The figure shows the stages of evolution in control theory. Control theory dates from 1868, when Maxwell made a theoretical analysis of the functions of a speed governor used to maintain a constant rotational speed in steam engines. Classical control theory was refined in the 1940s and 1950s, and is based on a mathematical model which gives the relationship between the input and output of the process to be controlled. Classical control theory deals with only one input and one output. However, control equipment called the general-purpose PID controller, which permits proportional, integral, and differential control, has been developed and is used in many processes and plants. Subsequently, modern control theory has been developed to handle multivariable systems with many inputs and outputs. Control based on classical and modern control theories requires the preparation of an appropriate mathematical model of the process to be controlled. If the mathematical model is satisfactory, the expected control can be obtained in a quantitative way. However, with control systems involving substantial nonlinearity, there are many parameters that must be determined by experiment in order to construct the most appropriate mathematical model. It is thus difficult to prepare satisfactory mathematical models, and hence the effect of control is limited. Remarkable progress has been made in recent years in (i) fuzzy theory, which can handle ambiguity, (ii) expert systems, which utilize the knowledge of experts, and (iii) neural networks, which are very effective for pattern recognition and learning. These methods are called intellectual information processing techniques and are employed effectively in various problem-solving systems. Control theory has developed by adopting these methods, and it has become clear that the concepts and means of control can be applied not only to manufacturing products, but also to all problems related to human activities, products, and information, such as production planning, inventory, and material distribution. Control theory was first applied by the steel industry to control one machine/device, and then to control a manufacturing line and process involving multiple connected devices. Such theory is now applied in the optimization of the operation of a whole steel works incorporating many lines. Moreover, control theory is already applied to some extent in total optimization of company performance, including logistics, production, and sales activities.