One day at SFI, physicist and SFI Professor Sid Redner and computer scientists Melanie Moses and Stephanie Forrest found themselves discussing the difficulties of modeling ant foraging behaviors.

One of Redner’s research interests is stochastic search (in which the searcher has no information about the location of the target). In May the trio pulled together a number of experts, including SFI Postdoctoral Fellow Caterina de Bacco, to focus on stochastic search with reset. In these conditions, a search process starts with little information about the target’s location and is intermittently stopped and restarted.

An ordinary stochastic search in one or two dimensions will, with infinite time available, and its target. But a search in three dimensions may never and its target, even in infinite time; in this many dimensions, the searcher is too likely to go o on wild tangents.

Counterintuitively, though, in search with reset, the target is found infinite time in any number of dimensions, guaranteed; reset cuts short those fruitless expeditions to nowhere.

Reset not only adds these fascinating qualities to search probabilities, Redner says, it is a practical consideration, too – as when ants or a search- and-rescue teams call it quits for the night, but commence their searches each morning.

Future models might incorporate more realistic considerations, he says.