Mathematical models have been developed to simulate influenza epidemics to help public health officials evaluate different control policies. In these models, often severe influenza epidemics with a considerable mortality rate are considered. However, as was the case for the 2009 H1N1 pandemic, some of the influenza epidemics are mild with insignificant mortality rates. In the case of a mild epidemic, the cost of different control policies becomes an important decision factor in addition to disease-related outcomes such as the attack rate.We develop a continuous-time simulation model for the spread of a mild influenza epidemic based on the SEIR model (an epidemiological model with four classes: susceptible, exposed, infective, and recovered) which includes different interventions. To determine the epidemic mitigation policy with the minimum cost, we also develop an optimization model with two decision variables, vaccination and self-isolation fractions, and an upper-bound constraint for the attack rate. We use this model to evaluate the cost-effectiveness of different mitigation policies. Furthermore, we integrate the simulation and optimization models to identify the optimal mitigation policy. Finally, we conduct sensitivity analysis on the key input parameters to ensure robust results.The optimal policy depends on the target population and, as our results show, in general is a combination of vaccination and self-isolation. Further, for low (high) levels of intervention, vaccination (self-isolation) is incrementally more cost-effective. Therefore, public health officials should concentrate on vaccination at the beginning of the epidemic. However, if the epidemic continues to spread, they should promote self-isolation as a more effective intervention.