A mathematical model and application on the prevention of gerrymandering

Official URL: http://risc01.sabanciuniv.edu/record=b2058055 (Table of Contents)

The significance of the political districting also known as redistricting has been recognized by several countries across the world since electoral district boundaries can be manipulated for a political gain. This manipulation practice is known as gerrymandering and has serious influences on the results of an election. In this thesis, we tried to show how easy policymakers can misuse the redistricting practice to gain a political advantage such as increasing the number of their representatives in the parliament. Two different mathematical models have been developed for different types of election systems. Single-member district electoral system in which the only representative can be elected from each electoral district is one of them. Every county in ˙Istanbul has been tried to be divided into their single-member districts considering the total number of representatives of the county. In addition to the first model, another formulation has been developed to also cover the multi-member district systems. The main drawback of the mathematical models is that they are only working on the small cases in terms of the total number of political units. Tabu search algorithm has been developed to answer the cases that cannot be classified as small. The required algorithm steps such as initialization, neighborhood change structure etc. are explained in detail. The results of the mathematical models and the algorithm have been achieved and visualized aesthetically.