Search methods in dynamic networks usually cannot rely on a stable topology from which shortest or otherwise optimized paths through the network are derived. When no reliable search indices or routing tables are provided, other methods like flooding or random walks have to be considered to explore the network. These approaches can exploit partially available information on network paths, but the search effort naturally increases with the lack of precise paths due to network dynamics. The problem is especially relevant for wireless technology with strict limitation on power consumption. We compare the efficiency of random walks and flooding for exploring networks of small to medium size. Several scenarios are considered including partial path information support for search. Transient analysis and a bound are applied in order to evaluate the messaging overhead.