The ground station scheduling problem is a complex scheduling problem involving multiple objectives. Evolutionary techniques for multi-objective optimization are becoming popular among different fields, due to their effectiveness in obtaining a set of trade-off solutions. In contrast to some conventional methods, that aggregate the objectives into one weighted-sum objective function, multi-objective evolutionary algorithms manage to find a set of solutions in the Pareto-optimal front. Selecting one algorithm, however, for a specific problem adds additional challenge. In this paper the ground station scheduling problem was solved through six different evolutionary multi-objective algorithms, the NSGA-II, NSGA-III, SPEA2, GDE3, IBEA, and MOEA/D. The goal is to test their efficacy and performance to a number of benchmark static instances of the ground scheduling problem. Benchmark instances are of different sizes, allowing further testing of the behavior of the algorithms to different dimensionality of the problem. The solutions are compared to the recent solutions of a weighted-sum approach solved by the GA. The results show that all multi-objective algorithms manage to find as good solution as the weighted-sum, while giving more additional alternatives. The decomposition-based MOEA/D outperforms the rest of the algorithms for the specific problem in almost all aspects.
Worst-case scenario optimization deals with the minimization of the maximum output in all scenarios of a problem, and it is usually formulated as a min-max problem. Employing nested evolutionary algorithms to solve the problem requires numerous function evaluations. This work proposes a differential evolution with an estimation of distribution algorithm. The algorithm has a nested form, where a differential evolution is applied for both the design and scenario space optimization. To reduce the computational cost, we estimate the distribution of the best worst solution for the best solutions found so far. The probabilistic model is used to sample part of the initial population of the scenario space differential evolution, using a priori knowledge of the previous generations. The method is compared with a state-of-the-art algorithm on both benchmark problems and an engineering application, and the related results are reported.