A Computational Approach to Redistricting Reform
Legislative redistricting occurs every ten years in the United States, following the decennial census. Ideally, the resulting districts provide fair representation for every citizen. In practice, many district lines are drawn carefully to favor or ensure future election results.
Professor Cho’s research project aims to provide computational tools that will illuminate and open up the redistricting process. A complicating factor is that many possible redistricting plans are extremely similar. Moving a single census block from one district to another does not cause much change; and there are many such minor modifications possible in any redistricting plan, with the magnitude of the problem rising exponentially with the number of geographic units. While an exact optimal solution is computationally intractable, she aims to combine the idiosyncrasies of the redistricting process with a genetic algorithm to produce near-optimal redistricting maps.
During her Center appointment Professor Cho and her research group will extend their library of scalable parallel genetic algorithms for computational analysis of ways to optimize the process of redistricting. They will formulate the redistricting problem as a discrete optimization problem, introduce quantitative measurements to score maps on a variety of redistricting criteria, and develop and apply new algorithms using optimization tools. High-performance computing will allow them to examine the problem at considerably finer spatial scales than ever before.
The project lies at the threshold of applying statistical and mathematical modeling and computing technology to achieve societal tasks. Instead of tinkering with endless possibilities, Professor Cho will develop computationally intensive models to synthesize and organize massive amounts of computation/data to help evaluate redistricting schemes and tailor them to our notions of fairness and democratic rule.