Portfolio optimization as a learning platform for control education and research

This paper demonstrates the use of discrete time portfolio optimization as a mechanism for introducing students to key problems in systems theory: control, system identification, model reduction, and verification. Too often students are not introduced to systems theory until very late in their programs, frequently after they have already decided on majors and generated momentum toward specific career plans. One reason for this late introduction is the prerequisite material demanded by our systems courses, typically involving a chain of math, physics, and engineering courses. Using portfolio optimization as the vehicle for introducing systems theory, however, can provide an early introduction to some of the central issues in the field. In particular, the open-loop nature of portfolio optimization simplifies the decision-making context sufficiently to make these problems accessible to younger students. Moreover, the familiarity of financial decision making, regardless of technical background, allows a broad range of students to appreciate the importance and nature of these problems. Here we illustrate these ideas, using portfolio optimization to show how the presence of uncertainty and complexity in decision problems interconnect control, system identification, model reduction, and verification in the design of practical decision systems.