Virtual-Intelligence Applications in Petroleum Engineering: Part 2—Evolutionary Computing

Abstract

Distinguished Author Series articles are general, descriptive representations that summarize the state of the art in an area of technology by describing recent developments for readers who are not specialists in the topics discussed. Written by individuals recognized to be experts in the area, these articles provide key references to more definitive work and present specific details only to illustrate the technology. Purpose: to inform the general readership of recent advances in various areas of petroleum engineering. Summary Part 1 of this series discussed artificial neural networks and provided a general definition of virtual intelligence.1 The goal of this second article is to provide an overview of evolutionary computing, its potential combination with neural networks to produce powerful intelligent applications, and its applications in the oil and gas industry. The most successful intelligent applications incorporate several virtual-intelligence tools in a hybrid manner. Virtual-intelligence tools complement each other and are able to amplify each other's effectiveness. This article also presents the background of evolutionary computation as related to Darwinian evolution theory. This is followed by a more detailed look at genetic algorithms, the primary evolutionary-computing paradigm currently used. The article concludes by exploring application of a hybrid neural network/genetic algorithm system to a petroleum-engineering-related problem. Background Evolutionary computing, like other virtual-intelligence tools, has its roots in nature. It is an attempt to mimic the evolutionary process by use of computer algorithms and instructions. The question, however, is why mimic the evolutionary process. The answer becomes obvious once we realize what types of problems the evolutionary process solves and consider whether we would like to solve similar problems. Evolution is an optimization process.2 One of the major principles of evolution is heredity. Each generation inherits the evolutionary characteristics of the previous generation and passes those same characteristics to the next generation. These characteristics include progress, growth, and development. The passing of the characteristics from generation to generation is facilitated through genes. Since the mid-1960's, new analytical tools for intelligent optimization inspired by the Darwinian evolution theory have surfaced. The term" evolutionary computing" has been used as an umbrella for many of these tools. Evolutionary computing comprises evolutionary programming, genetic algorithms, evolution strategies, and evolution programs, among others. To many people, these tools (and names) look similar and their names appear to have the same meaning. However, the names carry quite distinct meanings to scientists who are deeply involved in this area of research. Evolutionary programming, introduced by Koza,3 is concerned primarily with solving complex problems by evolving sophisticated computer programs from simple, task-specific computer programs. Genetic algorithms, the subject of this article, are discussed in detail in the next section. In evolution strategies,4 the components of a trial solution are viewed as behavioral traits of an individual (not asgenes along a chromosome) as implemented in genetic algorithms. Evolution programs5 combine genetic algorithms with specific data structures to achieve their goals. Genetic Algorithms Darwin's theory of survival of the fittest (presented in his 1859 paper titled "On the Origin of Species by Means of Natural Selection"), coupled with the selectionism of Weismann and the genetics of Mendel, have formed the universally accepted set of arguments known as the evolution theory.4 In nature, the evolutionary process occurs when the following four conditions are satisfied.3An entity has the ability to reproduce.There is a population of such self-reproducing entities.There is some variety among the self-reproducing entities.This variety is associated with some difference in ability to survive in the environment. In nature, organisms evolve as they adapt to dynamic environments. The" fitness" of an organism is defined by the degree of its adaptation to its environment. The organism's fitness determines how long it will live and how much of a chance it has to pass its genes on to the next generation. In biological evolution, only the winners survive to continue the evolutionary process. It is assumed that, if the organism lives by adapting to its environment, it must be doing something right. The characteristics of the organisms are coded in their genes, and their genes pass these characteristics to their offspring through the process of heredity. The fitter an individual, the higher is its chance to survive and, hence, reproduce.