Book Reviews
Abstract
In Book Reviews, we review an extensive and diverse range of books. They cover theory and applications in operations research, statistics, management science, econometrics, mathematics, computers, and information systems. In addition, we include books in other fields that emphasize technical applications. The new book reviews editor, Wenjing Shen ([email protected]), will be pleased to receive an email from those willing to review a book, with an indication of specific areas of interest. If you are aware of a specific book that you would like to review, or you think should be reviewed, please contact Wenjing.
The following books are reviewed in this issue of Interfaces, 45(2), March–April 2015: Nonlinear Optimization Applications Using the GAMS Technology, Neculai Andrei; Integrating Renewables in Electricity Markets, Juan M. Morales, Antonio J. Conejo, Henrik Madsen, Pierre Pinson, and Marco Zugno; Modelling and Managing Airport Performance, Konstantinos G. Zografos, Giovanni Andreatta, and Amedeo R. Odoni.
Nonlinear Optimization Applications Using the GAMS Technology
Andrei, Neculai. 2013. Nonlinear Optimization Applications Using the GAMS Technology. Springer, 340 pp. $129.00.
In this book, Neculai Andrei presents an impressive anthology of nonlinear optimization applications from the natural sciences, engineering, and other areas. He focuses on the formulation of such problems in an algebraic modeling language (AML). His choice of an AML is GAMS, for which he gives an introduction to the language elements required for the models he presents. Although the book contains many GAMS examples, the programs are mostly monolithic declarative models, and their translation into other AMLs supporting nonlinear optimization, such as AIMMS, AMPL, LINGO, or MOSEL, are straightforward.
The 82 applications are grouped into 10 application areas, including optimal control and various engineering areas. For each application, the author provides a description of the problem, a detailed list of references for further reading, and a mathematical description of the optimization problem. In addition, he provides GAMS source code of the optimization problem and some performance indicators for different nonlinear programming (NLP) solvers. His focus is primarily on local NLP solvers; they guarantee only local optimality and cannot prove global optimality or infeasibility. In particular, the solvers Conopt, Knitro, Minos, and Snopt are used in most runs. The motivation for selecting NLP solvers is not provided (e.g., the open-source solver, Ipoptis, is missing from the solver list). However, the emphasis on solver performance has only temporal value. Solver technology is constantly advancing and great progress has been made using new algorithms that are based on interior point methods and global optimization.
I agree with the author that the emphasis should be on running solvers without setting any algorithm-specific options; however, this can lead to some misleading conclusions. For example, the application, Optimal Design of a Gear Train (Application 5.5), is one of the few applications with discrete integer-valued decision variables. The author solves the model using the global optimization solver BARON, which inherits its relative optimality tolerance from GAMS; GAMS has a very large default value of 0.1. Hence, the solution (14, 14, 37, 37) that the book reports with this loose tolerance is not optimal. The optimum solution is (19, 16, 49, 43).
After reading the introduction and the chapter on GAMS, which is written at an rudimentary level for an audience unfamiliar with the GAMS language, the reader can select the book’s chapters in any order. Each application is self-contained and uses only elementary GAMS syntax. The description and model algebra of some applications are terse. Therefore, a reader’s understanding of the material discussed will depend on that reader’s experience in the particular application domain. Other applications, such as Finding the Surface with Minimal Area that Lies Above an Obstacle with Given Boundary Conditions (Minsurf) (Application 3.9), are described in more detail, thus allowing a wider audience to grasp the problem and model formulation without consulting other literature.
The inclusion of actual GAMS source for the applications might motivate the reader to experiment with the given model formulation. Unfortunately, the GAMS source code is not available online. The freely downloadable GAMS system contains model collections from other books, and allows the user to solve small demonstration models (many of the 82 models in the book fit within the GAMS demonstration limits) with all NLP solvers without requiring a license. Making the GAMS source code of the models available to the reader would significantly increase this book’s value; however, the elementary use of GAMS code in many examples is neither consistent with current information technology standards for clean and systematic programming (e.g., separating data and model, avoiding hardcoded numerical constants or data), nor does it indicate the intrinsic power of the GAMS modeling language; the fuel-allocation optimization problem is one of the better examples. The other limitation is that most models are small and do not reflect sizes of problems that state-of-the-art solvers can handle.
What would an undergraduate student learn from this book? Mostly, that a broad range of real-world problems can be modeled as NLP problems and solved using GAMS; however, this would require some didactic involvement of instructors to teach the students how to use current coding and programming techniques.
Although the book’s scope is limited to models, their corresponding GAMS code, and some related mathematics, most readers will learn something from reading it. For example, we can learn that the pooling problem is not only about car sharing and driving to campus or work with an optimal partner, and catmix is not about four-legged, mice-eating animals, but about optimal policies of mixing two catalysts.
The models discussed can be useful to many people. Physicists will enjoy differential equations models, such as the Van der Pol oscillator, being discretized and implemented in GAMS. Partial differential equations resulting from heat transfer or fluid dynamics problems are solved in their discretized version using GAMS. For space enthusiasts, the book provides a small section on the dynamic optimization of a rocket, and the examples of discretized optimal control problems can serve as examples for more complicated real-world problems.
So, peruse the book, learn from it, and enjoy the models. If, along the way, you happen to find some useful hints that may help you in your GAMS-related life, so much the better.
Josef Kallrath
BASF SE, Scientific Computing, Ludwigshafen, Germany; Department of Astronomy, University of Florida, Gainesville, Florida, [email protected]
Integrating Renewables in Electricity Markets
Morales, Juan M., Antonio J. Conejo, Henrik Madsen, Pierre Pinson, Marco Zugno. Integrating Renewables in Electricity Markets. Springer, 429 pp. $99.00.
Integrating Renewables in Electricity Markets is Volume 205 in Springer’s International Series in Operations Research and Management Science. Its intended audience is advanced undergraduate and graduate students studying electric energy systems. Its focus is on the increasingly stochastic nature of grid operations, particularly renewable generation. Electrical energy systems worldwide are incorporating renewable energy-based generation such as wind and solar powered plants. Although these plants provide cheap electricity, their energy production can be unpredictable as a result of uncertain weather conditions—solar irradiance in the case of solar plants and wind speed and direction for wind farms.
This book has nine chapters. Chapter 1 provides an introduction to pool-based electricity markets: the day-ahead, intraday, and balancing markets. After a brief discussion of the historical evolution of these markets, this chapter illustrates the impact of renewables on these markets; in particular, it describes the uncertain nature of renewable generation and how it puts a premium on flexibility—from the perspective of both traditional generation units and consumer demand. Chapter 2 introduces models to characterize stochastic renewable production units. Its primary theme is that although conventional generation is deterministic in nature because it is controllable, renewable generation is unpredictable and uncontrollable; hence, a need exists for stochastic production forecasting models. This chapter describes some of these models. Chapter 3, after providing a brief introduction to the futures market and bilateral contracts, shifts the discussion to day-ahead markets. It describes various aspects of these markets, including the process of submitting consumption bids and production offers, reserve requirements, and features of market-settlement schemes. It also presents stochastic programming and robust optimization models for day-ahead auctions. Chapter 4 addresses real-time or balancing markets, which are used to balance the production and consumption close to energy delivery. These markets have become increasingly important in regions with high renewable penetration as a result of the inherent uncertainty in their power generation. The chapter describes both the market process and introductory formulations. Chapter 5 addresses how the integration of renewable production facilities necessitates flexible production capacity, which is provided primarily by combined cycle gas turbines. It then introduces a formulation that accommodates capacity, ramp rates, and minimum up- and downtime requirements. Chapter 6 characterizes the impact of stochastic renewable production on energy market outcomes; the predicted levels of renewable generation affect the day-ahead markets, whereas the uncertainty in forecasting renewable generation affects balancing markets. This chapter then gives a methodology for this analysis using a Northern European market with a high penetration of wind generation. Chapter 7 offers models for optimal trading strategies for renewable power producers. The context is that because renewable producers in day-ahead markets are forced to follow the same rules as conventional producers, they are penalized for real-time deviations from day-ahead schedules. Thus, they should consider a multiple-market strategy, which they may formulate as a multistage problem under uncertainty. Chapter 8 introduces virtual power plants that are enabled by distributed energy resources. The concept is that a combination of distributed (i.e., localized) generation and flexible loads, called a virtual power plant, can be used to replace a conventional plant and should participate in the energy markets. Chapter 9 describes the strategy of using demand response (i.e., demand management using financial incentives) to address production uncertainty by letting demand follow supply. It describes various marketing frameworks that can be used to control flexible demand, including dynamic pricing for which it includes models. The lesson from this chapter is that price-related demand actions enable overall economic improvement in the operation of the entire system.
This book’s content flows logically. After discussing various markets, it covers the impact of volatility in renewable production on these markets, and then introduces the reader to the emerging concepts of virtual power plants and demand response. Each chapter starts with an introduction to the topic, follows with relevant models, and ends with a section on additional readings to help the reader who wants more knowledge about that topic. One major omission in this book is that it has few real-world examples or case studies to support the models presented. In addition, sentence constructions were awkward at times, perhaps as a result of differences in regional styles; however, this did not detract from the material presented. I think this book will provide readers, particularly operations research (OR) professionals, with exposure to an important research area that has traditionally been in the power engineering domain. Many aspects of grid operations require power engineering principles and are best handled by that community; nevertheless, some applications require OR expertise. In addition, emerging technologies, such as demand response and virtual power plants, have opened up entirely new research areas.
Here, I illustrate a potential research problem based on my industry experience. Suppose a mining facility is facing hourly electricity price fluctuations, which were announced the previous day. Because the mine also needs heat resources, installing a combined heat and power (CHP) plant, which can provide both electricity and heat from steam, may be an economical decision. The mine operators may also choose to use battery storage as a backup emergency supply, and could also use it to address the short-term peaks that can adversely affect the demand-charges component of the mine’s electricity bill. Of course, some mine operations may have some scheduling flexibility. Therefore, we have two research questions: (1) how should the mine operators plan for distributed resources, that is, the size and capability of the CHP and battery, including their required ramp-up rates, and (2) how should the operators plan the next-day operations; that is, once they know the day-ahead hourly rates, and given the mine’s power needs (for both flexible and inflexible tasks) for the next day, how should they best use CHP (hour and output) and battery (charge and discharge schedule), considering that CHP incurs startup costs and battery storage incurs cycling costs. Another research problem is the design of a demand-response program such that both customers and the utility benefit financially from its implementation. These research problems are only two of the many that renewable generation and other new concepts in grid operations have spawned.
I recommend this book to readers who are interested in learning the fundamental analytical models pertaining to renewable generation in energy markets. However, I cannot recommend it to those who want a broad understanding of this topic with real-life illustrations and case studies.
Rajesh Tyagi
GE Global Research Center, Niskayuna, New York, [email protected]
Modelling and Managing Airport Performance
Zografos, Konstantinos G., Giovanni Andreatta, Amedeo R. Odoni, eds. 2013. Modelling and Managing Airport Performance. Wiley, 314 pp. $130.00.
Airport and airspace operations in Europe and the United States are rife with challenges because they must function at or near capacity during much of any given day. Even modest reductions in capacity may lead to significant delays and cancellations at constrained stations. In 2013, 28 percent of all flight delays were attributed to problems with the National Airspace System (NAS), which was the highest of all root causes of delays (United States Department of Transportation 2013). The demand in air transportation is expected to outpace that of capacity through 2034 (Federal Aviation Administration 2014); moreover, such problems will only exacerbate unless significant changes are made. Edited by a distinguished team of editors, Modelling and Managing Airport Performance is a collection of 10 chapters by renowned industry experts and academics who address challenges to the future of strategic and tactical airport operations.
Investing in airport infrastructure, with the objective of ultimately increasing station capacity, will play an important role in bridging the gap between the supply and demand of air transport. Thus, the importance of being able to analyze the benefits of airport performance is obvious; however, these benefits are often not developed sufficiently. The first subject this book explores is the development of a foundation to model and evaluate station performance. Chapter 1 explores modeling airport landside performance by introducing recommendations in level-of-service measures. The authors also explore how airports may evaluate their impact on relevant passenger characteristics by developing an analytical hierarchy process. Chapter 2 introduces a broad problem-oriented framework with considerable granularity at a systems level to address challenges to various stakeholders. The modeling capability is unique in its ability to capture the behavior of different entities, each with its own objectives, in a holistic manner to provide the ability to evaluate total airport performance.
The next three chapters are concerned with quantifying and forecasting airport delays. Chapter 3 compares the performance and policies in the largest airports in the United States and Europe. The authors propose ways to mitigate delay by sharing best strategies from each region. Chapter 4 explores the forecasting of near-term delays; the authors provide two approaches by comparing a simple and intuitive approach with a complex and system-wide alternative. Chapter 5 proposes an econometric model, which differs from traditional cost-factor models, to measure airline delay costs. The new paradigm has several advantages in that it makes fewer assumptions, is more data intensive, and is likely more accurate.
In addition to performance costs associated with delays, other important nonpecuniary costs are associated with airports; two are the focus of the subsequent two chapters. Chapter 6 discusses external factors associated with four airport pollutants: noise, water run-off, criteria pollutants, and greenhouse gas emissions. Its authors discuss how integrating environmental-impact policy models will become critical as environmental sensitivity becomes more ubiquitous. In Chapter 7, the authors consider airport safety performance by examining the evolution of safety performance for flight ground segments between 1990 and 2008. By focusing on six International Civil Aviation Organization (ICAO) category groupings, they identify the areas in which flight safety has improved and those in which the results are more ambiguous. They conclude by providing recommendations for further initiatives to address areas prone to greater risk as a result of growth in air transportation.
The final three chapters explore airport congestion management. Chapter 8 introduces a novel measure to predict airport performance based on a slot-coordination mechanism and validates it using data from five German airports. Chapter 9 examines European demand-management strategies; the authors identify shortcomings to existing policies motivated by mismatch and misuse of slots, poor allocation efficiency, declared capacity considerations, barriers to new entrants, and pricing effectiveness. They outline a nine-step initiative designed to guide the implementation process to alleviate such concerns at European airports. Chapter 10 concludes by exploring market-based mechanisms to control congestion at constrained U.S. airports. The authors begin by discussing the justification for slot control, which is far less prevalent at U.S. stations compared with their European counterparts. By using existing tools to model airline behavior in reaction to an increase in slot controls, they suggest that the cost savings associated with incorporating greater robustness more than offsets the costs of scheduled delays incurred by reducing or reallocating scheduled flights at several large airports.
The principal virtue of this book is the breadth of topics covered; several of these topics, which have hitherto not been addressed in the literature or in decision support systems, are pertinent to understanding airports of the future. The span of topics is such that even the most seasoned professional will inevitably learn something new from reading about an area close to his (her) area of expertise. Industry practitioners and academics have made considerable progress in understanding many individual problems within the overall complex network that comprises airport operations; however, they have generally done so in isolation of one another. The editors correctly point out that future research requires an integrated perspective allowing users to “account for the often conflicting needs and interests of the multiple stakeholders involved” (p. xxvi). This book lays a solid foundation in describing current processes, future challenges, and most importantly the fundamental trade-offs, as determined by various decision makers, which are an essential component to meeting future challenges.
One limitation of this book is that it does not provide a comprehensive overview of how airline behavior is anticipated to change as a result of the policies governed by airport or airspace authorities, as alluded to in several chapters. Given the need to holistically understand the system, the development of a general, rather than partial, equilibrium is critically important for understanding future systems. This important criterion is mostly ignored and is scarce on detail when it is present. A second limitation is the general inconsistency among the chapters. The contribution, level of technical detail, rigor of analysis, and impact on future research varies greatly among the chapters.
Jon Petersen
Taleris, Austin, Texas, [email protected]

