A stochastic simulation model to support designs of experiments on transportation evacuation planning.

摘要

This research presents a new discrete-event simulation model, the DOE_EVAC, that can (i) effectively simulate alternative modes of transportation during evacuations, (ii) support designs of experiments, thus, provide the users (e.g., emergency planners and traffic engineers) with means to investigate "what-if" scenarios with sound statistical analysis capabilities, and (iii) allow the users to build and execute these models without having to know complex simulation or coding languages.;The contributions of this research are threefold. First, this research adopts designs of experiments to furnish users with statistical support to investigate "what-if" scenarios. Second, the DOE_EVAC model resolves existing issues of current simulation transportation evacuation modeling approaches by improving the initial system setup and supporting the stochastic traffic loading process. It allows users to implement and analyze various traffic management strategies, and is capable of rerouting traffic due to critical infrastructure failures during evacuation. DOE_EVAC also supports user-interruptions during simulation runs so that changes on the system can be executed. Finally, the DOE_EVAC model does not require the user to have any knowledge of specialized simulation or coding language as it relies only on four required (and one optional) data files supplied by the users to execute. A sample design of experiment is illustrated to show the multiple simulation run capabilities of the DOE_EVAC model and the ability of the DOE_EVAC to allow users to manipulate data that are typically inaccessible in existing evacuation models.