Better Forecasts Through Cloud Computing



Better Forecasts Through Cloud Computing

Authors

Tor Vorraa, Citilabs

Description

This paper will present the idea of cloud computing for our transport and traffic models describe experiences from some real world examples.

Abstract

Sharing knowledge and collaboration is key to the success of any planning project. Cloud based solutions are often mentioned as the future for computing for its ability to make sharing and collaboration easily possible. Such solutions can and should be used to improve the quality of transport and traffic forecasting.

The cloud offers a scalable, high performance computing environment. This type of hardware platform allows for calculations to be spread across vast numbers of cores in an efficient way. This kind of horsepower dramatically reduces the run-time associated with large and complex models. It provides the perfect platform to apply large urban-area models.

Often, analysts needing to run a model several times within a short period are frustrated by the limitations of the hardware they have available to them. In a cloud based solution, users can start tens or even hundreds of scenarios simultaneously. Hardware is not a constraint. This feature can improve the quality of transport analysis by making it possible to rapidly and easily undertake a larger number of sensitivity and reliability tests.

The safe-sharing capability within a cloud based solution can provide a helpful feature for those maintaining one or more models for multiple public and private-sector end-users, or anyone needing to deliver a turn-key modelling solution. And, it provides an easy way to share maps, charts and tabular summaries with management and decision-makers.

This paper will present the idea of cloud computing for our transport and traffic models describe experiences from some real world examples.

Paper to be given by :
Tor Vorraa, Director at V&B Software Services Ltd , UK and Regional Director for Citilabs Inc., USA.

About the author :
Tor Vorraa graduated (MSc) from the Norwegian University of Science and Technology (now NTNU) in 1983. He has since worked as a transport planner and forecaster in Norway, Germany and the UK. Tor joined MVA in 1997 as the international manager for the software TRIPS and has been with Citilabs, the developer of the Cube travel demand forecasting system, from its start in 2001. Tor also runs a service company, V&B Software Services Ltd., set up to help public authorities and their consultants to make better use of Citilabs software.

Publisher

Association for European Transport