Keywords: computational intelligence, neural and Bayesian networks, genetic algorithms, neuro-fuzzy networks, steel and composite design and behaviour, patch load, semi-rigid joints, block shear, pre-stressed stayed steel columns.

Structural engineers have long been searching for the conception of structures that could express their ability to withstand their imposed loads and usage, not only with lightness and economy, but also with the aim of becoming aesthetic landmarks and being in line with the current sustainability demands. Various examples of these attempts, over the years, can be found worldwide like the Eiffel Tower, Tower Bridge, Sidney Opera House, Bilbao Guggenheim Museum, among several others. With the development of new materials and faster computing processes, a new frontier has been opened for the conception and development of new and bolder designs that will set the trend for future 21th century structures. Numerous methods, techniques and tools have been, and still are being, used to improve and design these structures like optimization processes, numerical modelling systems involving non-linear finite element analysis among others.

At the same time, the last decade of the 20th century has been associated with substantial improvement and development of the so-called Computational Intelligent techniques, which are computational systems that try to mimic human behaviour, such as perception, reasoning, learning, evolution and adaptation. These techniques encompass Neural Networks, Genetic Algorithm, Fuzzy Logic, and Hybrid Intelligent Systems, such as Neuro-Fuzzy, Neuro-Genetic & Fuzzy-Genetic models.

Intelligent systems have been successfully applied in forecasting, optimization, risk analysis, control, inference, modelling and fraud detection, in a wide range of knowledge fields. These systems offer comprehensive solutions for managers and decision makers to a large number of complex and extensive applications that are considered difficult, totally limited or even impossible by companies in different fields: Economy and Finance, Trade and Commerce, Engineering, Energy, Transport, Logistic, Marketing, Environment, Medicine, etc. Therefore, numerical tools based on these ideas have been widely developed and were successfully applied not only to engineering applications but also to a wide range of problems.

This chapter presents a brief introduction of neural networks, genetic algorithm and hybrid neuro-fuzzy systems, which already have been used to forecast, design and optimise the structural behaviour. As mentioned before, the authors did not intend to provide an exhaustive list of the use of Computational Intelligence techniques to solve Structural Engineering problems. The main idea was to focus on some of these methods to enable a deeper insight into the structural engineering problems that could be modelled with their proper use.

After presenting the basic concepts of Neural Networks, Genetic Algorithms and a special hybrid neuro-fuzzy model based on hierarchical partitioning (HNFB-Class - Hierarchical Neuro-Fuzzy BSP model for Classification problems), the paper details some Structural Engineering case studies. Examples of these case studies are patch load prediction, block shear design, Perfobond shear connectors, stayed steel columns, steel and composite joint design and response, among others. All these case studies corroborate the great potential of Computational Intelligence techniques for solving problems that were considered difficult, limited or even impossible by many researchers in different fields. All case studies provided a performance, in some cases, beyond that expected, suggesting that these techniques might be a good solution in many other applications.

purchase the full-text of this chapter (price £20)

go to the previous chapter
go to the next chapter
return to the table of contents
return to the book description
purchase this book (price £88 +P&P)