Eugenio Mijangos

On the first-order estimation of multipliers from Kuhn-Tucker systems

Abstract The minimization of a nonlinear function with linear and nonlinear constraints and simple bounds can be performed by minimizing an augmented Lagrangian function that includes only the nonlinear constraints subject to the linear constraints and simple bounds. It is then necessary to estimate the multipliers of the nonlinear constraints and variable reduction techniques can be used to carry out the successive minimizations. The viability of estimating the multipliers of the nonlinear constraints from the Kuhn–Tucker system is analyzed and an acceptability test on the residual of the estimation is put forward. The computational performance of the procedure is compared with that of the inexpensive Hestenes–Powell multiplier update. Scope and purpose It is possible to minimize a nonlinear function with linear and nonlinear constraints and simple bounds through the successive minimization of an augmented Lagrangian function including only the nonlinear constraints subject to the linear constraints and simple bounds. This method is particularly interesting when the linear constraints are flow conservation equations, as there are efficient techniques for solving nonlinear network problems. Regarding the successive estimation of the multipliers of the nonlinear constraints there is some doubt as to whether using the Kuhn–Tucker system could improve upon the inexpensive Hestenes–Powell update, especially considering that the Kuhn–Tucker system with partial augmented Lagrangians may not always lead to an acceptable multiplier estimation. Clarifying the computational efficiency of the multiplier update when there are linear or nonlinear side constraints is also a necessary previous step regarding the comparison between partial augmented Lagrangian techniques and either primal partitioning techniques for linear side constraints or projected Lagrangian methods in the case of nonlinear side constraints.

An implementation of Newton-like methods on nonlinearly constrained networks

The minimization of a nonlinear function with linear and nonlinear constraints and simple bounds can be performed by minimizing an augmented Lagrangian function, including only the nonlinear constraints. This procedure is particularly interesting when the linear constraints are flow conservation equations, as there exist efficient techniques for solving nonlinear network problems. It is then necessary to estimate their multipliers, and variable reduction techniques can be used to carry out the successive minimizations. This work analyzes the possibility of estimating the multipliers of the nonlinear constraints using Newton-like methods. Also, an algorithm is designed to solve nonlinear network problems with nonlinear inequality side constraints, which combines first and superlinear-order multiplier methods. The computational performance of this method is compared with that of MINOS 5.5.

An efficient method for nonlinearly constrained networks

Many nonlinear network flow problems (in addition to the balance constraints in the nodes and capacity constraints on the arc flows) have nonlinear side constraints, which specify a flow relationship between several of the arcs in the network flow model. The short-term hydrothermal coordination of electric power generation is an example of this type. In this work we solve this kind of problem using an approach in which the efficiency of the well-known techniques for network flow can be preserved. It lies in relaxing the side constraints in an augmented Lagrangian function, and minimizing a sequence of these functions subject only to the network constraints for different estimates of the Lagrange multipliers of the side constraints. This method gives rise to an algorithm, which combines first- and superlinear-order multiplier methods to estimate these multipliers. When the number of free variables is very high we can obtain a superlinear-order estimate by means of the limited memory BFGS method fitted to our problem. An extensive computational comparison with other methods has been performed. The numerical results reported indicate that the algorithm described may be employed advantageously to solve large-scale network flow problems with nonlinear side constraints.

Efficient solution of optimal multimarket electricity bid models

Short-term electricity market is made up of a sequence of markets, that is, it is a multimarket enviroment. In the case of the Iberian Energy Market the sequence of major short-term electricity markets are the day-ahead market, the ancillary service market or secondary reserve market (henceforth reserve market), and a set of six intraday markets. Generation Companies (GenCos) that participate in the electricity market could increase their benefits by jointly optimizing their participation in this sequence of electricity markets. This work proposes a stochastic programming model that gives the GenCo the optimal bidding strategy for the day-ahead market (DAM), which considers the benefits and costs of participating in the subsequent markets and which includes both physical futures contracts and bilateral contracts.

Approximate Subgradient Methods for Lagrangian Relaxations on Networks

Nonlinear network flow problems with linear/nonlinear side con- straints can be solved by means of Lagrangian relaxations. The dual problem is the maximization of a dual function whose value is estimated by minimizing approximately a Lagrangian function on the set defined by the network constraints. We study alternative stepsizes in the approximate subgradient methods to solve the dual problem. Some basic convergence results are put forward. Moreover, we compare the quality of the computed solutions and the efficiency of these methods.

Lagrangian Relaxations on Networks by ε-Subgradient Methods

The efficiency of the network flow techniques can be exploited in the solution of nonlinearly constrained network flow problems by means of approximate subgradient methods. The idea is to solve the dual problem by using ε-subgradient methods, where the dual function is estimated by minimizing approximately a Lagrangian function, which relaxes the side constraints and is subject only to network constraints. In this paper, convergence results for some kind of ε-subgradient methods are put forward. Moreover, in order to evaluate the quality of the solution and the efficiency of these methods some of them have been implemented and their performances computationally compared with codes that are able to solve the proposed test problems.

Solving electric market quadratic problems by Branch and Fix Coordination methods

The electric market regulation in Spain (MIBEL) establishes the rules for bilateral and futures contracts in the day-ahead optimal bid problem. Our model allows a price-taker generation company to decide the unit commitment of the thermal units, the economic dispatch of the bilateral and futures contracts between the thermal units and the optimal sale bids for the thermal units observing the MIBEL regulation. The uncertainty of the spot prices is represented through scenario sets. We solve this model on the framework of the Branch and Fix Coordination metodology as a quadratic two-stage stochastic problem. In order to gain computational efficiency, we use scenario clusters and propose to use perspective cuts. Numerical results are reported.

An Algorithm for Two-Stage Stochastic Quadratic Problems

An algorithm for solving quadratic, two-stage stochastic problems is developed. The algorithm is based on the framework of the Branch and Fix Coordination (BFC) method. These problems have continuous and binary variables in the first stage and only continuous variables in the second one. The objective function is quadratic and the constraints are linear. The nonanticipativity constraints are fulfilled by means of the twin node family strategy. On the basis of the BFC method for two-stage stochastic linear problems with binary variables in the first stage, an algorithm to solve these stochastic quadratic problems is designed. In order to gain computational efficiency, we use scenario clusters and propose to use either outer linear approximations or (if possible) perspective cuts. This algorithm is implemented in C++ with the help of the Cplex library to solve the quadratic subproblems. Numerical results are reported.

A Variant of the Constant Step Rule for Approximate Subgradient Methods over Nonlinear Networks

The efficiency of the network flow techniques can be exploited in the solution of nonlinearly constrained network flow problems (NCNFP) by means of approximate subgradient methods (ASM). We propose to solve the dual problem by an ASM that uses a variant of the well-known constant step rule of Shor. In this work the kind of convergence of this method is analyzed and its efficiency is compared with that of other approximate subgradient methods over NCNFP.

Efficient dual methods for nonlinearly constrained networks

The minimization of nonlinearly constrained network flow problems can be performed by exploiting the efficiency of the network flow techniques. It lies in minimizing approximately a series of (augmented) Lagrangian functions including only the side constraints, subject to balance constraints in the nodes and capacity bounds. One of the drawbacks of the multiplier methods with quadratic penalty function when is applied to problems with inequality constraints is that the corresponding augmented Lagrangian function is not twice continuously differentiable even if the cost and constraint functions are. The authors purpose is to put forward two methods that overcome this difficulty: the exponential multiplier method and the e-subgradient method, and to compare their efficiency with that of the quadratic multiplier method and that of the codes MINOS and LOQO. The results are encouraging.