The flow deviation method: An approach to store‐and‐forward communication network design

Abstract
Two problems relevant to the design of a store‐and‐forward communication network (the message routing problem and the channel capacity assignment problem) are formulated and are recognized to be essentially non‐linear, unconstrained multicommodity (m.c.) flow problems. A “Flow Deviation” (FD) method for the solution of these non‐linear, unconstrained m.c. flow problems is described which is quite similar to the gradient method for functions of continuous variables; here the concept of gradient is replaced by the concept of “shortest route” flow. As in the gradient method, the application of successive flow deviations leads to local minima. Finally, two interesting applications of the FD method to the design of the ARPA Computer Network are discussed.

This publication has 20 references indexed in Scilit: