Abstract
One of the most important aspects in the general search problem of finding randomly located target sites concerns how to characterize the role played by the non-revisitability delay time τ during which a previously found target becomes unavailable to the searcher. By using an appropriate parameterization of the number of random walk steps undertaken between successive targets, we show that for the case of sparse randomly distributed sites the optimal search strategy shifts from a superdiffusive to a ballistic strategy consisting of essentially rectilinear motion between the targets, as τ increases from τ → 0 to τ → ∞, respectively. The crossover between these limiting regimes occurs as a function of τ. These conclusions are shown to hold even if dissipative phenomena are considered in the searching dynamics. We discuss the results in the context of their application to animal foraging.