Computational Intelligence: An Introduction - Computational Intelligence in Time Series Forecasting - page 11

intelligent computational tools presented above guarantee their successful use in
solving the majority of high-complexity problems in the industrial world. This was
demonstrated on a number of examples published in the last decade.
The earliest use of fuzzy logic in the process industry was recorded in Japan,
where, in the late 1980s, fuzzy logic facilities capable of solving complex
nonlinear and uncertainty problems of a chemical reactor were used to replace the
skilled plant operator. Around the same time, neural networks were applied in
statistical analysis of huge sets of acquired sensor data by time series analysis and
forecasting. This application was later extended to include data mining for
managing very large amounts of more complex data using the methodologies of
soft computing based on pattern recognition and multisensor data fusion. This was
helpful in better understanding the process behaviour through analysis and
identification of essential process features hidden in data piles. In addition, it was
also possible to solve some accompanying problems related to plant monitoring
and diagnosis, product quality control, production monitoring and forecasting,
plant logistics and various services, etc .
In the iron and steel industry, enormous progress was made after introducing
intelligent computational approaches in process modelling, advanced process
control, production planning and scheduling, etc . For more than three decades the
steel producers have profited from advanced methods, starting with direct digital
control and finishing with the glorious distributed computer control systems
developed by systems and control engineers (Popovic and Bhatkar, 1990). With the
advent of intelligent computational technologies, fuzzy logic control, neural
networks-based modelling, intelligent sensing, evolutionary computing-based
optimization at various process and plant levels, etc. have been on the agenda
mainly because of high international competition in this industrial branch in
producing high quality product at the lowest production cost.
However, it was the electronic industry that has to the most remarkable extent
profited from the introduction of intelligent computational technology in chip
design and production processes.
Computational intelligence has also found wide application in manufacturing,
particularly in product design, production planning and scheduling, monitoring of
tool wear, manufacturing control and monitoring of automated assembly lines, and
product quality inspection (Dagli, 1994). The use of intelligent technologies in this
area was particularly accelerated after the discovery and massive applications of
the mechatronics approach in product development. This has also contributed to
extending the application field of intelligent technology to include rapid
prototyping, integration of smart sensors and actuators, design of internal
communication links oriented systems, etc. (Popovic and Vlacic, 1999).

References

[1] Akhimetiv DF and Dote Y (1999) Fuzzy system identification with general parameter

radial basis function neural network. In: Farinwata SS, Filev D, and Langari R (Eds)
Fuzzy control synthesis and analysis, Wiley, Chichester, UK, Ch. 4.