Satellite Scheduling Using Hybrid Genetic Algorithm and Simulated Annealing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for solving the Satellite Scheduling Problem (SSP) face challenges in efficiently scheduling satellite resources to optimize acquisition tasks, often resulting in suboptimal solutions due to high computational complexity and a tendency to get stuck in local optima, especially with tightly constrained resources.
Innovation Solution
A hybrid approach combining Genetic Algorithms (GA) and Simulated Annealing (SA) is employed to optimize satellite scheduling, where GA initializes conflict-free solutions and SA performs local optimization to expand the solution space and improve task scheduling within given constraints.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional scheduling methods are used to solve the Satellite Scheduling Problem, then the scheduling process is simpler to implement, but the solution quality deteriorates due to getting stuck in local optima and high computational complexity
Solution Approach 1:
The patent combines Genetic Algorithm (GA) and Simulated Annealing (SA) into a hybrid optimization framework. GA provides global search capability through population-based evolution, while SA adds local search refinement and helps escape local optima through probabilistic acceptance of worse solutions. This merging resolves the contradiction by achieving high solution quality through complementary algorithmic strengths without requiring an overly complex single algorithm.
Solution Approach 2:
The scheduling problem is segmented into multiple independent acquisition tasks that can be optimized separately and then combined. The hybrid algorithm processes tasks in discrete iterations, evaluating and optimizing subsets of tasks while maintaining feasibility constraints. This segmentation allows the complex problem to be solved through manageable steps, improving solution quality without overwhelming computational burden.
2Productivity
If more satellite resources are allocated to increase task scheduling, then the productivity improves, but the computational time increases due to tighter constraints
Solution Approach 1:
The algorithm uses partial action by optimizing a subset of acquisition tasks in each iteration rather than attempting to optimize all tasks simultaneously. The GA population evolves through generations with incremental improvements, and SA performs localized refinements on promising solutions. This approach achieves high productivity by making progressive improvements without requiring exhaustive computation on the entire problem space at once.
Solution Approach 2:
The system performs preliminary filtering and preprocessing of acquisition tasks before applying the full hybrid optimization. Tasks are pre-evaluated for feasibility, compatibility, and priority, allowing the optimization algorithm to focus computational resources on the most promising task combinations. This preliminary action reduces the effective problem size, enabling higher productivity with reduced computational time.
3Reliability
If conventional optimization methods are used, then the implementation is straightforward, but the ability to avoid local optima deteriorates
Solution Approach 1:
The patent merges the global exploration capability of Genetic Algorithm with the local exploitation and escape mechanism of Simulated Annealing. GA's population-based approach maintains diversity and explores the solution space broadly, while SA's probabilistic acceptance criterion allows escape from local optima by occasionally accepting worse solutions. This combination achieves superior global optimization capability by leveraging the strengths of both methods while managing complexity through their synergistic interaction.
Data Source
Figure 1~2
Figure 3
Figure 4
AI summary
The invention relates to a satellite scheduling method (1) including: a) producing initial scheduling plans (105) on the basis of input requests related to tasks to be performed within a given time period by one or more remote sensing satellites; wherein in each of said initial scheduling plans respective tasks are scheduled, which do not conflict with each other in time and in using satellite resources of the remote sensing satellite(s); and wherein each of the tasks to be performed is scheduled in at least one of the initial scheduling plans; b) applying a genetic-algorithm-based processing (108) to the initial scheduling plans to produce a genetic-algorithm-based scheduling plan which is optimized with respect to given mission objectives, and complies with given constraints related to the satellite resources, to the tasks to be performed, and to the given time period; and c) applying a simulated-annealing-based processing (109) to the genetic-algorithm-based scheduling plan to produce a simulated-annealing-based scheduling plan that fits the given mission objectives, that complies with the given constraints, and in which a larger number of tasks is scheduled than in the genetic-algorithm-based scheduling plan. In particular, the step b) includes carrying out a genetic-algorithm-based iterative procedure comprising: at a first genetic-algorithm-based iteration, selecting a subset of the initial scheduling plans on the basis of the given mission objectives (202,203), and applying crossover (205), mutation (206) and elitism (207) techniques based on respective predefined genetic evolution factors to the selected sub-set of the initial scheduling plans to produce evolved scheduling plans complying with the given constraints; at each genetic-algorithm-based iteration following the first one, selecting, on the basis of the given mission objectives, a subset of the evolved scheduling plans produced at the preceding genetic-algorithm-based iteration (202,203), and applying the crossover (205), mutation (206) and elitism (207) techniques to the selected sub-set of the evolved scheduling plans produced at the preceding genetic-algorithm-based iteration to produce new evolved scheduling plans complying with the given constraints. Moreover, said step b) further includes: stopping carrying out the genetic-algorithm-based iterative procedure when given genetic-algorithm-related stopping criteria are met (208,209); and automatically selecting, among the evolved scheduling plans produced at the last genetic-algorithm-based iteration performed, the one which best fits the given mission objectives. Additionally, the satellite scheduling method (1) includes also: computing an intersection matrix (103) representing conflicts in time and in using the satellite resources of the tasks to be performed within the given time period; computing a plan complexity (103) on the basis of the intersection matrix; and computing the given genetic-algorithm-related stopping criteria on the basis of the intersection matrix (104,104a); wherein the initial scheduling plans are produced on the basis of said intersection matrix (105).