Initial volume fluctuation (VF) arising from the participant fluctuation would be the background which should be subtracted experimentally from the measured higher-order cumulants. We study the validity of the Volume Fluctuation Correction (VFC) on higher-order net-proton cumulants by using simple toy model and UrQMD model in Au+Au collisions at \sqrt s_{NN} = 200 GeV for various centrality definitions. The results are compared to the conventional data driven method called Centrality Bin Width Correction (CBWC). We find VFC works well in toy model assuming independent particle production (IPP), but does not seem to work well in UrQMD model. It is also found that cumulants are strongly affected by the multiplicity correlation effect as well as the centrality resolution effect. These results show that neither VFC nor CBWC are perfect method. Thus, both methods should be compared in the real experiment.

8 pages and 8 figures