A Knowledge Analysis of an Engineering Application Related to Sustainable Development

For materials engineers, it is very important to select the best material to make an object. The criteria used for such selection may be classified in several categories, including the two following: criteria related to the durability of materials against degradations caused by environments and those expressing their impacts on the environment throughout their whole life cycle. Selecting a material is therefore a multi-criteria decision problem. Due to the diversity of criteria, material analysis is carried out by design offices using systems based on human knowledge when available and, otherwise, trial and error. Currently, the large amount of materials available and the number of properties make it necessary to investigate the use of generative AI for the selection of materials. However, the information provided by the available systems is insufficient to make the appropriate decision. Traditional operational research systems based on mathematical models, such as Formal Concept Analysis (FCA), provide a support for understanding the logic used to reach the solutions. They offer decision-makers a greater possibility of arguing their choice. Our work tackles this material selection decision problem using FCA. The results obtained after applying the FCA ConExp tool are analyzed in relation to the results obtained by the AI systems ChatGPT and the Fictiv-AI material selection assistant.