Selecting judgment aggregation rules for NAO robots: an experimental approach

Social choice rules can be used to reach group decisions in multiagent systems. We consider judgment aggregation, the problem of aggregating answers to binary logically related questions. In general "fairness" is usually considered to be the main concern when selecting a social choice rule, however we believe that in judgment aggregation often a more relevant property is how efficient the rule is in truth tracking, that is, how often does it return the correct answer to the binary questions. Whereas "fairness" can be studied axiomatically, truth tracking efficiency needs to be studied experimentally. We accomplish the experimental analysis by constructing a multi-robot system.