The representation of large numbers in neural networks and its application to economical load dispatching of electric power

The fundamentals of electric power load dispatching and Hopfield networks are briefly reviewed. A method for representing large numbers in Hopfield networks is theoretically examined and then applied to the problem of economical load dispatching. Simulation of stationary and time-varying loads is discussed, and results are given. The approach is found to be superior in the ease of formalization of the problem and in its memory and computation time efficiency. It is applicable to many problems other than economical load dispatching.<>