Seed-Fill-Shift-Repair: A redistricting heuristic for civic deliberation

Area of Science:

• Computational Social Science
• Political Science
• Computer Science

Background:

• Political redistricting redraws electoral boundaries, often to reflect population shifts.
• Gerrymandering, the manipulation of district lines for partisan advantage, leads to undemocratic outcomes and is a significant issue in the United States.

Purpose of the Study:

• To introduce the Seed-Fill-Shift-Repair (SFSR) algorithm for creating redistricting plans.
• To provide an accessible implementation of SFSR for public deliberation and research.
• To advocate for a constraint satisfaction approach to districting, emphasizing the generation of multiple feasible solutions.

Main Methods:

• The Seed-Fill-Shift-Repair (SFSR) algorithm is developed and implemented.
• SFSR is applied to generate redistricting plans for U.S. Congressional districts.
• The algorithm's output is analyzed using descriptive statistics.

Main Results:

• SFSR successfully generated a corpus of 11,206 valid redistricting plans for Pennsylvania.
• 1,000 plans were generated for Arizona, Michigan, North Carolina, Pennsylvania, and Wisconsin.
• Descriptive statistics on plan characteristics and computational methods are presented.

Conclusions:

• SFSR offers a novel computational approach to generating a plurality of valid redistricting plans.
• The algorithm and its outputs can facilitate public discourse and research on redistricting.
• Making the code publicly available supports broader use in civic and academic contexts.