This paper addresses the problem of designing a hierarchical single-allocation hub-median (SA-H-HM) network considering fuzzy flows between nodes. The problem is modeled as a fuzzy mathematical programming model and a hybrid algorithm of population-based iterated local search (PILS) and fuzzy simulation is employed. Results clearly show that PILS is efficient in reaching solutions with virtually all the errors less than one percent to the optimal solutions. Moreover, the proposed PILS is capable to escape local optima. Finally, the results of the hybrid algorithm give insights about the problem under uncertainty. © 2013 Springer-Verlag Berlin Heidelberg.