Effective pandemic management requires a series of mitigating steps that change over time as function of external conditions and previous mitigating steps. Modeling this effectively requires a multi-disciplinary integrative approach that combines epidemiological, economic and social considerations within a unified modeling environment. The paper covered focuses on incorporating elements from classical utility theory in combination with control theory and machine learning to better model the dynamics of the non-linear trade-offs inherent in managing the pandemic. We postulate a theoretical formulation on how these trade-offs can be modeled, and demonstrate empirical results to elucidate the limiting factors in finding efficient solutions.