Optimization for decision making. Linear and quadratic models (Q1047613)

The book comprises a graduate-level textbook on linear and quadratic optimization models for decision making. It illustrates how to formulate real-world problems using linear and quadratic models, how to use efficient algorithms for solving these problems, and how to draw useful conclusions and derive useful planning information from the output of these algorithms. Several illustrative examples and exercises from a variety of application areas are used to illustrate the concepts of the book. In contrast to the majority of the available books on linear programming (LP) that focus mainly on the mathematics and the simplex method, this book emphasizes the intelligent modelling of real-world problems, and discusses new methods based on using the sphere, which uses matrix inversion operations sparingly and seems well suited to solving large-scale LPs, as well as problems that may not have the property of being very sparse. The book contains 9 chapters starting with a brief history of mathematical modelling in chapter 1. Chapter 2 discusses methods for formulating real-world problems, while chapter 3 explains via three case studies, the need for intelligent modelling in order to get good results. Chapters 4 to 6 deal with traditional topics in LP. More specifically, chapter 4 discusses polyhedral geometry and its role in the study of LP and the development of algorithms for the solution of LPs, chapter 5 treats duality theory, optimality conditions for LP and marginal analysis, and chapter 6 discusses variants of the simplex method. Chapter 7 discusses interior point methods (IPMs) for LP, while chapter 8 discusses the sphere methods, i.e., new IPMs that have the advantage of using matrix inversion operations sparingly. Finally, chapter 9 discusses extensions of the sphere methods to convex and nonconvex quadratic problems, and to 0-1 integer programs through quadratic formulations.