A model of inexact reasoning in mechanical design evaluation

Abstract This paper presents an inexact reasoning model based on the union of extended D-S theory and fuzzy theory, which is applied to evaluate mechanical designs. D-S theory is the main reasoning tool. To take account of the evaluation factors and their relationships, we adopt the inexact reasoning model (extended D-S theory) for top level evaluation, and multilevel fuzzy synthetic evaluation for bottom level evaluation. Conclusions of the bottom level evaluation are taken for intermediate data. Evidence reasoning is executed from bottom to top, to implement the evaluation of system design. In fuzzy synthetic evaluation, the mode M (·, ⊕) and evaluation factors with tree hierarchical structure have been utilized. When this inexact reasoning model was applied to mechanical design evaluation, the time complexity was only O ( mn ). It is much smaller than the time complexity O (2 2| ∗ |) produced by D-S theory only.