Using stochastic directed graphs, a social network stochastic directed graph model allows for policy analysis and action. An activity generator may be used for creating agents that represent a population stratum. Agents may be proportionally selected to the size of the population stratum and representative activities that are associated with said population stratum. Agents have one or more conditional probabilities attached to the activities, which indicate the likelihood of interaction between agents and one or more agents or actors. Outcomes for the interactions may be accumulated. Based on these outcomes, which include benign and acute, a multinomial probability distribution may be estimated.