A Spatial Decision Support System Combining GIS and OR Tools to Optimize District Boundaries and Bus Routes for a Suburban School District

This paper describes a GIS-based spatial decision support system (SDSS) developed to optimize school district boundaries and bus routes in a large suburban school district, Sweet Home Central Schools. This school district is one of the three in the suburb Amherst (population 110,000) of the City of Buffalo, New York State, USA. The school district prompted the study as a result of demographic changes that have affected enrollment distribution among its four elementary schools as well as a recent construction project affecting traffic flow into and out of the school bus depot. In consultation with the school district officials two approaches are developed as part of the SDSS: a clean slate and a community approach. The clean slate approach focuses on minimizing transport costs, while the community approach focuses on the least disruption to the people in the community and balancing socio-economic and diversity issues. Scenarios for both approaches are developed based on current enrollment data. Introduction Problems of choosing bus stops, allocating students to those stops and optimizing routes to visit each of those stops is a rather common problem that many authors have offered solutions to (Bowerman et al., 1995; Corberan et al., 2002). There is also literature discussing applications on districting problems such as creating political voting districts and school districts (D'Amico et al., 2002; Crespin, 2005; Sutcliffe et al., 1984). These types of problems are inherently geographic and have been the interest of geographic information systems (GIS) and operations research (OR) researchers for some time. Recently, the idea of integrating GIS and optimization tools to address school redistricting has been introduced (Caro et al., 2004; Derekenaris et al., 2001; Taylor et al., 1999). This area is still being explored as improvements to existing tools and new developments continue to provide opportunities for growth. To contribute to this field, a spatial decision support system (SDSS) for redistricting is presented in this paper. The system proposed helps district officials create different redistricting scenarios for evaluation purposes, to identify the one that best suits the school district and community needs. Even though the idea of creating such a system might not be new, there are important differences in terms of implementation, use of the GIS, and approaches adopted that make the system unique. In conversations with the redistricting task force of the school district which prompted the creation of the SDSS (e.g. Sweet Home Central), it was clear that a pure quantitative OR approach was not adequate to address the social concerns of stability (children and their siblings going to the same school), economic and cultural diversity, and special needs of the schools. Therefore, the SDSS proposed differs from previous implementations in two distinct ways: (1) the data used for redistricting is at the most disaggregate level available, namely the children's home (grade, 1 [email protected] corresponding author. 410 Bell Hall, University at Buffalo, Buffalo, NY 14260, USA