Utilization of Multi Attribute Decision Making Techniques to Integrate Automatic and Manual Ranking of Options

An information fusion system with local sensors sometimes requires the capability to represent the temporal changes of uncertain sensory information in dynamic and uncertain situation to access to a hypothesis node which cannot be observed directly. One of the central issue and challenging problem is the decision of what combination and order of sensors allocation should be selected between sensors, in order to maximize the global gain in the flow of information, when the data association is limited. In this area, Bayesian Networks (BNs) can constitute a coherent fusion structure and introduce different options (the combination of sensors allocation) for achieving to the hypothesis node through a number of intermediate nodes that are interrelated by cause and effect. BNs can rank the options in terms of their probabilities from Bayes’ theorem calculation. But, decision making based on probabilities and numerical representations might not be appropriate. Thus, re-ranking the set of options based on multiple criteria such as those of multi-criteria decision aid (MCDA) should be ideally considered. Re-ranking and selecting the appropriate options are considered as a multi-attribute decision making (MADM) problem by user interaction as semi-automatically decision support. In this paper, Multi Attribute Decision Making (MADM) techniques as TOPSIS, SAW, and Mixed (Rank Average) for decision-making as well as AHP and Entropy for obtaining the weights of attributes have been used. Since MADM techniques give most probably different results according to different approaches and assumptions in the same problem, statistical analysis done on them. According to the results, the correlation between compared techniques for re-ranking BN options is strong and positive because of the close proximity of weights suggested by AHP and Entropy. Mixed method as compared to TOPSIS and SAW is the preferred technique when there is no historical (real) decision-making case; moreover, AHP is more acceptable than Entropy for weighting.