Digital computer simulation of an electronic analog computer

Abstract

In recent years there has been renewed interest in analog computation to simulate dynamic physical systems. This paper sought to combine the accuracy and reliability of digital computers, with the benefits of parallel computing offered by electronic analog computers. One of the benefits of simulating an electronic analog computer is that a one-to-one correspondence has already been established between the building block components of an electronic analog computer and common operations used in differential equations. Historically, electronic analog computers were limited by physical components which needed constant calibration due to drift and loading effects. This study sought to determine whether a single-processor digital computer, using a serial program with procedures that replicate analog components, could simulate this parallel environment. We predicted that if the serial program proceeds in small time increments, with the value of all components updated during each increment, then this process would act as a parallel program. We also predicted that replacing physical components with digital procedures would eliminate the effects of drift and loading. The methodology used a computer program that graphically and procedurally replicates the electronic components and layout of an electronic analog computer. The results were compared against differential equations for the physical system, since physical electronic analog computers are not readily available. Findings indicate that with an increment of 0.01 second or smaller, the digital program accurately and smoothly simulates a dynamic physical system by solving the descriptive differential equation(s) in a parallel, analog fashion, without the need for digital component calibration.