Offered under: 6.7200, 15.093, IDS.200
Term(s): Fall
Level: Graduate
Units: 12
Prerequisite: Foundations in linear algebra (e.g., 18.C06), basic competency in computational programming, and mathematical maturity
Instructors: Dimitris Bertsimas (Sloan), Alexandre Jacquillat (Sloan)

Optimization methods are central to prescriptive analytics across operations research, management science and engineering. Thanks to astonishing progress in optimization algorithms and computing hardware, large-scale problems can now be routinely solved in practice, enabling new decision support systems manufacturing, transportation, energy, finance, healthcare, machine learning, etc. This course covers the methods at the core of linear optimization, non-linear optimization, integer optimization, and optimization under uncertainty, as well as their applications to management, science, and engineering.
Go to catalog listing

  • Modeling: techniques to formulate practical decision-making problems into well-structured mathematical models that can be solved with modern algorithms.
  • Theory: core principles to understand the mathematical and geometrical structure of an optimization formulation, and its implications in terms of scalability.
  • Algorithms: technologies at the core of modern optimization solvers; design and implementation of algorithms for optimization problems that cannot be solved with off-the-shelf solvers.
  • Practice: how to model complex problems, to build decision support tools, and to analyze results toward insights and recommendations for management, science and engineering

Prepares students for 15.083 Integer Optimization, 15.094[J]/1.142[J] Robust Modeling, Optimization, and Computation, 15.095 Machine Learning Under a Modern Optimization Lens, 15.084[J]/6.7220[J] Nonlinear Optimization.

Satisfies requirements in master’s programs in AeroAstro, Brain and Cognitive Sciences, Business Analytics, Chemical Engineering, Chemistry, Civil and Environmental Engineering, EECS, Finance, IDSS, Mechanical Engineering, Operations Research, Transportation, and the EECS PhD program. Undergraduates majoring in AI+D (Course 6-4) may use the class to satisfy a decision-centric or computation-centric elective.