Stochastic Methods  

Stochastic methods are instrumental in industrial and systems engineering  due to the uncertainties inherent to real-world applications, and to advances in high-performance computing that have made stochastic methods increasingly practical and powerful. Frequent domains of application include cyber-physical systems, energy systems, supply chains, healthcare, manufacturing, revenue management, finance, machine learning, and marketing.

Stochastic methods are used to analyze processes and design systems involving randomness. Examples include scenario-tree methods for decision analysis and stochastic optimization, option pricing methods in financial engineering, simulation techniques for system analysis and reliability assessment, uncertainty quantification methods for engineering design, as well as risk-aware and robust optimization techniques in operations research. Stochastic methods also provide powerful computational tools where randomness is artificially introduced to approximate quantities that are too complicated to evaluate otherwise. Examples include Monte Carlo methods to evaluate multivariate integrals, stochastic gradient methods to optimize over large datasets, and probabilistic methods to approximate hard combinatorial optimization problems.

At Lehigh ISE, we tackle the challenges of real-world operations research and investigate a broad range of stochastic methods relevant to  optimization and decision-making under uncertainty. Our contributions span areas such as stochastic processes, queueing systems, sample average approximation, optimal quantization, sequential analysis, uncertainty quantification, stochastic gradient methods, and stochastic optimization

Lehigh's research groups / institutes:
Illustrative publications by Lehigh's ISE faculty: