Integrating GIS and location modeling: A relational approach

Abstract

AbstractPlanners and geospatial analysts have long been using GIS and optimization methods to solve a wide range of spatial decision problems from routing and facility location to political redistricting. However, there has been a notable disconnection between the tools of GIS and optimization, constraining the capability of spatial decision‐making in GIS. This article presents relational linear programming (RELP), a new framework and tool for bridging the gap between the two fields by integrating a generic optimization solver into GIS: a solver for integer linear programming (ILP). In particular, we choose relational/spatial databases as the GIS platform. We demonstrate that not only data, but also programs for spatial problems (in the form of ILP) can be managed and executed in SQL inside spatial databases. The main focus of this article is on location modeling, and we show how classic models can be formulated succinctly in a declarative way using SQL. The declarative approach of RELP allows researchers to focus on formulating new spatial decision problems without being bogged down in sophisticated programming. The integrated approach allows data processing and decision‐making in one language. Therefore, it brings the power of location modeling to non‐specialist GIS practitioners.