TY - JOUR TI - Numerical methods for probabilistic constrained optimization problem where random variables have degenerate continuous distribution DO - https://doi.org/doi:10.7282/T35D8V1M PY - 2016 AB - Several probabilistic constrained problems (single commodity stochastic network design problem and water reservoir problem) are formulated and solved by use of different numerical methods. The distribution considered are degenerate normal and uniform distributions. The network design problem is to find optimal node and arc capacities under some deterministic and probabilistic constraints that ensure the satisfiability of all demands on a given probability level. The large number of feasibility inequalities is reduced to a much smaller number of them and an equivalent reformulation takes us to a specially structured semi-infinite LP. This, in turn, is solved by a combination of inner and outer algorithms providing us with both lower and upper bounds for the optimum at each iteration. The flood control and serially linked reservoir network design with consecutive k-out-of-n type reliability problems are formulated, simplified and solved. Alternative, derivative-free methods, are proposed and implemented. Various numerical examples are presented and solution methods software library is developed. KW - Operations Research KW - Stochastic processes KW - Mathematical optimization LA - eng ER -