Multilevel optimization for enterprise driven decision-based product design

Research in Decision-Based Design (DBB) has traditionally modeled the interaction between enterprise product planning and engineering product development in a single optimization problem and solved it using an All-in-One (AIO) approach. Such an approach is not practical in a typical industry design environment, due to the organizational and computational complexities involved. In this paper, we present a hierarchical multilevel optimization approach, based on the principles of DBD and the concept of analytical target cascading (ATC), to integrate enterprise-level product planning with engineering-level product development. A disaggregate probabilistic demand modeling approach based on the Discrete Choice Analysis (DCA) is employed to capture the impact of engineering design on the choices of individual customers and to serve as the link between engineering product development and enterprise-planning. A search algorithm that coordinates the enterprise-level product planning and the engineering-level product development in a multilevel optimization solution process is presented. Such an algorithm systematically explores the engineering attribute space that may consist of a number of disconnected feasible domains due to engineering constraints. A case study that involves the design of an automotive suspension system is used to illustrate the effectiveness of the proposed approach.