Evolutionary combination of connected event schemas into meaningful plots

Abstract Many of the stories we are exposed to built from small schemas connected events involving a set characters– boy meets girl leads relationship or crime revenge . The present paper proposes an evolutionary solution task putting together story by combining such schemas. This approach presents three challenges: how mix up elements in different schemas, instantiate characters across and tell acceptable combinations rest. applies that relies on genetic representation for these as fitness functions metrics compatibility constraints schema combinations. Outputs this procedure evaluated human judges comparison with baseline solutions which values genes assigned at random. proposed generates population drafts resemble plot descriptions simple stories. results comparative evaluation positive. pattern pairs provide valid constructing plots.