Quadratic Parametric Programming for Portfolio Selection with Random Problem Generation and Computational Experience

For researchers intending to investigate midto large-scale portfolio selection, good, inexpensive and understandable quadratic parametric programming software, capable of computing the efficient frontiers of problems with up to two thousand securities without simplifications to the covariance matrix, is hardly known to be available anywhere. As an alternative to Markowitz’s critical line method, a full explication of a simplexbased quadratic parametric programming procedure, utilizing well-known components, is coded in Java for public domain use on modern desktops and laptops. The advantage of the different design is to provide an algorithm that can ultimately be extended (not a part of this paper) to portfolio problems with objectives beyond mean and variance. Using the code, aspects of portfolio selection problems are investigated and computational experience is reported.