Dynamic Programming and Hybrid Computation

Abstract

Dynamic programming finds its use in operations research, chem ical engineering, control systems, and other fields of applied mathematics; and is primarily concerned with the optimisation of multistage decision processes. Dynamic programming will split up the problem into a succession of N one-dimensional optimisation problems. The optimising action takes place at each discrete stage of the process.

The technique is based on the so-called principle, of opti mality, as expressed by Bellmann.

The paper describes the various techniques involved.

A hybrid analogue/digital computer will serve as an ideal tool, having the ability to integrate differential systems of equations, and also to perform the logical operations required in connec tion with the optimisation of the return function.