An Interior Point Method for Solving a Class of Linear-Quadratic Stochastic Programming Problems

J. Sun, K.-E. Wee, J.-S. Zhu

The quadratically convergent polynomial algorithm of Ye and Anstreicher is suggested for solving a class of two-stage stochastic programs in which both the present cost function and the recourse problem are linear- quadratic. Such stochastic programs, although are nonsmooth in nature, can be reduced to a linear com- plementary problem with a special structure. The proposed algorithm takes advantage of this structure and performs well in computational tests.

Recent Advances in Nonsmooth Optimization Editors D.-Z. Du, Liqun Qi, R.S. Womersley, (World Scientific, Singapore, 1995) 392-404.