Title: Exploring complex networks by walking on them

Abstract: We carry out a comparative study on the problem for a walker searching on several typical complex networks. The search efficiency is evaluated for various strategies. Having no knowledge of the global properties of the underlying networks and the optimal path between any two given nodes, it is found that the best search strategy is the self-avoid random walk. The preferentially self-avoid random walk does not help in improving the search efficiency further. In return, topological information of the underlying networks may be drawn by comparing the results of the different search strategies.