Sequence Automata for Researching Consensus Levels

摘要

Having a group of ranking lists (ordered sequences) we can use different theories to research a scale of consistency according to the assessments contained in it. The agreement of ordered (queue) objects can be estimated as the maximal length of preference subsequences i.e. a maximal set of objects with the same queue of dominance relation. It is not important that among the chosen objects others exist but it is important that the i-th object is before j-th objects (has greater validation). The proposed sequence automata can be use to find different kinds of sequence agreements or disagreement among a sequenced group represented by for example by a ranking list. An exploitation ranking list has an essential advantage over other methods of object judgment (assessments) because different criteria or their sets are reduced only to positions on lists. In such a situation the problem of assessing the scale of consistency (agreement among authors or algorithms of ranking lists) remains. A unified-form of complex criteria presentation helps us to elaborate the tools for consistency estimation. We expect that sequence automata will resolve this and similar problems connected with sequence analysis. Our proposition is the tool supporting multi-agents system in areas of negotiation, partner recognition, creation hierarchy of criteria, planning strategy and finding hidden, imperceptible nuances (details), having dependence (or preference) character. It can be also use for coding and decoding information by sequence interlacement.