Research Topics

Redistricting

Redistricting occurs every 10 years following the decennial census. Ideally, the resulting districts are created to provide fair representation for each citizen. in the current system, the redistricting process is easily manipulated so that future election results are essentially known even before votes are cast. Although the process will always involve partisan and interested parties, a free and widely accessible computational tool that provides access to all relevant data and enables users to explore the universe of possible redistricting plans that satisfy certain criteria would make the process eminently more fair and transparent and would engage a much broader array of interested citizens.

This project will develop and implement computational tools that can objectively evaluate proposed redistricting plans and can automate the creation of redistricting plans to satisfy desired criteria. Our tool will provide a forum for state representatives and decision-makers to use when discussing and negotiating redistricting plans. It will allow for the consideration of multiple objectives and parameter weights when evaluating redistricting plans; plans can be compared based on all these criteria. Having such a tool available would eliminate the inherent bias that arises when the data and the ability to propose plans are available only to a few political interests.

To achieve these objectives, this project
1. will formulate the redistricting problem as a discrete optimization problem,
2. will introduce quantitative measurements to score existing and proposed maps with respect to a wide variety of redistricting criteria,
3. will create novel algorithms using optimization tolls customized for the redistricting problem to identify maps that score well on given criteria, and
4. will provide a widely accessible and free computational tool that allows states, individuals, and political parties to negotiate redistricting plans.

Map generalization

Coarse grained parallel genetic algorithm applied to a vector based land use optimization problem