CONTINUOUS PROCESS IMPROVEMENT THROUGH DESIGNED EXPERIMENTS AND MULTIATTRIBUTE DESIRABILITY OPTIMIZATION

Process optimization is a problem with many dimensions. Attributes of interest compete with one another and are affected by a host of variab les. It is impossible to achieve the best possible values for all proc ess outputs simultaneously. For this reason, it is important to define what should be achieved from the process. Once the objectives are kno wn, statistically designed experiments can be used effectively to dete rmine the optimal levels of controllable process variables that will p roduce the desired result and make the process robust to variations in the influential parameters that cannot be controlled. This paper desc ribes an approach to establishing values for process variables to cons istently achieve the optimal set of process outputs. It is an iterativ e process that produces continuous improvement. Principles of statisti cal experimental design and multi-attribute desirability optimization methodology are employed. The benefits of this approach include better products, less variability, lower costs, and more efficient process d efinition.