Decision making under uncertainty in nuclear engineering design

Abstract

Engineering design is a complex and ill defined procedure. Design is a highly heuristic problem solving activity which involves problem definition, solution or strategy selection (search among alternatives), trade off analysis, engineering analysis and optimization. Trade off analysis is inherent to the engineering design process. The constraints given to the design engineer as part of the problem definition are often expressed linguistically or in an imprecise manner. Classical decision or optimization techniques do not perform at their best in such a vague environment. In this work, a paradigm of the decision process under uncertainty has been developed. This paradigm uses Fuzzy Set Theory for representing vague and linguistic qualifiers. The paradigm developed was programmed into an interactive software with a graphical user interface. Software experiment against deterministic and linguistic cases are given. A comparative study of the performances of the formalized paradigm versus a panel study for selection of a compact space reactor among a set of proposed systems is performed. The software handled successfully two main uncertainties: linguistic variables and conflicting constraints. The paradigm developed allows for a formal tracking, representation and implementation of imprecision in the decision process.