Multi-Objective Design Indicators: A Quantitative Paradigm for Early Engineering Design of Deeply Integrated Products

The functional integration of products is increasing. This is the result of the unprecedented pace of the technological evolution and the demands from a diverse global market. Regardless of the reasons, it has been shown that the increasing integration bears a significant risk of increasing the lead time of new product launches. Since the products are simultaneously designed towards multiple functional objectives, the unavoidable dependencies between their components increase the number of needed iterations in the development process, because these multi-objective and constraint dependencies restrict the solution space. Small solution spaces are particularly critical when products are transferred from development to full production, where variation is inescapable and both expensive and hard to control. The underlying reasons for the need for more iterations in the development of deeply integrated products are both technical and cognitive. In deeply integrated products the design of the individual components is dependent, which can lead to restricted solution spaces. The result may be suboptimal performance of the single functions and higher sensitivity towards variation in production and use. Therefore, design of deeply integrated products is a challenge of modern manufacturing companies.

In the research field of Engineering Design, the literature often categorise design methods as either focusing on conceptualisation or engineering analysis for detailed designs. While conceptualisation is the creative synthesis of product concepts, engineering analysis focus on quantitative analysis of detailed designs. Due to the multi-objective dependencies, which are created in the early embodiment of design concepts, the development of deeply integrated products is challenging to address with existing methods. Even Robust Design methods, which support the reduction of sensitivity towards variation in production and use to ensure consistent product performance, do not currently address the selection of early design alternatives based on the properties of internal dependencies. Consequently, these dependencies may often not be identified before late development stages where detailed quantitative analyses can be used for exploration of the solution space one point at a time.

In this research project, the industrial challenges as well as the gap between methods for conceptualisation and detailed engineering analysis are addressed. Therefore, the aims of the research was to develop a method to support: (1) The assessment and comparison of multiple design alternatives, (2) the exploration of multi-objective solution spaces, and (3) the reduction of the effect of biases in the selection between alternative designs. These aims are addressed through a literature review, an interview study, a prescriptive study for method development, and an initial evaluation of the method based on a quasi-experiment.

The core contributions of the research project are: (i) Empirical evidence that the challenge of developing deeply integrated products are generalisable across different industrial sectors, (ii) the development of the Multi-Objective Design Indicators based on mathematical modelling and optimisation theory, and (iii) qualitative examples of engineering designers’ approach to selection and optimisation of three alternative designs as well as indication that visualisation of Objective Spaces may support the process.

The Multi-Objective Design Indicators support the assessment and comparison of multi-objective solution spaces based on achievable objective performance, trade-off severity, objective sensitivity towards design variables, and flexibility in the Design Space. Their theoretical validity is demonstrated based on two case studies, while the empirical validity is a topic of future work based on industrial application. The development of the indicators address the three aims listed above and represents a quantitative paradigm for early engineering design of deeply integrated products. The proposed method promise a potential for supporting early selection of design alternative to reduce late-stage iterations and product lead time.