Traffic flow simulation and optimization using evolutionary strategies

No Thumbnail Available

URL

Journal Title

Journal ISSN

Volume Title

School of Science | Master's thesis
Checking the digitized thesis and permission for publishing
Instructions for the author

Date

2011

Major/Subject

Informaatiotekniikka

Mcode

T-61

Degree programme

Language

en

Pages

xii + 80 s. + liitt. 8

Series

Abstract

Different studies, such as the survey that IBM yearly conducts about commuting to work, verify the importance of a well-known problem: the traffic congestion in big cities. Solving this problem has concerned professionals from many scientific and technological disciplines, including Artificial Intelligence (AI). However, we found that the use of Artificial Neural Networks (ANN)-an important tool in this field due to their great features-to control city lights has never been fully seized by any of the past researches, in our opinion, as a consequence of the adaptation process adopted in these investigations. In this thesis, we study the effect of different neuroevolutionary methods in adapting ANNs to efficiently control traffic semaphores. These methods include biological, cultural and linguistic evolution. Furthermore, the performance of our methods is compared with previous approaches using a microscopic traffic simulator enlarged to include different realistic scenarios in a square shaped city. The model has been implemented using a combination of Java language, Netlogo social simulation environment and Matlab. The results of this work illustrate the potential of our concept, which opens the door to further research in the topic and possible expansions to other research areas.

Description

Supervisor

Honkela, Timo|Sáez Achaerandio, Yago

Thesis advisor

Honkela, Timo

Keywords

traffic flow, traffic light, social simulation, evolutionary algorithm, neuroevolution, biological evolution, cultural evolution, language emergence, cellular automata, learning algorithms, artificial neural network

Other note

Citation