MSc. Thesis Defense:Dicle Yağmur Özdemir
  • FENS
  • MSc. Thesis Defense:Dicle Yağmur Özdemir

You are here



Dicle Yağmur Özdemir
Industrial Engineering,  MSc. Thesis, 2017


Thesis Jury

Prof. Dr. Ali Rana Atılgan, Assoc. Prof. Dr. Güvenç Şahin

, Assist. Prof. Dr. Ahmet Onur Durahim



Date & Time: 3rd January, 2017 –  13:40

Place: FENS G015

Keywords : Segregation, Trustworthiness, Agent-Based Simulation




In this thesis, we assume that there is a population of individuals, whom we refer to as agents. Each agent is colored green or red. We think of the two types as representing some immutable characteristic, for example, race or ethnicity. The agents are located in the cells of a lattice (and in general a complex network), intended as a toy model to focus on the interactions between agents. The neighbors of a cell on the lattice are those that touch it, including the diagonal contacts; thus, a cell has eight neighbors. Each agent desires to have at least some other agents of its type as neighbors. Black cells denote empty spaces where unfulfilled agents may move. Schelling assumed that there is a threshold t common to all agents: if an agent discovers that less than t of its neighbors are of the same type as itself, then it has an interest in moving to a new cell [1]. First, each unhappy agent moves to an empty space ranadomly. This may cause other agents to become unsatisfied; then, the second round of movements begins. After a finite number of rounds, representing a fixed time period during which unsatisfied agents have changed where they live, we do not find any agent who is unhappy. We then analyze what type of patterns emerge in the “equilibrated” community and how trustworthiness affected. This simple model of Schelling illustrates how the forces leading to segregation (phase separation) are remarkably robust.  Even if people only have a mild preference for living with neighbors of the same color, a significant level of segregation may occur.

Inspired from the fact that some respondents want to live in mixed neighborhoods rather than where there are mostly populated by the same color, we included in the model an upper bound of like-colored agents for each agent’s satisfaction function. In our model, therefore, the agents are not only tolerant to opposite colors, but deliberately seek diversity. A real life counterpart of this diversity desire is that in multi- ethnic communities where the agents reportedly said that they want their children to learn to get along with others.In our work, we also introduce several measures to quantify different patterns emerging from both tolerant and diversity seeking agents. The conventional segregation measures cannot appreciate the differences. We have defined new segregation indices that represent well the distinction between the segregated and integrated regions that appear due to variety of happiness (utility) functions.

[1] Schelling, T., “Dynamic models of segregation,” Journal of Mathematical Sociology, Vol. 1, 1971, pp. 143 – 186.