Mathematical Theory of ReliabilityMathematical Theory of Reliability fgSIAM’s Classics in Applied Mathematics series consists of books that were previously allowed to go out of print. These books are republished by SIAM as a professional service because they continue to be important resources for mathematical scientists. Editor-in-Chief Robert E. O'Malley, Jr., University of Washington Editorial Board Richard A. Brualdi, University of Wisconsin-Madison Herbert B. Keller, California Institute of Technology Andrzej Z. Manitius, George Mason University Ingram Olkin, Stanford University Ferdinand Verhulst, Mathematisch Institwut, University of Utrecht Classics in Applied Mathematics C.C. Lin and L. A. Segel, Mathematics Applied to Deterministic Problems in the Natural Sciences Johan G, F Belinfante and Bernard Kolman, A Survey of Lie Groups and Lie Algebras with Applications and Computational Methods James M. Ortega, Numerical Analysis: A Second Course Anthony V. Fiacco and Garth P. McCormick, Nonlinear Programming: Sequential Unconstrained Minimization Techniques F.H. Clarke, Optimization and Nonsmouth Analysis George F. Carrier and Carl E. Pearson, Ordinary Differential Equations Leo Breiman, Probability R. Bellman and G. M. Wing, An Introduction to Invariant Imbedding Abraham Berman and Robert J. Plemmons, Nonnegative Matrices in the Mathematical Sciences Olvi L. Mangasarian, Nonlinear Progranuming *Carl Friedrich Gauss, Theory of the Combination of Observations Least Subject to Errors: Part One, Part Two, Supplement. Translated by G. W. Stewart Richard Bellman, Introduction to Matrix Analysis ULM. Ascher, R. M. M. Maitheij, and R. D. Russell, Numerical Solution of Boundary Value Problems for Ordinary Differential Equations K. E. Brenan, S. L. Campbell, and L. R. Petzold, Numerical Solution of Initial-Value Problems in Differential-Algebraic Equations Charles L. Lawson and Richard J. Hanson, Solving Least Squares Problems J. E. Dennis, Jr. and Robert B. Schnabel, Numerical Methods for Unconstrained Optimization and Nonlinear Equations Richard E. Bartow and Prank Proschan, Mathematical Theory of Reliability Cornelius Lanczos, Linear Differential Operators Richard Bellman, Introduction to Matrix Analysis, Second Edition Beresford N. Parlett, The Symmetric Eigenvalue Problem Richard Haberman, Mathematical Models: Mechanical Vibrations, Population Dynamics, and Traffic Flow *First time in print.Classics in Applied Mathematics (continued) Peter W. M. John, Statistical Design and Analysis of Experiments Tamer Bagar and Geert Jan Olsder, Dynamic Noncooperative Game Theory, Second Edition Emanuel Parzen, Stochastic Processes Petar Kokotovié, Hassan K. Khalil, and John O'Reilly, Singular Perturbation Methods in Control: Analysis and Design Jean Dickinson Gibbons, Ingram Olkin, and Milton Sobel, Selecting and Ordering Populations: A New Statistical Methodology James A. Murdock, Perturbations: Theory and Methods Ivar Ekeland and Roger Témam, Convex Analysis and Variational Problems Ivar Stakgold, Boundary Value Problems of Mathematical Physics, Volumes I and IT J. M. Ortega and W. C. Rheinboldt, Iterative Solution of Nonlinear Equations in Several Variables David Kinderlehrer and Guido Stampacchia, An Introduction to Variational Inequalities and Their Applications FE Natterer, The Mathematics of Computerized Tomography Avinash ©. Kak and Malcolm Slancy, Principles of Computerized Tomographic Imaging R. Wong, Asymptotic Approximations of Integrals ©. Axelsson and V. A. Barker, Finite Element Solution of Boundary Value Problems: Theory and Computation David R. Brillinger, Time Series: Data Analysis and Theory Joel N. Franklin, Methods of Mathematical Economics: Linear and Nonlinear Programming, Fixed-Point Theorems Philip Hartman, Ordinary Differential Equations, Second Edition Michael D. Intriligator, Mathematical Optimization and Economic Theory Philippe G. Ciarlet, The Finite Element Method for Elliptic Problems Jane K. Cullum and Ralph A. Willoughby, Lanczos Algorithms for Large Symmetric Eigenvalue Computations, Vol. I: Theory M. Vidyasagar, Nonlinear Systems Analysis, Second Edition Robert Martheij and Jaap Molenaar, Ordinary Differential Equations in Theory and Practice Shanti S. Gupta and S. Panchapakesan, Multiple Decision Procedures: Theory and Methodology of Selecting and Ranking Populations Eugene L. Allgower and Kurt Georg, Introduction to Numerical Continuation Methods Leah Edelstein-Keshet, Mathematical Models in Biology Heinz-Otto Kreiss and Jens Lorenz, Initial-Boundary Value Problems and the Navier- Stokes Equations J. L. Hodges, Jr. and E. L. Lehmann, Basic Concepts of Probability and Statistics, Second Edition George FE Carrier, Max Krook, and Carl E. Pearson, Functions of a Complex Variable: Theory and TechniqueThis page intentionally left blankGS Mathematical Theory of Reliability Sh Richard E. Barlow University of California Berkeley, California Frank Proschan Florida State University Tallahassee, Florida with contributions by Larry C. Hunter Sylvania Electronic Defense Laboratories Mountain View, California . Slam Society for Industrial and Applied Mathematics PhiladelphiaCopyright © 1996 by the Society for Industrial and Applied Mathematics. This SIAM edition is an unabridged republication of the work first published by John Wiley & Sons, Inc., New York, NY, 1965. 109876543 Alll rights reserved. Printed in the United States of America. No part of this book may be reproduced, stored, or transmitted in any manner without the written permission of the publisher. For information, write to the Society for Industrial and Applied Mathematics, 3600 University City Science Center, Philadelphia, PA 19104-2688. Library of Congress Cataloging-in-Publication Data Barlow, Richard E. Mathematical theory of reliability / Richard E. Barlow, Frank Proschan with contributions by Larry C. Hunter. p. cm. — (Classics in applied mathematics ; 17) Includes bibliographical references and index. ISBN 0-89871-369-2 (pbk.) 1. Reliability (Engineering)—Mathematical models. 2. Mathematical statistics. 3. Probabilities. L Proschan, Frank, 1921-. 1 Hunter, Larry C. IIL. Title. IV. Series. TA169.B37 1996 620'.00452'015118—dc20 96-13952 SIAM is a registered trademark,Contents Preface to the Classics Edition xi Preface xiii Acknowledgments xvii Chapter 1 Introduction I 1. Historical Background of the Mathematical Theory of Reliability, 1 2. Definitions of Reliability, 5 Chapter 2 Failure Distributions 9 1. Introduction, 9 2. Typical Failure Laws, 12 3. The Exponential as the Failure Law of Complex Equipment, 18 Monotone Failure Rates, 22 Preservation of Monotone Failure Rate, 35 Additional Inequalities, 39 General Failure Rates, 41 NAWS Chapter 3. Operating Characteristics of Maintenance Policies 46 1. Introduction, 46 2. Renewal Theory, 48 3. Replacement Based on Age, 61 4, Comparison of Age and Block Replacement Policies, 67 5. Random Replacement, 72 6. Repair of a Single Unit, 74 viiVili Chapter 4 Chapter 5 Chapter 6 Chapter 7 CONTENTS Optimum Maintenance Policies 1. 2. 3. Introduction, 84 Replacement Policies, 85 Inspection Policies, 107 Stochastic Models for Complex Systems L 2. 3. 4. 5 Introduction, 119 Markov Chains and Semi-Markov Processes, 121 Repairman Problems, 139 Marginal Checking, 151 Optimal Maintenance Policies under Markovian Deterioration, 156 Redundancy Optimization 1. 2. Introduction, 162 Optimal Allocation of Redundancy Subject to Constraints, 163 Application to Parallel Redundancy Model, 170 Application to Standby Redundancy Model, 175 Complete Families of Undominated Alloca- tions, 180 Optimal Redundancy Assuming Two Types of Failure, 185 Qualitative Relationships for Multicomponent Structures Penner “ Introduction, 198 Achieving Reliable Relay Circuits, 199 Monotonic Structures, 204 S-shaped Reliability Functions for Monotonic Structures, 211 k-out-of-n Structures, 218 Relationship between Structure Failure Rate and Component Failure Rates, 221 84 LI9 162 198CONTENTS ix Appendix 1 Total Positivity 229 Appendix 2 Test for Increasing Failure Rate 234 Appendix 3 Tables Giving Bounds on Distributions with Monotone Failure Rate 238 References 243 Index 251This page intentionally left blankPreface to the Classics Edition Continuing interest in the subject of reliability and the heretofore unavailability of our book Mathematical Theory of Reliability have encouraged publication of this SLAM Classics edition. We have not revised the original version, although much has transpired since its original publication in 1965. Although many contemporary reliability books are now available, few provide as mathematically rigorous a treatment of the required probability background as this one. The book falls into two parts. The first part, which ends with Chapter 4, is essentially concerned with component reliability, while the second part concerns system reliability. Were the book to be done over, radical revision of the first part would not be required. However, a deeper derivation of the exponential distribution based on modern inference concepts would be included. This in turn would be supplemented by the line of argument centering on the idea that probability is essentially an analytical tool based on judgment useful for making decisions. In this way, there would be no need to separate probability models and the statistical tools required for the analysis of failure data. As it stands, the reader must look elsewhere for relevant statistical procedures for failure data analysis. Appendix 2, concerning tests for increasing failure rate, would be removed, as it was in the 1969 Russian edition. Chapters 5, 6, and 7 deal with system reliability problems that are as important today as they were in the 1960s. The only drawback to this work is that it was perhaps based too closely on electronic systems common at that time and mechanical failure modes as well as human error. Today, system reliability evaluation is usually preceded by a careful failure-oriented system review using fault tree and failure modes and effects analysis (FMEA). This approach, although less mathematical, is critical for credible system reliability analysis and would have to be included in any serious revision. Richard E. Barlow Frank Proschan xiThis page intentionally left blankPreface What is mathematical reliability theory? Generally speaking, it is a body of ideas, mathematical models, and methods directed toward the solution of problems in predicting, estimating, or optimizing the proba- bility of survival, mean life, or, more generally, life distribution of com- ponents or systems; other problems considered in reliability theory are those involving the probability of proper functioning of the system at either a specified or an arbitrary time, or the proportion of time the system is functioning properly. In a large class of reliability situations, mainte- nance, such as replacement, repair, or inspection, may be performed, so that the solution of the reliability problem may influence decisions con- cerning maintenance policies to be followed. Since reliability theory is mainly concerned with probabilities, mean values, probability distributions, etc., it might be argued that the theory is simply an application of standard probability theory and really deserves no special treatment. This view would be as shortsighted as arguing that probability theory itself is simply an application of standard mathematical theory and deserves no special development. Reliability problems have a structure of their own and have stimulated the development of new areas in probability theory itself. We shall see this most clearly in developing and applying the concept of monotone failure rates and in obtaining new results in renewal theory as a result of comparing competing replacement policies. The monograph presents a survey of some mathematical models useful in solving reliability problems. The models are probabilistic; that is, from information concerning the underlying distributions, such as for component lives, deductions are made concerning system life, optimum system design, etc. Statistical problems, such as the estimation of xiiixiv PREFACE component or system life from observed data, are generally not treated. A larger and possibly more useful volume could easily be devoted to statis- tical problems in reliability. We have, however, chosen to confine our- selves to the more developed subject of probabilistic models in reliability theory. Throughout the book the emphasis is on making minimal assumptions, and only those based on plausible physical considerations, so that the resulting mathematical deductions may be safely made about a large variety of commonly occurring reliability situations. Thus quite often we assume only that the life distributions have an increasing (decreasing) failure rate and that only a single parameter, such as the mean, is actually known. (Throughout the book we write “increasing” in place of “non- decreasing” and ‘‘decreasing” in place of ‘“‘nonincreasing.”) From these modest assumptions a surprising number of useful deductions can be made. This theme of monotone failure rate is one of the unifying influ- ences in our attempt to integrate the existing reliability theory. In selecting material we must admit we have favored reliability models we have worked on ourselves. This method of selection may not be com- pletely disastrous, for in originally turning our attention to these particular models we were influenced by their importance to reliability theory and by their mathematical interest. We are painfully aware of the resulting gaps and omissions, however. If the omissions are too egregious, we hope an irate reader will write a monograph completing the picture! To whom is the book directed? As is true of all the books in this series, this monograph is addressed primarily to the applied mathematician. We hope the monograph will be of interest not only to those already working on reliability problems, but also to scientists and engineers who are working in other areas but are interested in seeing some applications of probability theory in new areas. The mathematical subject chiefly used in the mono- graph is probability theory and, to a limited degree, the concept of total positivity. An appendix is provided to develop the theory of total positivity required. In the introductory chapter we present a historical survey of the subject ; such a survey we feel is greatly needed, for in the development of reliability theory, research has all too often been undertaken and published without an awareness or use of what has gone on before. The chapter is completed by constructing a single general mathematical model which by appropriate specialization yields the basic quantities of interest in reliability theory. Chapter 2 surveys probability distributions of interest in reliability. The point of view is to develop the properties not just of specific distributions, but of broad classes of distributions characterized by plausible physical assumptions. This point of view motivates the development of thePREFACE xv properties of the class of distributions characterized by an increasing failure rate, corresponding to the physical phenomenon of wear-out. From this modest assumption applying in so many actual situations, an amazingly rich set of consequences ensues. Chapters 3 and 4 are devoted to maintenance models; Chapter 3 is concerned with the stochastic aspects of various maintenance policies and Chapter 4 with the optimization of these policies. In Chapter 3 an unex- pected bonus is obtained; by comparing two maintenance policies, some useful bounds on the renewal function are obtained. For distributions with increasing failure rate with only the mean known, the use of the bounds permits determination of the renewal function within an error of one-half. In Chapter 4 optimum stationary and time-dependent policies are obtained or their characteristics described for a number of classes of replacement, repair, and inspection models. The assumption of an in- creasing (decreasing) failure rate will be seen to play a crucial role in simplifying the solution and in leading to useful qualitative results con- cerning the form of the optimum policy. Chapter 5 presents stochastic models for complex systems. The key probability tools are Markov and semi-Markov processes. By using these tools, optimal control rules are developed, repairman-type problems solved, and monitoring and marginal testing policies analyzed. Chapters 6 and 7 are devoted to models involving redundancy. Chapter 7 shows how to allocate redundancy among the various subsystems under linear constraints on weight, volume, cost, etc., so as to maximize system reliability. The model is applicable in both system design and spares provisioning. Chapter 7 develops qualitative relationships between system reliability and component reliabilities. As an example, conditions are presented under which system life possesses an increasing failure rate when each component life possesses an increasing failure rate. Finally, appendices are provided summarizing special topics needed in the development of the theory. Thus an appendix on total positivity is presented and shows, among other things, how the concept of an increasing failure rate falls into the more general framework of total positivity. Because the assumption of increasing failure rate is so basic in much of the theory, a statistical test is presented in a second appendix for deter- mining from a sample of observations whether the underlying population does have an increasing failure rate. This test is the one exception to the tule followed throughout the book of confining attention to probabilistic questions rather than statistical ones. Tables are presented in the final appendix giving bounds on survival probability from a knowledge of either one or two moments of an increasing (decreasing) failure rate dis- tribution.