Neutral Redistricting Using a Multi-Level Weighted Graph Partitioning Algorithm
In the 1960s, Nobel Prize winning economist William Vickrey posited that computers would soon be able to draw legislative districts without bias in an automatic and impersonal way. We have created a new algorithm for redistricting that shows no indication of bias that is also efficient enough to handle the largest jurisdictions analysts might examine.
The algorithm we propose is much more efficient than alternative approaches to the same problem. The improvement in efficiency means that
- analysts may use more detailed data to develop maps;
- analysts can evaluate maps in larger more complex settings;
- analysis can use standard desktop or laptop computers to evaluate maps as opposed to large, costly computing clusters or super computers. The algorithm exhibits no indication of bias in the way it draws maps of districts. Alternative approaches have been shown to draw particular types of districts which precludes a valid comparison to maps analysts may wish to evaluate.
Binghamton University RB538