Analytical Network Process in Decision Making

Abstract

Analytic Network Process (ANP) is a generalized form of the Analytical Hierarchical Process (AHP). AHP is an objective theory of deriving the decision makings on absolute scale values starting from subjective relative measurements perhaps on tangible or intangible or a mix of both rated by the experts. A consistency measure of the relative ratings is verified with a consistency ratio to assure the ratings of the experts are consistent. The clustering of criteria/sub-criteria with inner and outer dependencies as well as feedback is taken into account in ANP. In ANP rather than a hierarchical process, a network structure is assumed and all nodes in the network are treated equally with interaction between nodes and grouping into clusters. A super-matrix is constructed, then normalized it to get weighted super matrix and raised it to a sufficient power to converge this into a unique eigen vector representing the final weights in limit super-matrix. Step by step iterations of ANP was demonstrated with a selection of car example.