Improvement Of Biogeography-based Optimization Algorithm And Its Applied Research In Satellite Task Scheduling

Biogeography-Based Optimization algorithm(BBO)is an optimization algorithm constructed by studying the natural activities of biological organisms in space and time,and it belongs to a new type of swarm intelligence optimization algorithm.The BBO algorithm realizes information exchange within the population through the unique migration operators,and increases the diversity of the population through the mutation operators.The combination of the two operators makes the algorithm have the advantages of fast optimization and strong development ability,and have obvious advantages in solving complex combinatorial optimization problems.With the rapid development of China's aerospace industry,and user's requirements for satellite observation targets is also becoming more and more complicated.In order to make full use of satellite resources and effectively meet the needs of users,it is of great significance to explore the task scheduling of multiple satellites.The problem of satellite task scheduling is to work out a conflict-free task sequences to meet the needs of users,which needs fully consider the resources of multiple satellite and multiple ground stations while satisfying multiple constraints.The problem of satellite task scheduling belongs to a combinatorial optimization problem with characteristics of multi-objective and multi-constraint,so it is of certain application value to design an appropriate planning model and an optimization algorithm with better performance.In view of the optimization characteristics of BBO algorithm,this paper improves the algorithm and conducts research on satellite task scheduling based on this algorithm.The specific work is as follows:1.A Biogeography-Based Optimization algorithm based on perturbation-differential migration and mixed mutation(PDMBBO)was proposed.Since the traditional BBO algorithm cannot meet the excellent development ability while avoiding falling into local optimal solution,a differential mutation strategy and an adaptive perturbation factor are introduced to improve the search accuracy of migration operator,and the hybrid mutation operator is designed so that the algorithm has good global exploration ability in the early stage of iteration and local development ability in the later stage of iteration.The simulation experiments show that PDMBBO has obvious orders of magnitude advantages in three aspects of convergence speed,optimization accuracy and robustness whether compared with other intelligent optimization algorithms or the improved BBO algorithms in recent years.PDMBBO not only has excellent development ability,but also avoid the algorithm falling into the local optimal solution.2.An improved PDMBBO algorithm was proposed to solve the satellite task scheduling problem.By analyzing the principle of the satellite task scheduling problem,and taking full account of the constraints of storage,energy and time windows,the mathematical programming model with satellite planning time,task allocation balance and task revenue as objective functions is established.The dramatic increase in the size of satellite task scheduling problem has led to an exponential increase in the initial solution space.In response to this phenomenon,an improved PDMBBO algorithm is designed to solve the model.The algorithm adopts the coding rules based on discrete task sequences,the initialization strategy based on increasing the observation opportunities,the differential migration strategy based on exchanging the sequence position of tasks,and the hybrid mutation strategy based on replacing the time window of tasks.The applicability and effectiveness of improved PDMBBO algorithm in satellite task scheduling problem are verified by simulation experiments.In summary,the PDMBBO algorithm not only effectively enhances the optimization performance of traditional biogeography-based optimization,but also has certain applicability in solving satellite task scheduling problems.It provides a new research idea for solving this kind of discrete complex optimization problems.