Analog computer

Analog computer. An analog computer is a computational device that uses continuously variable physical quantities, such as electrical voltage, mechanical rotation, or fluid pressure, to model and solve problems. These machines represent numbers through analogous physical states, enabling them to simulate complex systems described by differential equations in real-time. Their operation is fundamentally based on the principles of mathematical analogy and physical system modeling, contrasting sharply with the discrete, symbolic processing of their digital counterparts.

Definition and basic principles

The core principle involves creating a physical analog, or parallel, of the mathematical relationships in the system under study. Key components include operational amplifiers configured to perform mathematical operations like summation, integration, and multiplication on input voltages. The behavior of these components is governed by the laws of electrical circuits and, historically, fluid dynamics or mechanical linkages. Computation is performed through the direct interaction of these physical phenomena, with the output often being a continuous graph plotted on a device like a potentiometer or displayed on an oscilloscope. This approach allows for the direct simulation of dynamic systems, such as the flight of a V-2 rocket or the stresses on a bridge design.

Historical development

Early precursors include mechanical devices like the planimeter and the differential analyser, a landmark machine developed by Vannevar Bush at the MIT in the 1930s. The Second World War accelerated development, with machines like the Heath Robinson and Colossus used at Bletchley Park, though these incorporated digital concepts. The post-war era saw the zenith of large-scale electronic analog computers, such as those built by General Electric and used for aerospace simulation at NASA and for designing the Apollo guidance systems. Institutions like the UCLA and the Royal Aircraft Establishment were significant centers of research and application during this period.

Types and examples

Analog computers are broadly categorized by the physical medium used for computation. Mechanical analogs included the Monte Carlo simulation tools and various naval fire-control computers like the Mark I Fire Control Computer. Electronic analogs, the most common, ranged from small educational devices to room-sized hybrid computers that combined analog and digital elements, such as the EAI Pacer. Special-purpose examples are numerous, including the DRAKON visual language for the Buran spacecraft and the Water integrator, a fluid-based computer used in the Soviet Union. Other notable machines were developed by companies like Applied Dynamics and Telefunken.

Comparison with digital computers

The primary distinction lies in representation: analog computers use continuous signals, while digital computers like the ENIAC or modern microprocessor-based systems use discrete binary digits. Analog machines excel at real-time simulation and solving complex differential equations but suffer from lower precision and reprogramming difficulty. Digital computers, championed by figures like John von Neumann, offer high precision, stored-program flexibility, and are underpinned by Boolean algebra. The debate between the paradigms was prominent in the mid-20th century, with events like the Moore School Lectures helping to establish digital dominance for general-purpose computing by the 1970s.

Modern applications and revival

While largely supplanted for general computation, analog principles remain vital in specialized domains. They are embedded within mixed-signal integrated circuits for tasks like radio frequency processing and sensor conditioning. There is a growing research interest in novel analog computing for neural network acceleration and quantum simulation, explored at institutions like the Georgia Tech and in the European Union's FET projects. The field of neuromorphic engineering, pursued by companies like Intel with its Loihi chip, draws direct inspiration from the parallel, continuous-time processing of analog systems.