Optimization of Life Support Systems Reliability by an Integer Programming Method

Abstract
The problem considered is to minimize the weight of the subsystems of a life support system subject to several separable nonlinear constraints while maintaining an acceptable level of reliability of the system. Zero-one integer programming is used to solve this problem. The subsystems of the life support system designed by the Space Division of the North American Rockwell Corporation are studied examples. They are the refrigerant circuit, the cryogenic oxygen circuit, the carbonization cell, and the water reclamation unit. The results obtained are compared with the original design proposed by the North American Rockwell Corporation. The comparison indicates that the use of the integer programming for determining the optimization reliability would result in an improved arrangement of the standby units.