Neurocontrol. Learning control systems inspired by neuronal architectures and human problem solving strategies (Q1201979)

The book represents an elaborated technical report on a project, utilizing ideas of various paradigms of the neural network theory closely following neurophysiological findings, automatically interpolating associative memories with local generalization properties (including ideas of finite element approach), genetic algorithms and others. All that gives an original mixture of techniques for a design of learning control loop, implemented in an explicit process model called LERNAS. The approach is illustrated on the main example (pH neutralization process), which continues throughout the book, but other applications are discussed as well. The presentation, divided into 3 chapters [with somewhat nontypical titles: II. Microintelligence, III. Macrointelligence] is not always easy to follow in details, some parts being repetitious and not well- structured, but for some applications it may pay off to study the material carefully. It incorporates many figures (graphs, tables of experimental results), references (both to well-known and specific literature) and comparisons with more standard models.