Online conflict resolution: algorithm design and analysis (Q6084129)

This paper studies algorithm design and analysis of online conflict resolution. A probability \(p_a(u)\) is taken to measure the likelihood of an attacker \(u\) begin persuaded to stop the conflicts by its close friends. This probability follows the Bernoulli process in which u is persuaded by each close friend one by one with the same probability. The paper formulates the online conflicts to a discrete optimization of choosing seeds to spread mediation influence over social networks to minimize the attack to all sufferers. The problem is NP-hard and the objective is \(\#\)P-hard to compute. An estimation method is proposed for the objective function employing the reverse sampling technique. Some experiments are conducted using real-world datasets.