Daniel Farenzena

Boolean Satisfiability Problem + Sudoku + Protests in Social Networks: three distinct social experiments to construct an improved human behavior model for Social Computing


We present a qualitative study about 3 distinct experiments with online social networks that we designed, performed and analysed during the past 4 years. The first two experiments challenged individuals to solve instances of the Boolean Satisfiability Problem and Sudoku game while allowing cooperation among peers. Preliminary analysis showed a cooperative behavior among peers that hinted on how humans take decisions individually in certain conditions. A third experiment was devised to verify the decision making properties found on the previous experiments and to construct an improved model for Algorithmic Game Theory studies. Interestingly, the third experiment was modeled as a cultural truth problem about the recent bus fares hikes in Brazil, as opposed to an objective truth problem (such as 3-SAT or Sudoku solving), indicating that our results hold for problems that only humans are able to currently execute, such as political decision making.