Optimization with Uncertainty

Classical algorithms (exact/ approximation) work with an input which is entirely specified up front. While this offline model is useful for static optimization problems, there are several domains which need algorithms to make decisions with partial/uncertain information which evolves over time. We seek to design algorithms in such uncertain environments and also design frameworks to evaluate their quality using different metrics. Models such as online algorithms, stochastic optimization, and algorithms with recourse fall into this area of study.