Overview of Computational Intelligence

Overview of Computational Intelligence

Copyright: © 2014 |Pages: 19
DOI: 10.4018/978-1-4666-4908-8.ch002


This chapter presents an overview of computational intelligence. The chapter starts with an introduction about the issue of computational intelligence. Then, the related methodologies used in the book are discussed in the next section. Right after this, the use of computational intelligence methodologies to deal with various remanufacturing/reverse logistics problems are conducted. Finally, the conclusion drawn in the last section closes this chapter.
Chapter Preview


What is Computational Intelligence?

A major impetus in algorithmic development is to resolve increasingly complicated problems by designing various algorithmic models. Tremendous successes have been achieved through the modelling of biological and natural intelligence, resulting in so-called “computational intelligence (CI)”. In fact, the term “CI” was introduced for the emulation of “intelligent” functions of animal brain by digital electronic computers. It is a fairly new research field, which is still in a process of evolution. At a more general level, CI comprises a set of computing systems with the ability to learn and deal with new events/situations, such that the systems are perceived to have one or more attributes of reason and intelligence (Marwala & Lagazio, 2011). According to the degree of acceptance, we have divided CI into two categories, namely conventional and innovative CI. Some representative techniques under each class are shown in Figure 1.

Figure 1.

CI paradigms

Conventional CI Methods

Conventional CI primarily focuses on artificial neural network (ANN), fuzzy logic (FL), multi-agent system (MAS), evolutionary algorithms (EA) (e.g., genetic algorithm (GA), genetic programming (GP), evolutionary programming (EP), and evolutionary strategy (ES)), artificial immune systems (AIS), simulated annealing (SA), tabu search (TS), as well as two variants of swarm intelligence (SI), i.e., ant colony optimization (ACO) and particle swarm optimization (PSO). The main idea of conventional CI methods is self-organizing principles which inspired from biological systems or SI that have stronger probabilities in exhibiting autonomy in different ways. For example, the task allocation process in the insect colonies is collaboratively decided and performed according to the willingness of an individual such that the overall task is optimized with a global intelligence comprised of simple individual responses (Bonabeau, Dorigo, & Theraulaz, 1999). Similarity, experts have shown experimentally that immune system does not need any externally controlling entity of the brain to protect the organism from the pathogens (Timmis, Neal, & Hunt, 2000). Furthermore, due to the problems’ complexity, the CI methods can be either used individually or in combination with other techniques to form complex hybrid methodologies for achieving systems with enhanced capabilities, e.g., a single system can make decisions under uncertainty by using FL, learn and adapt by using ANN, and undergo evolutionary optimization by using GA.

Complete Chapter List

Search this Book: