A Heuristic Construction Neural Network Method for the Time-Dependent Agile Earth Observation Satellite Scheduling Problem

The agile earth observation satellite scheduling problem (AEOSSP), as a time-dependent and arduous combinatorial optimization problem, has been intensively studied in the past decades. Many studies have proposed non-iterative heuristic construction algorithms iterative meta-heuristic to solve this problem. However, spend relatively shorter time at expense of solution quality, while accomplish high-quality with lot time. To overcome shortcomings these approaches efficiently utilize historical information task characteristics, paper introduces new neural network model based on deep reinforcement learning algorithm (DRL-HA) AEOSSP proposes an innovative algorithm. DRL-HA is composed (HCNN) arrangement (TAA), where HCNN aims generate planning sequence TAA generates final feasible order tasks. In study, examined other by series experiments. results demonstrate that outperforms competitors possesses outstanding generalization ability for different scenario sizes distributions. Furthermore, HCNN, when used generating initial solutions algorithms, can achieve improved profits accelerate interactions. Therefore, verified be effective method solving AEOSSP. way, high-profit high-timeliness guaranteed, further explored improved.