Modeling of an Agent Society with Probabilistic Approaches

In the problem of task allocation in multiagent systems, it is very desirable to be able to guarantee the error rate of a solution in an agent society. In this paper, a novel approach for this problem has been introduced by devising a mathematical model with probabilistic approaches. It is also shown that the multiagent system is able to achieve any desired and predetermined threshold of correctness for the final solution, regardless of the performance of selfish agents in the society or any stipulation about their honesty, which is extremely critical in the design of open and real-world multiagent systems.