MAPS (Multi-mission Analysis and Planning Support) is a tool dedicated to Earth Observation (EO) satellites to support Mission planners in maximizing both users needs and resources utilization. MAPS comes from an ESA research project with the aim to go beyond the current mono-mission dedicated mission planning systems and offers a useful support tool able to handle ESA/non-ESA missions and also futures missions simulation. It provides an integrated environment for defining and managing on-board

more »

... and on-ground resources in a multi-mission context. Introduction MAPS permits to specify a scenario as a set of areas of interest, time windows and revisiting times. In the scenario the dedicated resources will be selected in terms of satellites/sensors and acquisition stations/antennas. Such resources are configurable in the system though a dedicated and powerful GUI. MAPS enables to specifies many characteristics associated to the resources such as the sensor view angles, satellite duty cycle, instruments modes constraints, antenna AOS/LOS, station unavailability periods but also On-board recorder capacity/rate and Data Relay Satellites in alternative to ground stations. Taking into consideration all these parameters MAPS will deduce first the candidate observations for the scenario, using different kind of propagation (nominal, TLE). One of the main goals of MAPS is then to detect and automatically resolve acquisition conflicts optimizing the user request fulfillment. This complex optimization problem needs a modeling of a wide set of on-board and on-ground constraints, which requires very flexible methods. Systematic search methods (Linear Programming, Constraint Programming), even if they provide an optimal solution, are computational expensive and show a rigid and poor constraint managing, not suitable for EO scheduling [1]. Thus, we have implemented a specific greedy algorithm. The success of the greedy search methods depends largely on the heuristics used to decide the order in which the constraints are solved and which rules are applied [2] [3] [4] [5] [6] . For this purpose, besides the simple priority, we have taken into account other criteria, such as: the number of opportunities of an observation and the percentage of the remaining area to cover in order to fulfill the user request. The scheduler will prefer observations having the fewest remaining opportunities and, also, it will favour the complete fulfillment of a higher-priority user request before beginning the scheduling of another one. We have also developed a stochastic evolution of the greedy method in order to better explore the solution space following the Heuristic-Biased Stochastic Sampling method proposed by Bresina [7].