This study suggests a method for performance evaluation in a multiprogramming environment. The joint effect of two major cost elements - the processing time and the memory space - have been investigated. This involves two issues; 1) optimal memory loading and sizing, and 2) central site timing analysis. The latter has been approached mathematically with a cyclic queuing model; the memory space utilization was studied with a linear optimization model. Varying the values of the major cost carrying factors, through iteration, the authors investigate marginal utilities and search for the concept of an optimally balanced system.