Structure of this Book

The purpose of this book is to provide discussions of a variety of AI methods and techniques along with source code applications to demonstrate them. Each method is discussed at a high level, as well as the low level, detailing the properties and flow of the algorithms. In many cases, practical and useful applications are selected to demonstrate the algorithms. In others, applications that are more theoretical are chosen to identify the usefulness of the algorithms.

Chapter 2 introduces simulated annealing, the modeling of problem solving after the physical process of annealing (cooling of a molten substance to a solid). The constraint problem of N-Queens was chosen to illustrate the algorithm's capabilities.

Chapter 3 discusses adaptive resonance theory (or ART) and its clustering capabilities. The modern problem of personalization was chosen to demonstrate its properties.

Chapter 4 details the newer technique of ant algorithms and their path -finding properties. The traveling salesman problem was chosen to illustrate how simulated ants could navigate a graph and identify Hamiltonian paths.

In Chapter 5, neural networks using the back propagation learning algorithm are discussed. To illustrate the generalization capabilities of the trained networks, the problem of creating game AI behaviors is detailed.

Chapter 6 introduces genetic algorithms and its subfield of genetic programming. The optimization capabilities of the genetic algorithm are demonstrated through the creation of instruction sequences to solve specific numeric problems.

In Chapter 7, the field of artificial life is discussed through the evolution of winner-takes-all neural networks. Simple organisms are evolved within a closed environment to demonstrate food webs and the evolution of survival skills within the organism's neural network controller.

Chapter 8 discusses an older technique from symbolic AI of rules-based systems. A simple forward-chaining, rules-based system is constructed and a fault-tolerant subsystem is encoded within rules and facts to manage sensors.

Chapter 9 introduces fuzzy logic and its strengths in building control systems. A number of examples are provided, detailing a simple battery charge control system using fuzzy control.

In Chapter 10, hidden Markov models are discussed along with other probabilistic graph methods. The technique is used to mimic existing literary works using them as a model for text generation.

Chapter 11 introduces the newer field of intelligent agents along with a discussion of the different types and their characteristics. A simple web-filtering agent is constructed that autonomously collects news items of interest based upon predefined search criteria.

Finally, in Chapter 12, some of the new AI methods and future of AI are discussed.