Search

We develop a class of lower bounds on the expectation of a convex function. The bounds utilize the first two moments of the underlying random variable, whose support is contained in a bounded interval or hyper-rectangle. Our bounds have applications to stochastic programs whose random parameters are known only through limited moment information. Computational results are presented for two-stage stochastic linear programs.

The success of multiskilling greatly relies on the foreman's ability to assign workers to appropriate tasks and compose crews effectively. The foreman assigns tasks to workers according to their knowledge, capabilities and experience on former projects (Rodriguez 1998). To obtain the maximum benefits from multiskilling, practitioners have stated that automated tools that optimize allocation and assignment of workers would be valuable (Rodriguez 1998). This research describes efforts directed toward developing tools that optimize the allocation and assignment of multiskilled workers to activities in a construction project.

Operations Research, 47, pp. 943-956. We consider the problem of bounding the expected value of a linear program (LP) containing random coeffecients, with applications to solving two-stage stochastic programs. An upper bound for minimizations is derived from a restriction of an equivalent, penalty-based formulation of the primal stochastic LP, and a lower bound is obtained from a restriction of a reformulation of the dual. Our "restricted recourse bounds" are more general and more easily computed than most other bounds because random coefficients may appear anywhere in the LP, neither independence nor boundedness of the coefficients is needed, and the bound is computed by solving a single LP or nonlinear program. Analytical examples demonstrate that the new bounds can be stronger than complementary Jensen bounds. (An upper bound is "complementary" to a lower bound, and vice versa). In computational work, we apply the bounds to a two-stage stochastic program for semiconductor manufacturing with uncertain demand and production rates.

We describe a multi-period optimization model, implemented in GAMS, to help the U.S. Air Force improve logistical efficiency. It determines the maximum on-time throughput of cargo and passengers that can be transported within a given aircraft fleet over a given network, subject to appropriate physical and policy constraints. The model can be used to help answer questions about selecting airlift assets and about investing or divesting in airfield infrastructure.