Stochastic calculator

The stochastic concept of calculator is already old (for the young history of data processing) and contemporary of research and applications developed at the any end of the decade 1950 and in the decade 1960 to mitigate the insufficiencies of speed of treatment and in capacity of autocorrection of the computers based on the principles of the Machine of Turing and the Boolean logical - inter alia the imperative Programmation represented by the language Lisp (1958/1960) and its derivative Logo (1966), and the fuzzy Logique (1965), which has besides relationship with the stochastic Calcul.

Two problems of the analog computers were the cost of realization of the circuits of multiplication and their slowness. At the end of years 1960 of the researchers of the company Alsthom had the idea to represent a size between 0 and 1 by a Probabilité of presence of one 1 on a line. A multiplication then required nothing any more but one simple and logical (very cheap circuit to produce) of a very good vitesse : the estimate with 1% of the probability required only 10ms with a frequency of 10.000 Hz. Still it for that was necessary that the entries are décorrélées, which was carried out by a small auxiliary attachment containing a generator random and of a circuit or exclusive (XOR) .

The calculators stochastic , good market, reliable and easy to produce in great quantity, were regarded in the years 1970 as promising substitutes with the fluidic circuits used in the military and industrial world. However the production at increasingly weak prices of Microprocesseur S carried a fatal blow to them. The theory keeps nevertheless a ludic attraction today, and it is possible to manufacture with few expenses (or to emulate on computer) stochastic calculators. Their excellent reliability and their fast reaction time could make them again useful a day in the industrial processes or the military applications - if it is not already done.

Random links: