Modern Optimization with R

Table of Contents

1. Frontmatter

2. Chapter 1. Introduction

   Paulo Cortez

   Abstract

   This chapter first introduces the motivation for using modern optimization via the R tool. Then, three relevant aspects are discussed: how to represent a solution, how to evaluate the quality of solutions, and how to handle constraints. Next, an overall view of modern optimization methods is presented, followed by a discussion of their limitations and criticism. Finally, this chapter presents the optimization tasks that are used for tutorial purposes in the book.

3. Chapter 2. R Basics

   Paulo Cortez

   Abstract

   This chapter presents the necessary knowledge for non-R-experts to understand and test the book code examples. First, more basic knowledge is introduced, namely objects and functions, how to control execution, and how to import and export data. Then, more advanced features are discussed, including command line execution, parallel computing, interfacing with other computer languages, and interactive web applications.

4. Chapter 3. Blind Search

   Paulo Cortez

   Abstract

   This chapter discusses what is blind search and how it can be implemented using the R tool. In particular, it details two full blind search implementations, two grid search approaches (standard and nested) and a Monte Carlo (random) search.

5. Chapter 4. Local Search

   Paulo Cortez

   Abstract

   This chapter introduces several local search methods and their R implementations, namely hill climbing (pure and steepest ascent and stochastic variants), simulated annealing, and tabu search. Then, the chapter demonstrates how local search methods can be compared. Finally, it shows how modern optimization methods (including local search ones) can be tuned in terms of their internal parameters by using an iterated racing.

6. Chapter 5. Population Based Search

   Paulo Cortez

   Abstract

   This chapter introduces population based search methods and their R implementations, namely genetic and evolutionary algorithms, differential evolution, particle swarm optimization, ant colony optimization, estimation of distribution algorithm, genetic programming, and grammatical evolution. The chapter also presents examples of how to compare population based methods, how to handle constraints, and how to run population based methods in parallel.

7. Chapter 6. Multi-Objective Optimization

   Paulo Cortez

   Abstract

   This chapter is dedicated to multi-objective optimization. First, three demonstrative multi-objective tasks are presented. Then, three main multi-objective approaches are discussed and demonstrated: weighted-formula, lexicographic, and Pareto (e.g., NSGA-II and NSGA-III, SMS-EMOA, AS-EMOA).

8. Chapter 7. Applications

   Paulo Cortez

   Abstract

   This chapter presents real-world applications of previously discussed modern optimization methods: traveling salesman problem, time series forecasting, and wine quality classification.

9. Backmatter