A　novel　optimization　algorithm,　Group　Area　Search　(GAS),　is　proposed,　which　is　inspired　by　searching　behavior　patterns　of　human　beings　and　social　animals.　In　GAS,　the　search　area　of　each　individual　is　automatically　adjusted　and　gradually　shrunk　to　the　most　promising　region.　A　cruising-following　mechanism　is　introduced　to　GAS,　which　allows　individuals　with　low　fitness　chances　to　follow　the　historical　best　individual.　The　algorithm　strikes　a　good　balance　between　global　search　and　local　search.　The　experimental　results　on　6　benchmark　functions　show　that　GAS　has　good　performance　on　both　unimodal　and　multimodal　test　functions,　especially　on　multimodal　ones.　It　significantly　outperforms　six　other　population-based　algorithms.　It　shows　potential　to　solve　complicated　function　optimization　problems.