Solving Satellite Range Scheduling Problem with Learning-Based Artificial Bee Colony Algorithm

Yanjie Song, Luona Wei, Lining Xing, Yi Fang, Zhongshan Zhang, Yingwu Chen

Abstract: Satellite range scheduling problem (SRSP) is a critical and challenging scheduling problem due to the oversubscribed and sequence dependency characteristics. The artificial bee colony algorithm (ABC) is one of the popular evolutionary algorithms to solve large-scale scheduling problems. A new artificial bee colony algorithm is proposed, named the learning-based artificial bee colony algorithm (LB-ABC). We proposed two new learning strategies, named error-based learning strategy and position-based learning strategy, to improve traditional ABC's exploration and exploitation performance. Error-based learning strategy uses differences of each individual in the population to improve the population structure. Position-based learning strategy through the experience of scheduling tasks guide the generation process of new bees. Experiments and algorithm analysis show that the learning strategy we proposed is of great help to planning tasks, and the performance of the new algorithm is better than state-of-the-art algorithms.