Increasing the Quality of Multi-step Consensus

Abstract

Determining the consensus of a collective is becoming a popular problem-solving method in our society. However, given that determining the consensus of large collectives is time-consuming, a multi-step consensus approach is necessary. Thus, one important problem is to determine the number of steps required to obtain a reliable consensus in an acceptable time. Execution time depends on the number of steps; determining the number of steps relies on the quality of the consensus in each step. The overall consensus quality depends on the problem of determining consensus in each step. Therefore, it is important to improve the consensus quality and investigate the quality according to the number of smaller collectives in each step. Herein, we improve the basic algorithm used for the multi-step consensus approach. The experiment result shows that the approach based on the improved algorithm is more efficient than that of the basic algorithm in terms of consensus quality (4.9%). Furthermore, the consensus quality was investigated according to the number of smaller collectives in each step.