Learning in Mechatronic Conceptions

článek v časopise - ostatní, Jost

en

Mechatronic conceptions are most frequently characterized as synergistic conjunction of the mechanics, electrotechnics and computer science. Computer science as a platform of the realization of control algorithms especially increasingly runs the soft computing algorithms. Soft computing differs from conventional (hard) computing in the basic principle: it exploits the tolerance for imprecision, uncertainty and partial truth to achieve tractability, robustness and low solution cost. The most important components of soft computing are fuzzy logic, neural network theory, probabilistic reasoning, genetic algorithm, chaos theory and parts of machine learning theory. Fundamental issue is that the principal contributions of cited components are complementary, not competitive (leading on hybrid systems creation, etc.). The survey of the most interesting ideas of learning used in soft computing is introduced in this contribution.

Mechatronic conceptions are most frequently characterized as synergistic conjunction of the mechanics, electrotechnics and computer science. Computer science as a platform of the realization of control algorithms especially increasingly runs the soft computing algorithms. Soft computing differs from conventional (hard) computing in the basic principle: it exploits the tolerance for imprecision, uncertainty and partial truth to achieve tractability, robustness and low solution cost. The most important components of soft computing are fuzzy logic, neural network theory, probabilistic reasoning, genetic algorithm, chaos theory and parts of machine learning theory. Fundamental issue is that the principal contributions of cited components are complementary, not competitive (leading on hybrid systems creation, etc.). The survey of the most interesting ideas of learning used in soft computing is introduced in this contribution.

Q-learning, computer science, control algorithms

2001

01.12.2001

1210-2717

8

6

12