Efficient Searching by Bias and Raising Threshold Algorithm Using Multiple Voting in the Best-of-n Problem

摘要

The Bias and Raising Threshold (BRT) algorithm is one of the methods for the best-of-n problem (BSTn) that allows a group of robots to find out the appropriate collective option among a set of n alternatives. This paper improves the BRT algorithm by using multiple voting for shortening the search time. Concretely, each robot is considered that might be able to vote multiple times in a selection. The experimental results revealed that the search time was only dramatically reduced but also the search accuracy was improved, especially in difficult problems where there are a large number of options (