The article describes an algorithm for finding the most relevant element in a finite set of answers. The search is performed on a finite set of queries and with a given relevance function, which has a significantly lower computational complexity than exhaustive search. To achieve the performance goal we created a relevance function induced coordinate system on the set of answers and, subsequently, created a metrized small world graph on its basis. The validity of the proposed heuristic algorithm was confirmed by computer simulations.
|
