A constraint programming approach to automatic layout definition for search results

Abstract

In this paper we describe a general framework based on constraint programming techniques to address the automatic layout definition problem for Web search result pages, considering heterogeneous result items types (e.g., web links, images, videos, maps, etc.). Starting from the entity type(s) specified in the search query and the result types deemed more relevant for the given entity type, we define an optimization problem and a set of constraints that grant the optimal positioning of results in the page, modeled as a grid with assigned weights depending on the visibility.