Akademska digitalna zbirka SLovenije - logo
Institut Jožef Stefan, Ljubljana (IJS)
  • The lunar lander landing site selection benchmark reexamined [Elektronski vir] : problem characterization and algorithm performance
    Vodopija, Aljoša ; Cork, Jordan ; Filipič, Bogdan
    The benchmark problem of lunar lander landing site selection is a nonlinear constrained optimization problem introduced as a challenge for the 2018 Evolutionary Computation Symposium competition. In ... the single-objective variant, the task is to find the latitude and the longitude of the landing site maximizing the total communication time with Earth, and satisfying constraints for consecutive shade days and the landing site inclination angle. The requirements of the multiobjective variant are to maximize the total communication time, and minimize the consecutive shade days and the landing site inclination angle, while satisfying the constraints for the latter two objectives. We reexamine the problem by first characterizing it by means of problem landscape analysis, then performing a comparative study of algorithm performance, and finally relating the results of the two phases. The characterization relies on both the features known from the literature and the recently constructed ones for analyzing constrained and multiobjective problem landscapes, while the algorithmic study includes several single- and multiobjective evolutionary algorithms. The findings are insightful and allow for a better understanding of both the problem and the optimization algorithm behavior. They help identify appropriate algorithms for landing site selection and are advantageous for solving comparable problem instances.
    Vir: GECCO'24 [Elektronski vir] : proceedings of the 2024 : Genetic and Evolutionary Computation Conference : Melbourne, Australia, 14-18 July 2024 (Str. 1381-1389)
    Vrsta gradiva - prispevek na konferenci ; neleposlovje za odrasle
    Leto - 2024
    Jezik - angleški
    COBISS.SI-ID - 202434563

    Povezava(-e):

    https://dl.acm.org/doi/pdf/10.1145/3638529.3654229
    DOI

Vnos na polico