Bayes Inference in Multicriteria Analysis for Hybrid Electrical Transportation Systems Design

Abstract

An advanced stage has been reached in the research and development of alternative vehicle propulsion systems, aiming at assuring acceptable levels of both consumption and emissions. Many different criteria, often expressed in incommensurable units, such as weight, reliability or availability, cost, etc., must be taken into account in order to obtain an “optimal” design for such systems. The design of such vehicles is the result of a complex synthesis, requiring high degree of methodologies involving tradeoffs among multiple criteria. This paper proposes a probabilistic multicriteria decision analysis applied to the design of a hybrid electric vehicle. In particular, a Bayesian estimation approach is proposed and numerically illustrated, with the purpose of determining the optimal decision for the systems’ design under uncertainty. In the course of the numerical application, it is shown how such methodology is able to integrate and update available information, even if poor and uncertain, on the basis of new experimental data.