Recent improvements in dynamic reservoir modeling have led to an increase in the application of model-based optimization of hydrocarbon bearing reservoirs. Numerous studies and articles have indicated the possibility of improving reservoir management using these dynamic models, coupled with methods to reduce uncertainties in the static models, to optimize reservoir performance. These studies have focused on maximizing the life-cycle performance of the project. Thus life cycle optimization is essentially a single-objective optimization problem. In reality, short-term targets usually drive operational decisions. The impact of short-term targets should be included in the optimization to achieve a more realistic solution. The process of optimizing these short-term targets constrained to life cycle targets is a form of multi-objective optimization. Several methods have been suggested to achieve multi-objective reservoir flooding optimization (Van Essen et al. 2011). These methods have been implemented with the adjoint formulation. This thesis proposes the use of an ensemble-based optimization technique (EnOpt) for multi-objective optimization. The optimization of smart wells or production schedules (inflow control valve (ICV) settings) is the objective of this work. We also propose variations to the existing multi-objective algorithms suggested by Van Essen et al. (2011). We propose the use of the BFGS algorithm to improve the computational efficiency. Undiscounted Net Present Value (NPV) and highly discounted NPV are the long-term and short-term objective functions used in this thesis. We also propose an extension of the optimization functionality to better cope with model uncertainties. This robust ensemble-based multi-objective production optimization framework has been applied and tested on a synthetic reservoir model. In our test cases, the ensemble-based multi-objective optimization methods achieved a 14.2% increase in the secondary objective at the cost of only a minor decrease between 0.2-0.5% in the primary objective.