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

MSc Thesis #9617

Title:Multi-Agent Traffic Simulation - Street, Junction and Traffic Light
Date: 1996
Abstract:Distributed Artificial Intelligence (DAI) is a newly developed branch of AI, the research mainly deals with distributed and concurrent problem solving. DAI has generated two subfields, namely Distributed Problem Sovling (DPS) and Multi-Agent Systems (MAS). The implementation of the Multi-Agent Traffic Simulation System is based on a MAS-based architecture. An urban traffic simulation is a complex and highly interactive problem. The complex connections among streets and junctions, the number of road users with conflicting goals and the finding of optimum traffic light sequences are the main issues concerned. A multi-agent traffic simulation system has been developed to simulate urban traffic scenarios by integrating the domain knowledge in this area, the client/server architecture, the agent interpreter architecture (Agent-0 [Shoham 93]) and various other techniques used in MAS. The system demonstrates intelligent behaviours in the process of solving traffic problems. It provides a platform for future research and development in DAI, in particular MAS.

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