Keywords: data acquisition system, heterogeneous computing, human computer interaction, parallel and distributed processing, risk analysis, workflow manager.

In this paper, a workflow management system and a web-based cluster front-end developed for scheduling simulations is presented. The workflow, generated via a web interface is described as well as the integration from the web server into the cluster. User interaction was achieved on multiple stages. The user can interact with the system by specifying the input parameters that the simulations should vary. One can add, start and stop the simulation on the cluster, upload or create scripts and parse the resulting data to the server location. The user is also able to execute existing executable scripts located on the cluster. The process of every sub-simulation can be followed heterogeneously with a maximum of one minute delay. How the core allocation decision is being taken by the scheduler itself, which follows an agent-based protocol inspired by the FIPA-Contract-Net protocol of communication is discussed. Thus, the scheduler itself always inspects the level of coverage of the preconditions that each of the jobs has to meet.

The strategy that was followed for integrating different simulating models that were developed for mesh generation, traffic simulation, airflow simulation, smoke propagation simulation, ventilation strategy simulation is presented with an evacuation simulation that are used to estimate the consequences of a fire in a tunnel and finally approximate its risk.

purchase the full-text of this paper (price £20)

go to the previous paper
go to the next paper
return to the table of contents
return to the book description
purchase this book (price £75 +P&P)