An Adaptive Local Grid Nesting-based Genetic Algorithm for Multi-earth Observation Satellites’ Area Target Observation

The Scheduling of the Multi-EOSs Area Target Observation (SMEATO) is an EOS resource scheduling problem highly coupled with computational geometry. The advances in EOS technology and the expansion of wide-area remote sensing applications have increased the practical significance of SMEATO. In this paper, an adaptive local grid nesting-based genetic algorithm (ALGN-GA) is proposed for developing SMEATO solutions. First, a local grid nesting (LGN) strategy is designed to discretize the target area into parts, so as to avoid the explosive growth of calculations. A genetic algorithm (GA) framework is then used to share reserve information for the population during iterative evolution, which can generate high-quality solutions with low computational costs. On this basis, an adaptive technique is introduced to determine whether a local region requires nesting and whether the grid scale is sufficient. The effectiveness of the proposed model is assessed experimentally with nine randomly generated tests at different scales. The results show that the ALGN-GA offers advantages over several conventional algorithms in 88.9% of instances, especially in large-scale instances. These fully demonstrate the high efficiency and stability of the ALGN-GA.