Long Term Space Program Scheduling and System Design Optimization

The human planetary exploration goal asks for both a strong technology development and a great scientific knowledge enhancement that can be achieved only through a rational and well tuned sequence of different highly related missions, from demonstrators up to scientific probes, in the framework of a limited resources scenario. Therefore the programmatic division has to face the hard task of producing a long term planning deeply connected with the technical design of each unit belonging to a multi-missions scenario. The paper suggests a possible tool to solve the long term space missions planning problem by working on the single mission preliminary sizing, while taking into account the complex constraint net, both in the design and temporal domains. Thanks to the proposed multi-objective optimisation the engineers are given a pruned and ranked solution space, to work on to refine the programmatic. The Evolutionary Algorithms have been selected to deal with mixed domains and get the globally optimal solution set on the time and in design alternatives search space. Appropriate preliminary models are proposed to deal with the criteria vector elements here assumed to be the most relevant for the problem. The architecture has been tested on the NASA Apollo program scenario. The obtained results are consistent with the real program and possible discrepancies helped better tuning the tool. Simulations for ongoing exploration scenarios, such as the ESA Aurora, are offered to highlight the benefits of the tool in identifying a set of optimal, preliminary plans. A critical discussion is also offered.