Modeling Social Media Effect on Civil Revolutions
state , agent interactions , and environment . When its grievance level is higher than its threshold and assuming its grievance level exceeds its net risk perception , the agent wants to move and protest ( Lemos , Coelho , & Lopes , 2013 ).
After Epstein ’ s main contribution to the field , there are some variations of his model . Lemos et al . ( 2013 ) has a wide survey paper on civil revolutions models . Therefore , we would like to cite their work in order to show all of the former papers on civil revolts until 2012 ( Lemos et al ., 2013 ).
Here , three civil revolutions papers , published after 2012 , are introduced .
Comer and Loerch ( 2013 ) suggest a new way for modeling agent ’ s actions . In their model , agents are not moving uniformly as they were in Epstein ’ s model . For Comer and Loerch ( 2013 ) model , agents move either asynchronously or randomly . For the asynchronous , agents ’ neighbors influence agents ’ attributes . At first , an agent moves to a state and checks its conditions by counting the numbers of active agents and cops in its vision radius . Then , according to its neighborhood , it looks at its own threshold and decides whether to act . This action is dependent on the agent ’ s neighborhood . For the random , an agent makes decisions on whether or not its next action is active . This paper opens an additional discussion for our research : if we have a social media effect on our agents , then there will be no constraint on their visions ; therefore , we can state that the social media effect will enlarge the agent ’ s vision .
New types of agents and their roles listed correspondingly by Lemos , Lopes , and Coelho ( 2016 ). Here , the number of previously active protesters who are in jail affects an agent ’ s grievance level . There is a feedback process in this model . If the number of jailed active agents increases , then government legitimacy decreases , and the agent ’ s grievance level subsequently increases . For Lemos et al . ( 2016 ), the point is more realistic than Epstein ’ s model . Instead of having a random distribution for grievance levels of the agents , they suggest an internal dynamic for an agent ’ s activations process . We would like to add this function into our model with the help of social media .
Compared with others , Moro ( 2016 ) has the greatest contribution to the literature with his model , as it can generate real world examples of civil revolutions like the Arab Spring ( Moro , 2016 ). His model predicted several rebellions that arose in similar ways with different scenarios . Outcomes of his model show different political conditions , such as the successful revolution in Tunisia , the failed protests in Saudi Arabia and Bahrain , and civil war in Syria and Libya . His model can create similar dynamics to these real civil revolutions .
In his model , there are three types of agents to characterize citizens : active ( A ), jailed ( J ), and revolutionary ( R ). In addition , there are police officers ( P ) who fight with the active citizens , who turn into active revolutionary citizens if they intend to kill police officers in their vision radius . Moro suggests three scenarios with different parameters for the probability of killing police : successful revolution , anarchy , and failed revolution .
When a revolutionary citizen is active , he kills a randomly selected police officer in his vision radius with a probability equal to r . For the police officers , if the randomly selected agent is a citizen , he arrests him ; if he is a revolutionary citizen , he kills him with a probability equal to p ( Moro , 2016 ). We give Moro ’ s probabilities in Table 1 .
18