We consider a two-echelon location-distribution problem arising from an actual application in fast delivery service. This problem belongs to the class of adaptive logistics problems, as the locations of the facilities (typically, parking spaces) are revised on a daily basis according to demand variations. We present and compare two formulations for this problem: an arc-based model and a path-based model. Since these formulations cannot be solved in reasonable time for large-scale instances, we introduce a heuristic method based on a variable neighborhood search approach.