The system design process is intractable, but robust.

Abstract

The system design process is intractable, but robust. The process is intractable because it is proven to be NP-complete. It is robust, because the tools used to do the system design are robust and allow for very good, but not optimal, solutions. The system design process is proven to be NP-complete by reduction of the Knapsack problem. The solutions used to solve the NP-complete problem called the Traveling Salesman Problem, will be examined with respect to how real designs are done. The design methodology is analyzed by a series of seventeen actual case studies. The tools evaluated are Quality Function Deployment and Design of Experiments. A sensitivity analysis of Quality Function Deployment is performed. It is proven to be insensitive to changes in weights and relationships. The needs and uses of modeling and prototypes in the system design process are examined. An actual prototype is analyzed using Design of Experiments. The Design of Experiments technique is shown to be very useful during the test and integration activity of the system design process.