Advanced Statistical Tools for Six Sigma and other Industrial Applications

Six Sigma is a methodological approach and philosophy for quality improvement in operations management; its main objectives are identifying and removing the causes of defects, and minimizing variability in manufacturing and business processes. To do so, Six Sigma combines managerial and statistical tools, with the creation of a dedicated organizational structure. In this doctoral thesis and the three years of study and research, we have had the purpose to advance the potential applications of the methodology and its tools; with a specific attention on issues and challenges that typically prevent the realization of the expected financial and operational gains that a company pursue in applying the Six Sigma approach. Small and medium sized enterprises (SMEs), for instance, very often incur into such issues, for structural and infrastructural constraints. The overall application of the methodology in SMEs was the focus of the initial research effort and it has been studied with a case study approach. Then, on this basis, most of our research has been turned to the rigorous methodological advancement of specific statistical tools for Six Sigma, and in a broader sense, for other industrial applications. Specifically, the core contribution of this doctoral thesis lies in the development of both managerial and/or statistical tools for the Six Sigma toolbox. Our work ranges from a decision making tool, which integrates a response latency measure with a well-known procedure for alternatives prioritization; to experimental design tools covering both planning and analysis strategies for screening experiments; to, finally, an initial effort to explore and develop a research agenda based on issues related to conjoint analysis and discrete choice experiments.