Designing a single-configuration-for-all product for population accommodation on the basis of user preference data

Creating a consumer product that is capable of anthropometrically accommodating a large proportion of a target population is known to be a non-trivial problem as individuals vary significantly in their physical characteristics (Parkinson et al., 2007). The problem of how to realize population accommodation for consumer product design seems to have much to do with the level of product variety or the type of product an enterprise decides to offer. The enterprise may decide to custom-design for each individual customer (e.g., a custom-tailored suit), create a reconfigurable product (e.g., a height-adjustable chair) or produce a product available in multiple varieties (Ulrich, 2011). Alternatively, it may opt for providing a singleconfiguration product for everyone in the population with rigorous optimization for product configuration design. These alternatives would require different methods for realizing population accommodation. Among the alternatives mentioned above, creating a single-configuration-for-all product (hereafter, a single configuration product) seems to offer multiple advantages over the others. From the enterprise’s point of view, the main advantages would be the relative simplicity of design problem and manufacturing process, which can lead to lower design and manufacturing costs. For the consumers, they could benefit from lower product price; also, the consumers would find it advantageous to be able to use the product as is without having to make an effort to choose the right product variant or reconfigure the product. Despite the importance, however, the problem of designing a single configuration product for population accommodation has not been fully investigated. Currently, there are two methods that are being widely utilized for single configuration product design: designing for the extremes and designing for the average. The first method focuses on accommodating the individuals at the extremes of the population distribution with an assumption that doing so would ensure accommodating less extreme individuals. The second method, on the other hand, aims to accommodate the people around the medium, with an implicit assumption that it will result in a good solution in terms of population accommodation. Taking the design of a door as an example, the first method is typically used for the design of the doorpost height, yet not for the design of the doorknob height, and, vice versa, for the second method. Despite their wide use, however, these methods do not always guarantee high-level population accommodation and could result in low-quality solutions for many design problems. As an effort to address the problem of single configuration product design, this paper presents a novel design optimization method. This method utilizes empirically obtained human preference data for optimizing a product’s configuration; and, in doing so, both the intra-individual as well as inter-individual variability in human preference are considered. A similar approach has been presented by Park et al. (2012) for the problem of designing reconfigurable products. In what follows, the design method will be described using an example design problem.