Evolutionary Computation: Fundamentals and Applications

By Fouad Sabry

What Is Evolutionary Computation

The term "evolutionary computation" refers to both a subfield of artificial intelligence and soft computing in the field of computer science that studies various optimization algorithms and a family of algorithms for global optimization that are inspired by biological evolution. They belong to a family of problem solvers known as population-based trial and error problem solvers, and they either have a metaheuristic or stochastic optimization flavor.

How You Will Benefit

(I) Insights, and validations about the following topics:

Chapter 1: Evolutionary computation

Chapter 2: Differential evolution

Chapter 3: Dual-phase evolution

Chapter 4: Evolutionary algorithm

Chapter 5: Evolutionary programming

Chapter 6: Neuroevolution

Chapter 7: Evolutionary robotics

Chapter 8: Grammatical evolution

Chapter 9: Human-based evolutionary computation

Chapter 10: Interactive evolutionary computation

(II) Answering the public top questions about evolutionary computation.

(III) Real world examples for the usage of evolutionary computation in many fields.

(IV) 17 appendices to explain, briefly, 266 emerging technologies in each industry to have 360-degree full understanding of evolutionary computation' technologies.

Who This Book Is For

Professionals, undergraduate and graduate students, enthusiasts, hobbyists, and those who want to go beyond basic knowledge or information for any kind of evolutionary computation.

• Language: English
• Publisher: One Billion Knowledgeable
• Release date: Jul 1, 2023

Book preview

Chapter 1: Evolutionary computation

The term evolutionary computation refers to both an area of artificial intelligence and soft computing in the field of computer science that studies various optimization algorithms and a family of algorithms for global optimization that are inspired by biological evolution. To speak more technically, they belong to the family of population-based trial and error problem solvers that have either a metaheuristic or stochastic optimization flavor.

The process of evolutionary computing begins by generating a preliminary pool of possible solutions, which is then iteratively improved upon. In order to develop each new generation, we first remove less desirable solutions using a stochastic process, and then we make minute random adjustments. In the language of biology, a population of potential solutions is mutated after being exposed to natural selection (or even artificial selection). As a consequence of this, the population will eventually evolve to grow in fitness, namely the fitness function that the algorithm has selected to represent fitness.

As a result of its ability to provide highly optimal solutions in a diverse variety of problem contexts, evolutionary computing approaches are becoming more prominent in the field of computer science. There are many other versions and extensions available, each of which is tailored to a particular family of issues or data structure. In the field of evolutionary biology, evolutionary computation is also used on occasion in the form of an in-silico experimental method to investigate various aspects of general evolutionary processes.

The idea of imitating evolutionary processes in order to solve issues dates back to a time before the invention of computers and includes examples such as Alan Turing's proposal in 1948 of a technique known as genetic search. Holland's primary objective was to utilize genetic algorithms to research adaptation and figure out how it may be reproduced, in contrast to the other techniques, which were largely focused on finding solutions to issues. An artificial selection technique was used to change populations of chromosomes, which were each represented as bit strings. This procedure selected for certain 'allele' bits included inside the bit string. Interactions between chromosomes were one of the many approaches that were used to replicate the recombination of DNA that occurs naturally between various kinds of animals. Holland's genetic algorithms monitored enormous populations, while earlier systems could only track a single optimum creature at a time (by having offspring compete with their parents) (having many organisms compete each generation).

In the 1990s, a novel method of evolutionary computing arose that would later be known as genetic programming. John Koza was one of the many people who supported this line of thought. Within this kind of algorithm, the thing that was evolving was really a program that had been created in a high-level programming language (there had been some previous attempts as early as 1958 to use machine code, but they met with little success). The programs that Koza used were called Lisp S-expressions, and you may conceptualize these expressions as trees made up of sub-expressions. This model makes it possible for programs to trade subtrees, which is similar to the mixing of genetic material. The degree to which a program is able to successfully carry out a predetermined assignment earns it a score, and that score is then employed in an artificial selection process. The concept of genetic programming has been successfully used in a number of areas, including