A Computer Aided Approach for Reformulating “Ill-Defined” Problems

It is proven that a low accuracy in setup or an insufficient attention during the problem structuring [1] may affect the accuracy of the final solution [2], while a problem that is properly defined is virtually solved [3], especially for “ill-defined” problems. Many efforts have been made since the 1970s for improving problem assessment, managing problem information (their functioning, constraints and requirements) and avoiding psychological barriers and memory loads [4]. Other studies have been focused on overcoming trivial points of view, such as lateral thinking [5] and to systematize an abstraction path for the initial situation by using theoretical models as functional models [6], cause effect analysis, and contradictions [7]. This work takes into account this background in problem solving methods, especially root cause analysis [8] and TRIZ [7], the theory of inventive problem solving. In this paper, a set of steps, called BOB-UP® have been conceived in order to support designers during the correct reformulation of the initial problem. This procedure has the aim to reformulate every technical problems, turning an ill-defined initial problem into a well-defined final problem. In this paper the framework of the procedure based upon an inedited cause-effect analysis is widely shown. For completing each design step, ontologies, linguistic rules, modeling tools and sketches are carefully integrated. Novel and key points are extensively presented in this paper, together with an exemplary case about a toaster implementation. The overall procedure was translated into software to ease the completion. The output of such computer aided reformulation consists of a “well defined” problem reformulation. Unlike the most famous and traditional cause-effect approaches, in the BOB-UP® environment the user is guided step by step to identify the best level of detail, focusing on the undesired effect(s), and finally identify the crucial element on which intervention is needed. A preliminary validation, conducted with 9 industrial cases by 30 users selected from the teaching staff of the University, research assistants and PhD and Master Degree students in Mechanical Engineering, has demonstrated the effectiveness of the method.