Research On Robust Scheduling Approach And Its Application Of Imaging Satellites

Imaging satellite is a kind of Earth Observation Satellite (EOS) acquiring remote sensing information from outer space. The imaging requirments will augment rapidly with the increasement of the imaging satellites, the time efficiency and accuracy of the imaging tasks will be more and more rigorous and the complexity of task management will be enhanced largely. Scheduling of imaging satellites means to allocate multiple satellite resources to multi competitive tasks without conflict and ascertain the execution times of those tasks according to the user's requirements and satellite capabilities, and makes the most use of the limited resources in order to satisfy multifarious observing requirements in future. So far, researches have been primarily focused on imaging satellites scheduling problem without uncertainty. In practice, the satellites are working in a complex environment, faced with different uncertainties such as insertion of new tasks, cancellation of arranged tasks, change of task properties, change of weather condition, unanticipated changes in satellite resources and etc. Based on the summarizing and analyzing of research works inside and outside, this thesis devides the whole problem into two phases including the creation of robust solution phase and the dynamic adjustment of robust solution phase. The main research contents and achievements of this thesis are as follows:(1) Putting forward the imaging task gain computation method and the neighborhood-based robustness measure, and presenting the imaging satellites robust scheduling model. On the basis of analysis on the main contributing factors, we extend the task gain computation method to take into account the importance of observing target, image quality and remaining feasibility opportunities at the same time. We then propose a neighborhood-based robustness measure for scheduling which is inspired by advances in robust optimization of continuous functions, analyse the main constraints and present the imaging satellites robust scheduling model.(2) Proposing a preference-based hierachical multi-objective genetic algorithm (PHMOGA) to solve the imaging satellites robust scheduling model. This thesis designs PHMOGA which is inspired by hierachical places of the culture algorithm and preference-based weighted Pareto. According to the character of imaging satellites robust scheduling, we use the appropriate encoding method to depict the solution, give a greedy randomized insert algorithm to construct initial feasible population for evolution, propose a knowledge evolution strategy based on the knowledge evolution space, design the task sequence-based crossover operator and multi-mode mutation operator, and analyze the convergence for PHMOGA.(3) Presenting the computation methods of earlist start execution time and latest start execution time, and providing the feasibility analysis and parameters update methods of genetic operators. Time constraint and energy constraint are primary constraints in imaging satellites scheduling problem. This thesis provides the notions of critical task sequence and backward energy load and presents the computation methods of earlist start execution time and latest start execution time. The feasibility analysis and parameters update methods are then provided for task transferring operation and task exchanging operation.(4) Presenting the imaging satellites dynamic scheduling model and its dynamic insert tasks heuristic algorithm (DITHA) with regard to the dynamic adjustment problem of imaging satellites robust solution. This thesis describes the problem with a unified form of dynamic problem with tasks insertion considering the different perturbations, and presents the imaging satellites dynamic scheduling model and its dynamic insert tasks heuristic algorithm. The demonstration calculation and experiments on test instances show that result of the adjustment depends both on adjusting approach and robustness of the solution.(5) Introducing the design and realization of the imaging satellites planning and scheduling system with the application of the aforementioned achievements. Also an example application is given and solved by the system, the result confirms the validity of imaging satellites robust scheduling approach provided in this paper.