Stochastic programming. Modeling decision problems under uncertainty (Q2325678)

This book aims to provides an essential introduction to stochastic programming, especially intended for graduate students. As already the authors pointed out in their preface: ``This book is based on the lecture notes that we have developed for the course Stochastic Programming for master's students in Econometrics and Operations Research at the University of Groningen, The Netherlands''. The book contains 8 chapters. The chapter headings are as follows: 1.~Introduction; 2.~Random objective functions; 3.~Recourse models; 4.~Stochastic mixed-integer programming; 5.~Chance constraints; 6.~Integrated chance constraint; 7.~Assignments; 8.~Case studies. The book ends with 5 appendices ( A.~Probability, B.~Elementary convex analysis, C.~Deterministic LP, D.~Maximum calculus, E.~Algorithms for convex non-linear optimization), a bibliography of 52 references and an index of terms. Each chapter is dedicated to a different model in stochastic programming. In each chapter, the authors start with explaining the main concepts of the models, and discuss small examples in full detail to provide insights into the mathematical properties of the models. Later, these properties are proved in general and presented in different theorems. Each chapter ends with algorithms for solving the models. In Chapter 1, the authors discuss a linear programming problem with random parameters, representing a decision problem under uncertainty. Three small examples of linear programming problems in which some coefficients in the matrix of coefficients, the right-hand side vector, and in the cost vector, respectively, are random variables are discussed: a blending problem, a production planning problem in wich the variable cost of production is minimized, under the condition that the production x should satisfy the demand omega, and a portfolio optimization problem. Chapter 2 presents some clasical approaches for dealing with randomness in the objective function: maximization of expected utility and the mean-variance model and the relation with the expected utility theory. Chapter 3 treats the most important class of models in stochastic programming: recourse models. First, models are presented with penalty costs in deterministic LP (individual penalty costs, refined individual penalty costs, joint penalty costs and penalty costs based on recourse actions) and then recourse models in stochastic linear programming (representations of recorse models, modeling aspects, special recourse structures and properties of recourse models). Finally, algorithms for recourse models are presented. Chapter 4 represents a generalization of the recourse model presented in Chapter 3, obtained by allowing integrality restrictions on some or all of the decision variables. The authors give some motivation why such mixed integer recourse models are useful and interesting and give several examples of applications, discuss mathematical properties of the general model as well as the so-called simple integer recourse model and give an overview of available algorithms. Chapter 5 treats the case of chance constraints, i.e., the modeling with chance constraints, the mathematical properties of chance constraints, the case of discrete distributions and algorithms for models with a joint chance constraints. In Chapter 6, some differences as well as similarities between chance-constrained and recourse models are presented. Mixing the ideas of both model types, the authors introduce the concept of integrated chance constraints. This chapter ends with algorithms for models with integrated chance constraints. Chapter 7 contains theoretical assignments which can be considered as homework assignment during this course of stochastic programming. Chapter 8 contains several recourse models with different recourse structures which are more practically oriented and represent a wide class of the variety of stochastic programming applications and are intended to be solved using a computer. The book is well writen. The book will be of interest to mathematicians, engineers, economics and especially graduate students.