MATHEMATICAL MODEL FOR COMPUTER SYSTEM CONFIGURATION PLANNING.

This paper determines the speed of the devices constituting a computer system which will maximize system throughput given a fixed budget. The system composed of a central processor and peripheral devices is modeled as a closed queueing network of exponential servers. The problem is formulated as a nonlinear optimization problem where the objective is to maximize throughput subject to a nonlinear cost constraint. Individual device costs are represented by continuous power functions of device speed. Main memory costs are considered in terms of the degree of multiprogramming. The Lagrange multiplier technique is employed to set up a system of simultaneous nonlinear equations which are solved to produce the optimal solution.