ON SOME ASPECTS OF OPTIMIZATION OF RELIABILITY FUNCTION OF SERIES-PARALLEL NETWORK

Abstract

Reliability is one of the important measures of effect-iveness to be achieved in the change of new and improved techno-logical systems. Its achievements must be viewed as a process in itself and as a part of the total process in which size, cost and performance are equally important measures of syfem effectiveness. The total innovation process, therefore, may be thought of as three or many closely interrelated sub pro-cesses that may be aimed at cost, performance and reliability. For maximum overall effectiveness it is obvious that these sub processes must stand in balanced relationship to one an-other. For many systemr reliability will be the dominant measure of overall economic effectiveness. Therefore, attention must be focused heavily on the reliability subprocess and costs. The performance may be compromised if needed. This requires optimization of reliability function, taking care of cost, performance etc. as constraints. In Chapter three, an algorithm is developed to find reliability of a time-dependent system which is easy and requires less time. In Chapter four, mathematical model of the reliability problem is discussed. Only series system is considered for finding mathematical model. In Chapter five th<_ two techniques, "Interior point algorithm" and "Geometric Programming", are applied for maximizing the reliability of a system subject to linear constraints while Chapter six "Generalised -Lagrange's Multiplier", "Convex Programming", "Concave Programming", Lagrangian algorithm and Penalty function method dxs used for optimizing reliability of a series network of system subject to linear or nonlinear constraint. Except Generalised Lagrange's Multiplier technique, all other methods are new to the field of Reliability® All the methods are iterative and hence can be programm-ed on computer easily. At the end in the Appendix computer programmes in FORTRAN II for all methods are given.