On the worst case complexity of potential reduction algorithms for linear programming

Abstract

There are several classes of interior point algorithms that solve linear programming problems in O(x/ffL) iterations. Among them, several potential reduction algorithms combine both theoretical (o(vr~L) iterations) and practical efficiency as they allow the flexibility of line searches in the potential function, and thus can lead to practical implementations. It is a significant open question whether interior point algorithms can lead to better complexity bounds. In the present paper we give some negative answers to this question for the class of potential reduction algorithms. We show that, even if we allow line searches in the potential function, and even for problems that have network structure, the bound O(v/'ffL) is tight for several potential reduction algorithms, i.e., there is a class of examples with network structure, in which the algorithms need at least ll(v/ffL) iterations to find an optimal solution. (~) 1997 The Mathematical Programming Society, Inc. Published by Elsevier Science B.V.

On the safe side of stochastic programming: bounds and approximations

“…These shapes follow from the parametric shape of LPs (Bertsimas and Tsitsiklis, 1997). Notice that bounds based on convexity are generally not applicable to mixed-integer linear programs .…”

Energy exchange among base stations in a Cellular Network through the Smart Grid

“…At CBS i, the constraints imposed by its battery involve only decision variables associated with CBS i. Therefore, (5) can be solved by solving m independent optimization problems at 2 Although LPs can be solved in polynomial time [13], the computational complexity of the proposed strategy grows with the number of CBSs in the set. As a reference, each CBS introduces 6n inequalities, where n is the number of time slots considered in the optimization.…”

Beam controlled arc therapy–a delivery concept for stationary targets

“…The LP was formulated in the general algebraic modeling system (GAMS Development Corporation, Washington, DC), a high-level modeling system for mathematical programming and optimization and solved with an integrated high-performance commercial solver CPLEX (ILOG Inc., Sunnyvale, CA). The advantage of using the LP approach is that if the problem is feasible and bounded, which is the case for the dose optimization problem, formulating and solving it using LP guarantees the optimal solution (Bertsimas and Tsitsiklis 1997). Besides LP, other approaches such as quadratic programming can also be applied.…”