390 likes | 818 Views
CLOUD COMPUTING FOR AGENT-BASED URBAN TRANSPORTATION SYSTEMS. VIVEK.R ROLL NO:18 DATE:17-02-12 S1 MCSE SLOT NO: 2. OVERVIEW. INTRODUCTION HISTORY AGENT-BASED TRAFFIC MANAGEMENT SYSTEMS CHALLENGES INTELLIGENT TRAFFIC CLOUDS REFERENCES. INTRODUCTION.
E N D
CLOUD COMPUTING FOR AGENT-BASED URBAN TRANSPORTATION SYSTEMS VIVEK.R ROLL NO:18 DATE:17-02-12 S1 MCSE SLOT NO: 2
OVERVIEW • INTRODUCTION • HISTORY • AGENT-BASED TRAFFIC MANAGEMENT SYSTEMS • CHALLENGES • INTELLIGENT TRAFFIC CLOUDS • REFERENCES
INTRODUCTION • Agent-Based Traffic Management Systems • Cloud computing can help such systems to deal with large amounts of storage and computing resources • Development of Traffic control systems within evolving computing paradigm • Agent-Based Distributed and Adaptive Platforms for Transportation Systems (Adapts)
HISTORY • IBM 650 was first introduced to an urban traffic-management system in 1959 • Traffic control and management paradigm has five phases
At second stage a microcomputer could handle a single user’s requirements • Traffic Signal Controller (TSC) had enough capacity to control one intersection
In phase three, LANs appeared for resource sharing. • Traffic model became hierarchical
In the internet era, users could retreive data and from remote sites and process them locally • To reduce loss of bandwidth, mobile agents were introduced
Fifth computing paradigm- Cloud computing • Users do not need to know the infrastructure in the “clouds”
Parallel transportation Management System (PtMS) • Term ‘Parallel’ means the parallel interaction between an actual transportation system and their virtual counterparts • PtMS use Artificial Transportation Systems (ATS)
AGENT-BASED TRAFFIC MANAGEMENT SYSTEMS • Agent technology was used since 1992 • Multiagent systems came later • Mobile agents became popular in 2004 • Move through the network • Traffic device only need an operating platform
In 2005, ‘Adapts’ was proposed as a hierarchical urban traffic management system • It has three layers • Organization • Coordination • Execution • Currently Adapts is part of PtMS
Organization Layer • Functions • Agent-oriented task decomposition • Agent scheduling • Encapsulating traffic strategy • Agent management • Consists of • A management layer • Three databases • Artificial transportation system
Databases are • Control strategy • Typical traffic scenes • Traffic strategy agent • Code of new traffic strategy is saved in traffic strategy database • It is encapsulated into a traffic strategy agent and saved in it’s database
Traffic strategy agent tested with typical traffic scenes • Management agent embodies organization layer’s intelligence • Agent’s scheduling and agent-oriented task decomposition is based on MA knowledge base
When an unknown traffic scene is encountered • Urban management system sends a traffic task to organization layer • It is decomposed into a combination of traffic scenes • MA will return a combination of most appropriate agents and a map about their distribution
Testing System Performance • Set up an ATS to test performance of the urban-traffic management system • Computational experiments are faster than real world • If unsatisfactory, both systems will be modified
NEW CHALLENGES • Agent-distribution map and relevant agents need to be sent to ATS for experimental evaluation • A test was conducted to find the cost of this operation • If the time to complete evaluation exceed a threshold, results will become useless and meaningless
In the test, they used a 2.66-GHz PC with 1GB memory to run both ATS and Adapts • It took 3600s in real time • Number of intersections increased from 2 to 20
The time required to run ATS and Adapts experiments on one PC
When number of traffic control agents is 20, experiment takes 1,130 seconds • If time threshold is set to 600 seconds, maximum number of intersections in one experiment is only 12 • This is insufficient for major cities like Beijing • We will need several PCs or a high performance server
Future Systems • The future systems must have the following capabiliites • Computing Power • Testing a large amount of typical traffic scenes requires lot of computing resources • If a traffic strategy trains on actuator, it will damage the performance of the traffic AI agent • Better to train AI agent before moving it to the actuator
Storage • Vast amount of traffic data like configuration of traffic scenes, regulations and information about agents in ATS need vast amount of storage • Two solutions • Implement a super computer with all centers of urban-traffic management systems • Use cloud computing technologies. • For eg: Google’s Map-Reduce, IBM’s Blue Cloud and Amazon’s EC2
INTELLIGENT TRAFFIC CLOUDS Overview Of Urban-traffic management systems based on cloud computing
Prototype • Urban-traffic management using intelligent traffic clouds • It will go far beyond other multiagent traffic management systems • It has two roles • Service provider and • Customer
Service providers include ATS, traffic strategy database and traffic strategy agent database • They are all in system’s core: intelligent traffic clouds • Customers include urban-traffic management systems and traffic participants • They exist outside the cloud
Could provide traffic-strategy agents and agent-distribution maps to the traffic management systems • Numerous traffic management systems could connect and share cloud thereby saving resources • New strategies can be converted to mobile agents
Architecture • Intelligent traffic clouds have four architecture layers • Application • Platform • Unified source • Fabric
REFERENCES [1] D.C. Gazis, “Traffic Control: From Hand Signals to Computers,”Proc. IEEE, vol. 59, no. 7, 1971, pp. 1090–1099 [2] F.-Y. Wang, “Toward a Revolution in Transportation Operations: AI for Complex Systems,” IEEE Intelligent Systems, vol. 23, no.6, 2008,pp. 8–13 [3] F.-Y.Wang, “Parallel Control and Management for Intelligent Transportation Systems:Concepts,Architectures, and Applications,” IEEE Trans.Intelligent Transportation Systems, vol.11,no.2, 2010,pp.485-497
REFERENCES CONT’D…. [4] B. Chen and H. H. Cheng, “A Review of the Applications of Agent Technology in Traffc and Transportation Systems,” IEEE Trans. Intelligent Transportation Systems, vol. 11, no. 2, 2010, pp. 485–497.