The report describes a method of formal conversion of address algorithms, with the goal of reducing the initial address representation and thereby the memory space required. The theorem of the possibility under certain conditions of compressing individual groups of addresses used in the address algorithm is proven. The method suggested is illustrated by an example of sequential transformations of a program for multiplication of a matrix by a vector.
Edited trans. of mono. Kibernetika i Tekhnika Vychislenii, Kiev, 1964 p84-93.