Using the Stochastic Fleet Estimation Model to Compute Optimized Fleet Mixes
Intervenant(s) : Slawomir Wesolkowski (Centre for Operational Research and Analysis, Defence R&D Canada)
An organization involved in the transportation of people and freight needs to decide how many platforms or vehicles it needs to accomplish the tasks that it wants to perform. This presentation is motivated by the need to provide solutions based on task requirements in order to assist decision makers. The Stochastic Fleet Estimation (SaFE) model was devised to estimate the size and composition of an average fleet of platforms, based on task requirements (stochastic in nature), for an air mobility/airlift problem. Within the SaFE model, tasks need only occur at some point within the considered time interval (as opposed to a typical discrete event simulation), and tasks are not considered to have closure time requirements, i.e., the maximum allowable time to complete a task. The model is used within a multiobjective optimization framework to analyze platform demand based on a large number of future possible scenarios in order to obtain the overall minimum platform requirements. Various objectives such as total task duration (performance), fleet mix cost (cost), and probability of not being able to complete a scenario (risk) will be discussed. Optimization results including example fleet mixes will be presented.
Sahar.Changuel (at) nulllip6.fr