This note briefly describes a partitioning algorithm (FRACTOP) for global optimization, which, in the search for the global optimum of simple bounded multimodal functions, evaluates nonoverlapping partitions of the feasible region by random search. Fuzzy measures are used in the assessment of samples taken from all partitions, and the partition with the highest tendency of containing the global optimum is repartitioned to refine the search in that region. The performance of FRACTOP where fuzzy measures are imbedded is compared with a statistical promise measure proposed by Tang. Computational results are reported on an extensive set of 77 test functions collected from the literature.