions. It effectively avoids the problems of traditional optimization algorithms easily getting trapped in local optima and slow convergence, thereby greatly improving the feasibility and overall quality of the final task planning scheme.

[0031] 3. This invention unifies multiple objectives through objective function weight vectors and uses distance-based heuristics to generate an initial population, effectively balancing the competition between objective functions of different physical dimensions and orders of magnitude, while providing high-quality starting search points for the algorithm. This measure significantly accelerates the algorithm's convergence process and improves overall solution efficiency, making this invention particularly suitable for practical space-based observation mission planning scenarios with large star clusters, numerous observation tasks, and complex constraints, exhibiting good engineering applicability and computational robustness. Attached Figure Description

[0032] Figure 1 This is a flowchart of the space-based resource intelligent mission planning method based on the intensity Pareto evolutionary algorithm in an embodiment of the present invention;

[0033] Figure 2 This is a schematic diagram of the domain search operator in an embodiment of the present invention;

[0034] Figure 3 This is a schematic diagram of the logical flow of the space-based resource intelligent mission planning method based on the strength Pareto evolutionary algorithm in an embodiment of the present invention;

[0035] Figure 4 This is a schematic diagram of the fitness allocation scheme for a multi-objective optimization problem in an embodiment of the present invention;

[0036] Figure 5 This is a first curve showing the estimation of the detector attitude maneuvering time in an embodiment of the present invention;

[0037] Figure 6 This is a second curve showing the estimation of the detector attitude maneuver time in an embodiment of the present invention;

[0038] Figure 7 This is a schematic diagram of timeline expansion in an embodiment of the present invention;

[0039] Figure 8 This is a schematic diagram of the task planning results under simple conditions in an embodiment of the present invention;

[0040] Figure 9 This is a resource scheduling Gantt chart for the optimization results of the c101 dataset in this embodiment of the invention. Detailed Implementation

[0041] The present invention will now be further described in conjunction with the accompanying drawings and specific embodiments. It should be noted that, without conflict, the various embodiments or technical features described below can be arbitrarily combined to form new embodiments.

[0042] This invention provides a space-based resource intelligent mission planning method based on the strength Pareto evolutionary algorithm. This method can simultaneously optimize mission allocation, resource utilization, and observation coverage under various constraints such as time windows, attitude maneuvers, energy load, and mission priorities. The space-based resource intelligent mission planning method of this invention models the space-based mission planning problem as a multi-objective optimization model with time windows. Combining the strength Pareto evolutionary algorithm with diverse local search mechanisms, it achieves multi-objective coordination and feasibility assurance of mission schemes, and features rich solution sets, good convergence, and strong robustness.

[0043] like Figures 1 to 9As shown in the figure, the space-based resource intelligent mission planning method based on the strength Pareto evolutionary algorithm of this invention is applicable to the autonomous planning and resource optimization allocation of missions such as multi-satellite formations, Earth observation constellations, and space exploration platforms, and includes the following steps:

[0044] S1. Model the space-based resource mission planning problem as a multi-objective vehicle routing problem with a time window, and define the set of virtual customer nodes, the set of space-based resources, the planning period, and the attribute parameters of each node and resource.

[0045] In this embodiment, problem modeling and parameter initialization model the space-based resource mission planning problem as a multi-objective vehicle routing problem with a time window. Based on the user's spatial information requirements, the set of virtual client nodes that need to be served is determined. The system's schedulable set of space-based resources (virtual vehicles) And the planning period. Define the basic parameters of the problem model:

[0046] An undirected graph G=(N,E) is given, where E is the set of edges connecting virtual client nodes.

[0047] Attributes of virtual client node i: including earliest service start time Latest service start time The shortest service time required And the virtual vehicle service capabilities required to serve this customer. .

[0048] Attributes of virtual vehicle k: including maximum customer service capacity .

[0049] Connecting edges Attributes: including service switching costs Service transition time In this invention, = .

[0050] Total system resource constraints: Maximum available number of virtual vehicles .

[0051] S2. Based on the multi-objective vehicle routing problem model with time window, establish a multi-objective optimization model with the optimization objectives of minimizing the total resource switching cost, minimizing the amount of resources used, and maximizing the total benefit of task service, and including resource constraints, path continuity constraints, and time window constraints.

[0052] In this embodiment, considering the simultaneous existence of multiple satellite resources and multiple observation missions, the space-based mission planning problem is modeled as a multi-objective optimization problem with a time window. Joint optimization is performed using the following three integrated objectives:

[0053] 1. Minimize the total cost of resource switching, i.e., the cost of attitude or orbital maneuvers between tasks:

[0054]

[0055] 2. Minimize resource usage to improve the overall task concentration of the system:

[0056]

[0057] 3. Maximize the overall service benefits of the mission to improve observation coverage and mission gains:

[0058]

[0059] The constraints of this problem include:

[0060] Satellite resource constraints:

[0061]

[0062] Path continuity constraints:

[0063]

[0064] Satellite return constraints:

[0065]

[0066] Customer single service constraints:

[0067]

[0068] Satellite capability constraints:

[0069]

[0070] Service time window constraints:

[0071]

[0072]

[0073]

[0074]

[0075] in, This represents the waiting time at virtual client node i;

[0076] This represents the service start time at virtual client node i;

[0077] This represents the end time of the service at virtual client node i.

[0078] S3. Set the population size, external file size, maximum number of iterations, and objective function weight vector for the Pareto evolutionary algorithm, and use heuristic rules to generate an initial population corresponding to the multi-objective optimization model.

[0079] Set the parameters for the strength Pareto evolutionary algorithm, including population size N, external archive (non-dominated solution set) size, and maximum number of iterations. and the objective function weight vector Used for standardization of the objective function. Generates the initial population. When generating the initial solution, a heuristic rule is used, which tends to direct the satellite to the virtual client node closest to its current location.

[0080] S4. Based on the multi-objective optimization model, iteratively evolve the initial population to solve the problem, including: assigning fitness values, selecting environments and updating files, mating selection, combining mutation and recombination with local search operators, updating the population and repairing paths, until the termination condition is met, and obtaining the final non-dominated solution set.

[0081] In this embodiment, the initial population undergoes iterative evolution and data updates, as detailed below:

[0082] 1. Fitness assignment: First, assign fitness values ​​to the current population. and external archives Fitness was calculated for all individuals.

[0083] Calculate the intensity value S(i) for each individual i:

[0084]

[0085] Calculate the density estimate D(i) for individual i:

[0086]

[0087] The final fitness F(i) of individual i is:

[0088]

[0089] The running time of the fitness assignment process is determined by the density estimator ( ) decided, and and The computational complexity is ,in .

[0090] 2. Environmental Selection and Archive Update: First, compare the current population with all non-dominant individuals in the external archive, i.e., their fitness values. Each individual is copied to a temporary file. Then, the size of this temporary file is determined and adjusted: if its size exceeds the preset file capacity limit, a method based on the first... The nearest neighbor distance truncation method iteratively removes solutions with the smallest distance to neighboring individuals until the file size meets the requirements. If the size does not reach the upper limit, individuals with the best fitness are selected from the remaining dominated individuals in the current population and the file to fill the file. Finally, this updated temporary file replaces the original external file, completing this round of file update.

[0091] 3. Termination condition judgment: If If other convergence conditions are met, then the iteration stops and... Output the final non-dominated solution set; otherwise, proceed to the next step.

[0092] 4. Mating selection: A binary tournament selection method is used, selecting from the merged populations. and archives Individuals are selected to enter the mating pool.

[0093] 5. Mutation and Recombination: Crossover and mutation operations are applied to individuals in the mating pool to generate a progeny population. To significantly improve the quality of the solution, this method integrates a variety of local search operators during the mutation and recombination process. These operators aim to fine-tune the structure of the solution and mainly include: the 2-opt operator to optimize the customer access order within a single path by swapping two edges in the path; the node swapping operator to swap two customer nodes between different paths; the node insertion operator to remove a customer node from the original path and insert it into another path; the path crossover operator to swap a segment of two paths; and the path splitting and reconnecting operator to split a complete path and reconnect its segments to other paths. The synergistic effect of these operators enables the algorithm to effectively explore different regions of the solution space.

[0094] 6. Population Update and Local Search: The generated offspring population is set as the new current population. Subsequently, a local search is selectively performed on solutions within this new population, primarily employing the computationally efficient 2-opt algorithm for local path optimization. Simultaneously, to ensure the new solution satisfies all problem constraints, a path repair algorithm is designed. This algorithm first identifies and removes all customer nodes in the current path that violate constraints (such as time windows and service capacity), connecting the remaining compliant nodes to form partially feasible path segments. Next, according to the order of the customer node time windows, it attempts to re-insert these removed nodes one by one into feasible positions in all existing paths. If, after the above attempts, there are still nodes that cannot be accommodated by any existing path, a new path is created, and the remaining nodes are inserted sequentially into the new path according to their time windows, thus ultimately obtaining a fully feasible task planning scheme.

[0095] S5. Select the final mission planning scheme from the final non-dominated solution set. The scheme specifies the space-based resources corresponding to each observation mission, the service path of each resource, and the service time of each mission.

[0096] In this embodiment, the final task planning scheme is selected from the final non-dominated solution set. Specifically, after the algorithm terminates, the external file is output. Multiple Pareto non-dominated programming schemes are stored in the database. A final scheme is selected from these schemes based on actual mission preferences. The scheme explicitly specifies: which satellite (space-based resource) serves each virtual client node (observation task), the service path of each satellite (i.e., the order in which client nodes are accessed), the specific service start and end times for each client node, the total service cost of the entire scheme, the total number of resources used, and the cumulative service time.

[0097] This invention presents an intelligent mission planning method for space-based resources based on the Intensity Pareto Evolutionary Algorithm (SPEA). For arbitrary multi-satellite, multi-mission observation scenarios, a mission-resource mapping model is first established based on observation requirements and resource parameters. Observation satellites are abstracted as virtual resource units, and observation requirements are transformed into virtual mission nodes with time window constraints. Then, a multi-objective optimization model is constructed with the objectives of minimizing mission service costs, minimizing resource usage, and maximizing cumulative service time. During the solution process, a global search framework based on the Intensity Pareto Evolutionary Algorithm (SPEA) is introduced. Non-dominated solutions are saved through an external archive mechanism, and local search operators are used to refine and restore the feasibility of the solutions. Through multiple iterations of evolution and local optimization, a high-quality mission planning scheme that satisfies multiple constraints is obtained, achieving intelligent allocation of space-based resources and multi-objective collaborative optimization. This method has advantages such as diverse planning results, strong robustness, and high computational efficiency.

[0098] The following uses c101 from the internationally recognized standard example set of vehicle routing problems as a simulation scenario to verify the effectiveness of the space-based resource intelligent mission planning method based on the intensity Pareto evolutionary algorithm in this embodiment of the invention. This example simulates the mission planning problem of observing 100 ground targets under the condition of multiple Earth observation satellites. Its implementation process includes the following steps:

[0099] Step 1: First, based on the C101 example data, the space-based resource mission planning problem is modeled as a multi-objective vehicle routing problem with a time window. This model abstracts space-based observation resources (such as satellites) as virtual vehicles and the observation tasks to be executed as virtual client nodes. The edges between client nodes represent transitions between tasks, and their costs... With time Typically set to Euclidean distance. Each client node has a specific service time window. Service needs Each virtual vehicle has a fixed service capacity. In this embodiment, the maximum available virtual vehicle capacity is set to 25 km, and the single-vehicle service capacity is... =200.

[0100] Step 2: Based on the above model, this embodiment constructs an optimization problem with three objectives:

[0101] 1. Minimize the total cost of resource switching, i.e., the cost of attitude or orbital maneuvers between tasks:

[0102]

[0103] 2. Minimize resource usage to improve the overall task concentration of the system:

[0104]

[0105] 3. Maximize the overall service benefits of the mission to improve observation coverage and mission gains:

[0106]

[0107] At the same time, the solution must meet all the constraints given in step two, including path continuity, satellite return, customer single service, satellite capacity, and strict time window.

[0108] Step 3: Set the core parameters of the Pareto evolutionary algorithm: population size =50, External file size =100, maximum number of iterations =200. The objective function is standardized using a weight vector λ=(0.4,0.4,0.2) to balance the dimensions and importance of different objectives. Subsequently, an initial population P0 is generated. During the generation process, heuristic rules are used to prioritize serving targets closer to the satellite, thereby quickly obtaining a high-quality initial solution.

[0109] Step 4: Iterative evolution and file updates, the overall process is as follows: Figure 3 As shown, the specific process is as follows:

[0110] 1. Fitness assignment: Calculate the current population fitness value according to formulas (13)-(15). and archives The fitness of all individuals is calculated. This process comprehensively considers each individual's Pareto dominance S(i) and its density estimate in the target space. Thus, the final fitness is obtained. This effectively balances convergence and diversity.

[0111] 2. Environmental selection and record updating: This involves all non-dominated individuals ( Copy to a temporary file. (e.g., ...) Figure 4 As shown, when the number of non-dominated solutions exceeds the file capacity, a truncation method based on the k-th nearest neighbor distance is used to iteratively remove the solution with the smallest distance to its neighbors, in order to retain a uniformly distributed set of non-dominated solutions and update the result. .

[0112] 3. Termination Check: If the number of iterations is... If the condition is met, the loop terminates and proceeds to step five; otherwise, execution continues.

[0113] 4. Mating selection: A binary tournament selection method is used, selecting from the merged populations. and archives Individuals are selected to enter the mating pool.

[0114] 5. Mutation and Recombination with Local Search: Crossover and mutation operations are applied to individuals in the mating pool. To significantly improve the quality of the solution, this method integrates multiple local search operators, whose operation methods are as follows: Figure 2 As shown, the operators include 2-opt, node swapping, node insertion, path intersection, and path splitting and reconnection. These operators can finely adjust the path structure and task allocation. Among them, 2-opt is the core operator, and its execution logic flow is as follows: Figure 2 As shown in (a), it optimizes the internal access order by swapping two edges in the path, thereby effectively reducing the total cost of a single path.

[0115] 6. Population Update and Feasibility Repair: Set the generated offspring population as the new current population. .right The new solutions are selectively subjected to local search (primarily using 2-opt). For solutions that may violate constraints, a path repair algorithm is initiated: the violating node is removed, and the remaining nodes are connected; then, the violating node is re-inserted into all feasible positions in the order of the time window; if there are still remaining, a new path is created and inserted in order, ultimately ensuring that all solutions satisfy the constraints.

[0116] Step 5: After the algorithm terminates, output the final file. The Pareto non-dominated solution set in [the context of the problem]. Decision-makers can choose the final solution based on practical preferences (such as prioritizing cost reduction or reducing resource usage). For example... Figure 9 As shown, this embodiment presents an optimal solution obtained on the C101 example. The Gantt chart clearly illustrates the time scheduling of the nine satellites (vertical axis), with each square representing an observation of a target, and the labels indicating the mission sequence. This solution successfully served all 100 targets with a total service cost of 3424.6.

[0117] The above is a detailed description of the preferred embodiments of the present invention. However, the present invention is not limited to the embodiments described. Those skilled in the art can make various equivalent modifications or substitutions without departing from the spirit of the present invention. All such equivalent modifications or substitutions are included within the scope defined by the claims of this application.

Claims

1. A space-based resource intelligent mission planning method based on the intensity Pareto evolutionary algorithm, characterized in that, Includes the following steps: S1. Model the space-based resource mission planning problem as a multi-objective vehicle routing problem model with a time window, and define the set of virtual customer nodes, the set of space-based resources, the planning period, and the attribute parameters of each node and resource; S2. Based on the multi-objective vehicle routing problem model with time window, establish a multi-objective optimization model with the optimization objectives of minimizing the total resource switching cost, minimizing the amount of resources used, and maximizing the total benefit of task service, and including resource constraints, path continuity constraints, and time window constraints. S3. Set the population size, external file size, maximum number of iterations, and objective function weight vector for the Pareto evolutionary algorithm, and use heuristic rules to generate an initial population corresponding to the multi-objective optimization model. S4. Based on the multi-objective optimization model, the initial population is iteratively evolved to solve the problem, including: assigning fitness values, selecting environments and updating files, mating selection, combining mutation and recombination with local search operators, updating the population and repairing paths, until the termination condition is met, and the final non-dominated solution set is obtained. The fitness assignment includes: calculating the strength value of each individual based on the objective function value of the multi-objective optimization model. and density estimation And according to the formula Calculate the final fitness, where, For the final fitness, For individuals The quantity related to being dominated by other individuals; The mutation and recombination operation combining local search operators specifically includes: integrating multiple local search operators, including a 2-opt operator for optimizing the access order within a single path, a node swapping operator for exchanging client nodes between different paths, a node insertion operator for inserting a client node from one path into another path, a path intersection operator for exchanging two path segments, and a path splitting and reconnecting operator for splitting and reconnecting paths. S5. Select the final mission planning scheme from the final non-dominated solution set. The scheme specifies the space-based resources corresponding to each observation mission, the service path of each resource, and the service time of each mission.

2. The space-based resource intelligent mission planning method based on the intensity Pareto evolutionary algorithm according to claim 1, characterized in that, In step S1, the multi-objective vehicle routing problem model with time window abstracts space-based observation resources into virtual vehicles and observation tasks into virtual client nodes. The attributes of virtual client nodes include the earliest service start time, the latest service start time, the shortest service time, and service capacity requirements. The attributes of virtual vehicles include the maximum service capacity. The attributes of the edges between nodes include service switching cost and time.

3. The space-based resource intelligent mission planning method based on the intensity Pareto evolutionary algorithm according to claim 1, characterized in that, In step S3, the heuristic rule is: when generating the initial solution, the space-based resources should prioritize serving the virtual client node that is closest to its current location.

4. The space-based resource intelligent mission planning method based on the intensity Pareto evolutionary algorithm according to claim 1, characterized in that, In step S4, the environment selection and file update include: copying non-dominated individuals from the current population and external files to a temporary file; if the temporary file exceeds the preset capacity, removing some solutions using a truncation method based on the k-th nearest neighbor distance; if the capacity is not reached, supplementing with the best-fit dominated individual, and finally replacing the original file with the updated temporary file.

5. The space-based resource intelligent mission planning method based on the intensity Pareto evolutionary algorithm according to claim 1, characterized in that, In step S4, the path repair includes: based on the constraints of the multi-objective optimization model, identifying and removing customer nodes that violate the constraints in the current path, connecting the remaining nodes to form a partially feasible path; attempting to re-insert the removed nodes into feasible positions in the existing path according to the time window order; if there are still nodes that cannot be inserted, creating a new path and inserting them in sequence, finally obtaining a fully feasible task planning scheme.

6. The space-based resource intelligent mission planning method based on the intensity Pareto evolutionary algorithm according to claim 1, characterized in that, In step S4, the termination condition is that the number of iterations reaches the preset maximum number of iterations.

7. The space-based resource intelligent mission planning method based on the intensity Pareto evolutionary algorithm according to claim 1, characterized in that, In step S5, the final task planning scheme includes: which specific space-based resource serves each virtual client node, the service path and access order of each resource, the service start and end time of each node, the total service cost, the total number of resources used, and the cumulative service time.

Images