Analog computer analysis of Markov random processes in linear and nonlinear systems

Conclusions

The use of an analog computer to solve boundary value problems for one-dimensional diffusion equations with either constant or variable coefficient depending on the coordinate has been considered. It has been shown with actual examples that the error of the finite-difference solution is less than 5% for a rather small number of operational amplifiers (from 10 to 20). In the cases considered the analog computers are superior to digital machines not only in convergence but also in the speed of obtaining a solution. Moreover, the programming requires less time and the equations can be modified with little effort.