Decision Analytics Lab develops intelligent decision-making framework for multifaceted systems originated from healthcare, manufacturing, energy and other enterprises to make reliable, resilient, and efficient decisions. Our research uses an extensive array of problem-solving methods such as data driven large-scale optimization under uncertainty handled partly by mixed integer stochastic programming, decision analysis by leveraging cutting edge data-mining, machine learning, and artificial intelligent techniques. These areas offer unique research potential in formulating new theories, finding challenging application areas, and their intersections. Our interdisciplinary research approach investigates how big data analytics can be used to assess healthcare policies and recommend intervention for improved healthcare service and patient outcome. We analyze several state and national hospital discharge data along with Massachusetts All Payer Claim Data set (MA APCD), which includes claims from private insurers, to detect prescribing patterns of opioids and identify the causes of hospital readmission. To analyze and make robust decisions for detecting treatment effects from large-scale observational studies, we develop robust causal inference framework to harness the potential of big data.
The gradually expanding application areas of our work include development of resilient strategies to combat the ravaging crisis of opioid addiction epidemic including risk assessment tools for physicians to predict the likelihood of drug dependency and incentive allocation model for prescription opioid users to deter the flow of excess Opioid to black market and secondary users; designing robust supply chain and resilient humanitarian logistic networks under the upshot of disruptions; energy smart production plan for an integrated manufacturing and remanufacturing systems operating under uncertainties; development of pragmatic decision making framework to handle inherent uncertainty and complexity of multi-criteria decision making process; designing machine learning assisted decision making framework for wind energy power plants. We assess the trade-offs in several parameters including cost, service quality, systems reliability, and resiliency to help decision makers best mirror their true preferences while deriving high impact societal, business, and engineering decisions.