| Under the increasing pressure of the evolving customers' expectations, the speed and competitiveness of the competitors, automakers have become customer-oriented. They continuously survey the customers' needs in order to early identify the desired or utopian vehicle performances and strive to fulfill these expectations by designing and marketing quickly new innovative products. The development of a new vehicle supposes the translation of the vehicle performances into its components' characteristics. Such approach requires making critical design decisions that can impact noticeably the competitiveness and profitability of the company.;In the early stages of the vehicle development process, the engineers lack precise and complete information about the possibility to meet the initial utopian vehicle performances due to many factors (technological, regulation, resources, etc.). For that reason, identifying, quantifying and handling the inherent uncertainty throughout the vehicle development process (V D P) became a serious issue, which affects the effectiveness of the design process.;This study proposes a methodology for target allocation and decision-making under uncertainty during the VDP. The method starts by the decomposition of the vehicle in hierarchical multilevel structure, which represents the basic framework required for the definition of the vehicle multilevel model (VMM). We have considered that each component in the VMM may have several characteristics, and that a target is defined for every component and characteristic in accordance with the utopian vehicle performances. Experts' opinions are expressed with uncertainty regarding the feasibility of achieving each target. Experts' opinions are given in the form of probability distributions or intervals associated with their subjective beliefs for the possible values of the characteristics and then are aggregated and propagated from the leaf nodes of the multilevel model up to the vehicle level. Evidence theory has been used to express uncertainty in the form of belief and plausibility measures. Using this information, two measures regarding the desirability and the achievability of the characteristics are defined. An approach for targets allocation under uncertainty based on the maximization of achievability and desirability measures of the characteristics is proposed and discussed.;A methodology to handle large-scale problem based on the merging of intervals by the control of the information granularity without affecting the precision of the belief and plausibility measures is presented.;A decision-making framework based on the integration of both exposed techniques for uncertainty characterization and target allocation under uncertainty in the vehicle multilevel model is proposed. This framework consists in modeling the design process in the form of a series of parallel stage-gate processes that alternate knowledge generation and decision-making. Iteratively the characteristics are set and refined in such a way to orient the design process towards an achievable and desirable final design.;In brief, including uncertainty in target allocation and decision-making processes presents many potential benefits to the automakers. In fact, in the case of vehicle design under certainty, the characteristics of the vehicle and its components are supposed to be known with certainty and the design process is restricted to only a choice among few existing alternatives. Such approach may yield unreliable designs that are pushed to the limits of design constraints boundaries. Consequently the design team may miss the opportunity: to improve the design to a desired level because of lack of exploration of all offered possibilities, to satisfy the customers' expectations and to meet the company's goals. In return considering uncertainty in targets allocation and decision-making processes implies exploring new possibilities, collecting more information and developing new knowledge that leads to new concepts, new designs and technologies allowing at best the fulfillment of the customers' needs and the attainment of the company's objectives.;In conclusion, the present work represents a step in the formulation of an integrated methodology to take into account uncertainty during the early stages of the vehicle development process. The proposed methodologies and the approaches for uncertainty management, target allocation under uncertainty and decision-making under uncertainty were developed for the design of complexes systems. Since, the VDP is only a specific application of the proposed systems. This later can be directly applied to any engineering field concerned by the development of complex system such as aeronautical, aerospace and naval industries. |
