Probabilistic Robust Parallel Design of the Subsystems Constituting a Complex System
| Authors: | Mahmoud Haitham, University of Michigan, United States
Kabamba Pierre, University of Michigan, United States
Ulsoy A. Galip, University of Michigan, United States
Brusher Gerald, Ford Motor Company, United States |
|---|
| Topic: | 5.4 Large Scale Complex Systems |
|---|
| Session: | Large Scale Complex Systems - Applications |
|---|
| Keywords: | Monte Carlo simulation, Optimization, Parallel processing, Probabilistic models, Random searches, Vehicle suspension. |
|---|
Abstract
The design of complex systems, consisting of several subsystems and with performance specifications from multiple disciplines, in parallel was addressed in a previous publication using a Robust Parallel Design (RPD) approach. In this paper, RPD is extended and a Probabilistic Robust Parallel Design (PRPD) approach is proposed to handle cases where the statistical properties of uncertainties are known. Monte Carlo simulation is used to determine the value of a subsystem objective, given the known statistical distributions of uncertainties. Random search techniques (e.g., Simulated Annealing) can then be used to minimize the subsystem objective. PRPD is illustrated using a passive suspension design example of a half-car model.
