| The mission of an Earth Observation Satellite (EOS) is to acquire images of specified areas on the Earth surface, in response to observations requests from customers. These images include abundant information of resource and environment on the Earth, which are becoming important to resource exploitation and economic development. The planning of EOS is to allocate satellite system resource optimally, and make the most use of the limited resources in order to satisfy multifarious imaging requests from all kinds of customers.The thesis research the problem from two aspects: modeling and algorithm. And also try to implement the result into a satellite planning management software system.Firstly, this thesis introduced the characteristics and work mode of EOS. Then the imaging requests of customer are analyzed and lots of constraints existing in planning of EOS are summarized. After study the basic theory of Artificial Intelligent Planning, the truth that the planning of EOS is a complex problem integrate many time and resource constraints is obviously. The thesis simplifies the problem of planning of EOS and only the most important constraints in actual application are considered in succeeding analyzing. By referring to the constraint satisfaction problem (CSP) model, the corresponding model of planning of EOS is also established and the formular representation is provided.Three algorithms are studied to resolve the problem of planning of EOS, including greed algorithm, simulated annealing algorithm and genetic algorithm. The thesis detailed work flow of each algorithm and also develop the key operators. At last, the performance of each algorithm is compared through a simple application example.On the basis of this research, the planning management software system of EOS is developed with the application of the model and algorithm. The important function module in the system, planning function module, is introduced at the last part of the thesis. |
PDF Full Text Request