New Evaluation Methods for Conceptual Design Selection Using Computational Intelligence Techniques
Hong-Zhong Huang,Yu Liu,Yanfeng Li,Lihua Xue,Zhonglai Wang
DOI: https://doi.org/10.1007/s12206-013-0123-x
IF: 1.81
2013-01-01
Journal of Mechanical Science and Technology
Abstract:The conceptual design selection, which aims at choosing the best or most desirable design scheme among several candidates for the subsequent detailed design stage, oftentimes requires a set of tools to conduct design evaluation. Using computational intelligence techniques, such as fuzzy logic, neural network, genetic algorithm, and physical programming, several design evaluation methods are put forth in this paper to realize the conceptual design selection under different scenarios. Depending on whether an evaluation criterion can be quantified or not, the linear physical programming (LPP) model and the RAOGA-based fuzzy neural network (FNN) model can be utilized to evaluate design alternatives in conceptual design stage. Furthermore, on the basis of Vanegas and Labib’s work, a multi-level conceptual design evaluation model based on the new fuzzy weighted average (NFWA) and the fuzzy compromise decision-making method is developed to solve the design evaluation problem consisting of many hierarchical criteria. The effectiveness of the proposed methods is demonstrated via several illustrative examples.