The University of Edinburgh -
Division of Informatics
Forrest Hill & 80 South Bridge

MSc Thesis #9615

Title:Multi-Agent Traffic Simulation-Vehicle
Date: 1996
Abstract:As the number of vehicles increases on the roads in every major city of the world today, the daily life of a city commuter consists of waiting for long periods of time in heavy traffic congestion. With the current computer communication technologies, various methodologies have been sought and tried in an attempt to solve a traffic jam and achieve effective traffic control. Distributed artificial intelligence (DAI) has emerged recently and one of its subfields uses a multi-agent paradigm (MAS) to solve problems by coordinating the knowledge, goals and plans of semi-autonomous intelligent agents to solve a problem collaboratively. The principal goal of this project is to adopt the multi-agent paradigm and build an experimental environment, the Multi-Agent Traffic Simulation System, to model and simulate our traffic scenarios which can be used to develop a (distributed) intelligent system for urban traffic problems. The resulting Traffic Simulation System shows the flexibility to simulate and model different driving scenarios of varying complexity. It demonstrates the multi-agent paradigm is an applicable methodology, and provides a natural framework for modelling the entities in our traffic world.

[Search These Pages] [DAI Home Page] [Comment]