Cooperative searching for stochastic targets

TL;DR

This paper develops a formal framework for cooperative spatial search involving multiple agents, demonstrating how information sharing can enhance search efficiency in stochastic target scenarios through analytical and numerical methods.

Contribution

It introduces a novel formalism for cooperative search with information sharing, providing analytical and numerical insights for optimizing multi-agent search strategies.

Findings

Sharing information improves search efficiency in certain tasks.

Agents can achieve synergy by aggregating local information.

The framework guides designing optimal search algorithms.

Abstract

Spatial search problems abound in the real world, from locating hidden nuclear or chemical sources to finding skiers after an avalanche. We exemplify the formalism and solution for spatial searches involving two agents that may or may not choose to share information during a search. For certain classes of tasks, sharing information between multiple searchers makes cooperative searching advantageous. In some examples, agents are able to realize synergy by aggregating information and moving based on local judgments about maximal information gathering expectations. We also explore one- and two-dimensional simplified situations analytically and numerically to provide a framework for analyzing more complex problems. These general considerations provide a guide for designing optimal algorithms for real-world search problems.