|
Abstract:
|
Applying evolutionary computation in industry requires resolving many issues of technical and non-technical nature. This tutorial gives a systematic approach of evolutionary computation in industrial settings based on experience within The Dow Chemical Company. The competitive advantages of exploiting evolutionary computation are discussed as is internal marketing issues. The tutorial covers an implementation methodology based on integrating multiple complimentary computational intelligence methods, notably, genetic programming for symbolic regression, neural networks, support vector machines, and particle swarm optimization. Unique extensions to these techniques which have been found to be highly effective are also presented. Examples of successful applications that created significant value in the area of inferential sensors, process optimization, and new product development are given. The tutorial also gives a summary of the key open issues for future research in the area of evolutionary computation driven by the current industrial needs.
|
