Abstract
The growing complexity of technical solutions, which encompass knowledge from different scientific fields, makes necessary, also for multi-disciplinary working teams, the consultation of information sources. Indeed, tacit knowledge is essential, but often not sufficient to achieve a proficient problem solving process. Besides, the most comprehensive tool of the TRIZ body of knowledge, i.e. ARIZ, requires, more or less explicitly, the retrieval of new knowledge in order to entirely exploit its potential to drive towards valuable solutions.
A multitude of contributions from the literature support various common tasks encountered when using TRIZ and requiring additional information; most of them hold the objective of speeding up the generation of inventive solutions thanks to the capabilities of text mining techniques. Nevertheless, no global study has been conducted to fully disclose the effective knowledge requirements of ARIZ. With respect to this deficiency, the present paper illustrates an analysis of the algorithm with the specific objective of identifying the different types of information needs that can be satisfied by patents. The results of the investigation lay bare the most significant gaps of the research in the field. Further on, an initial proposal is advanced to structure the retrieval of relevant information from patent sources currently not supported by existing methodologies and software applications, so as to exploit the vast amount of technical knowledge contained in there. An illustrative experiment sheds light on the relevance of control parameters as input terms for the definition of search queries aimed at retrieving patents sharing the same physical contradiction of the problem to be treated.