Towards Measure of Semantic Correlation between Messages in Multiagent System

In this work an original method for the measure of semantic correlation between messages generated by communicative agents in multiagent system is presented. This method is applied to the algorithm for finding the set of agents with the maximal cognitive capacity to observe the states of the objects from external world. It is assumed that the generated messages are the result of the process of language generation, represented by the algorithm for the choice of relevant semantic messages. This algorithm relates belief formulas to the internal agents’ knowledge states.