A complex engineering system such as a xerographic marking engine is an aggregate of interacting subsystems that are coupled through a large number of constraints and design variables. The traditional way of designing these systems is to decouple the overall design into smaller subsystems and assign teams to work on these subsystems. This approach is critical to making the project manageable and enabling concurrent development. However, if the goal is to design systems that can deliver best possible performance, i.e. if the performance limits are being pushed to the extreme, characterizing the interactions becomes critical.
Multiobjective optimization is a design methodology that addresses the issue of designing large systems where the goal is to simultaneously optimize a finite number of performance criteria that come from one or more disciplines and are coupled through a set of design variables and constraints. This approach to design makes explicit and quantitative the inherent trade-offs that need to be made in doing coupled system design. It also enables the determination of the attainable limits of performance from a given system.
This paper will discuss the multiobjective optimization methodology and optimal methods of performing quantitative trade-off analysis. These design methods will be applied to problems from the xerographic design domain and results will be presented.