Applying lean principles and set-based approaches in product development

Doctoral thesis, 2022

The research described in this thesis addresses the problem of transformation to lean product development (LPD) and how to introduce and support the use of set-based design (SBD) in the concept development process. The original description of SBD does not define how to generate, evaluate and reduce a set of design solutions. Evaluation of solution candidates, which are too complex to be analytically verified, or are driven by qualitative criteria, has here been given special attention, particularly in cases when methods utilising human judgment may be needed. For some products, the solution space can consist of both principally different alternatives and parameterised variants of these. The question here is if established methods can be combined and introduced in an efficient way to support an SBD process for development of such products, when driven by both quantitative and qualitative criteria.

The research approaches used are:

-          a two-case study (Yin, 2009),

-          the design research methodology (Blessing and Chakrabarti, 2009), and

-          the scientific work paradigm (Jørgensen, 1992),

the last two combined with multiple case studies. Also, elements of action research (Oosthuizen, 2002) are used.

The results show that the principles and introduction of LPD were experienced as positive by participating practitioners in the conducted case studies. It was furthermore shown that SBD can be introduced and applied in a workshop at team level within a time frame of one or two working days if the design problem at hand is not too complex. Another result is that SBD can be combined with and supported by established methods such as creative and systematic methods for synthesis, enhanced function-means modelling, axiomatic design, extended causal diagrammes, interactive evolutionary algorithms (IEA) and Pugh matrices for generation, analysis, evaluation and reduction of a solution space of design alternatives and variants of these. Both qualitative and quantitative requirements can be handled.

The conclusions are that a transformation to LPD is facilitated by information about good examples and internal support by management. Also, the existence of a lean enthusiast in the organization and an appropriate implementation plan supports a transformation to LPD. A function to maintain the LPD system as well as influence of the lean principles are valuable guides on how to use LPD. 

Also concluded is that a seamless, efficient process, applying set-based principles, for synthesis, evaluation, and reduction of a solution space of design alternatives can be created by combining enhanced function-means modelling, morphological matrices, axiomatic design, causal diagrammes and Pugh matrices. Such a compound of methods can be introduced and applied in a workshop at team level within a time frame of one to two days when solving well-known and not too complex design problems. The workshop should be facilitated by an expert on the methods used and initiated and surveyed by a team manager. Furthermore, a solution space of parameterised design variants, with criteria that are either qualitative or too complicated to be numerically defined, can be generated, evaluated and reduced in such a process. By using a defined set of functional and constraining criteria, and applying axiomatic design and IEA, a variant solution space can be generated and refined. A set-up of the IEA that does not overburden the user should be preferred.