Colonial Algorithm: A Quick, Controllable and Visible One for Gerrymandering

Abstract

Gerrymandering is an issue of boundary delimitation for legislative purposes. In this paper, we presented a novel approach called Colonial Algorithm(CA) to draw legislative boundaries. Based on Voronoi Diagram, CA was developed originally for simulating colonies’ behavior in a culture dish. We focused on the expanding speed and the competitive strategies on edges to create our algorithm and used Cellular Automaton to validate it. CA can be well applied in gerrymandering. When generated from places with big population density, every region undergoes a process of expansion and competition to finally reach a partition. In optimization, population equality and regional compactness were adopt to form the value function. To illustrate the process, a simplified example was taken, whose result indicates that regions drawn by Colonial Algorithm attain continuity, compactness, small population variances and relative geometric simplicity within a short time. And the whole process is controllable and visible.