Probability Concepts

in Engineering
Planning and Design
VOL UME II DECISION, RISK,
AND RELIABILITY

ALFREDO H-S. ANG

Professor of Civil Engineering
University of Illinois at Urbana-Champaign

WILSON H. TANG
Professor of Civil Engineering
University of Illinois at Urbana-Champaign

JOHN WILEY & SONS

New York • Chichester • Brisbane • Toronto • Singapore
contents

1. Introduction 1

1.1 Aims and Scope of Volume II

1.2 Essence and Emphasis 1

1.2.1 Decision Analysis 1
1.2.2 Markov, Queueing, and Availability Models 2
1.2.3 Statistical Theory of Extremes 3
1.2.4 Monte Carlo Simulation 3
1.2.5 Reliability and Reliability-Based Design 3
1.2.6 Systems Reliability 4

2. Decision Analysis 5

2.1 Introduction 5
2.1.1 Simple Risk-Decision Problems 5
2.1.2 Characteristics of a General Decision Problem 9

2.2 The Decision Model 10

2.2.1 Decision Tree 10
2.2.2 Decision Criteria 13
2.2.3 Decision Based on Existing Information-Prior Analysis 16
2.2.4 Decision with Additional Information-Terminal Analysis 25
2.2.5 Preposterior Analysis 28
2.2.6 Value of Information 30
2.2.7 Sensitivity of Decision to Error in Probability Estimation 32
2.2.8 More Examples 33

2.3 Applications in Sampling and Estimation 47

2.3.1 Bayes Point Estimator 47
2.3.2 Optimal Sample Size 50

2.4 Elementary Concepts of Utility Theory 56

2.4.1 Axioms of Utility Theory 56
2.4.2 Utility Function 57
2.4.3 Determination of Utility Values 60
2.4.4 Utility Function of Monetary Value 63
2.4.5 Utility Function of Other Variables 66
2.4.6 Maximum Expected Utility , 67
2.4.7 Common Types of Utility Functions 70
2.4.8 Sensitivity of Expected Utility to Form of Utility Function 73

vii
viii CONTENTS

2.5 Assessment of Probability Values 74

2.5.1 Bases of Probability Estimation 74
2.5.2 Subjective Probability 75
2.5.3 Subjective Distribution of a Random Variable 76

2.6 Decisions with Multiple Objectives 77

2.6.1 Weighted Objective Decision Analysis 78
2.6.2 Multiattribute Utility Approach 84

Problems for Chapter 2 96

3. Markov, Queueing, and Availability Models 112

3.1 The Markov Chain 112

3.1.1 Introduction 112
3.1.2 The Basic Model 112
3.1.3 State Probabilities 113
3.1.4 Steady State Probabilities 118
3.1.5 First Passage Probabilities 122
3.1.6 Recurrent, Transient, and Absorbing States 125
3.1.7 Random Time Between Stages 136
3.1.8 Continuous Parameter Markov Chain 138

3.2 Queueing Models

3.2.1 Introduction
3.2.2 Poisson Process Arrivals and Departures
3.2.3 Poisson Arrivals and Arbitrarily Distributed Service Time
3.2.4 Distribution of Waiting Time

3.3 Availability Problems

3.3.1 Introduction ,
3.3.2 Continuous Inspection-Repair Process
3.3.3 Inspection and Repairs at Regular Intervals
3.3.4 Deteriorating Systems

Problems for Chapter 3

4. Statistics of Extremes

4.1 Introduction
4.1.1 Engineering Significance of Extreme Values
4.1.2 Objective and Coverage

4.2 Probability Distribution of Extremes

4.2.1 Exact Distributions
 CONTENTS ix

4.2.2 Asymptotic Distributions 194

4.2.3 The Symmetry Principle 204

4.3 The Three Asymptotic Forms 206

4.3.1 Gumbel's Classification 206
4.3.2 The Type I Asymptotic Form 207
4.3.3 The Type II Asymptotic Form 216
4.3.4 The Type III Asymptotic Form 222
4.3.5 Convergence Criteria 228

4.4 Extremal Probability Papers 235

4.4.1 The Gumbel Extremal Probability Paper 235
4.4.2 The Logarithmic Extremal Probability Paper 238

4.5 Exceedance Probability 246

4.5.1 Distribution-Free Probability 246
4.5.2 Significance of Probability Distribution 249

4.6 Estimation of Extremal Parameters 255

4.6.1 Remarks on Extreme Value Estimation 255
4.6.2 The Method of Moments 256
4.6.3 The Method of Order Statistics 256
4.6.4 Parameters of Other Asymptotic Distributions 261
4.6.5 Graphical Method 266

Problems for Chapter 4 268

5. Monte Carlo Simulation 274

5.1 Introduction 274

5.1.1 An Example of Monte Carlo Simulation 275

5.2 Generation of Random Numbers 278

5.2.1 Random Numbers with Standard Uniform Distribution 279
5.2.2 Continuous Random Variables 281
5.2.3 Discrete Random Variables 288
5.2.4 Generation of Jointly Distributed Random Numbers 290
5.2.5 Error Associated with Sample Size 291

5.3 Variance Reduction Techniques 292

5.3.1 Antithetic Variates 292
5.3.2 Correlated Sampling 296
5.3.3 Control Variates 298

5.4 Application Examples 301

Problems for Chapter 5 326

X CONTENTS

6. Reliability and Reliability-Based Design 333

6.1 Reliability of Engineered Systems 333

6.2 Analysis and Assessment of Reliability 334

6.2.1 Basic Problem 334
6.2.2 Second-Moment Formulation 340
6.2.3 Linear Performance Functions 347
6.2.4 Nonlinear Performance Functions 359

6.3 Modeling and Analysis of Uncertainty 383

6.3.1 Sources and Types of Uncertainty 384
6.3.2 Quantification of Uncertainty Measures 387
6.3.3 Analysis of Uncertainty 390
6.3.4 An Alternative Representation of Uncertainty 406

6.4 Probability-Based Design Criteria 412

6.4.1 On Design and Design Criteria 412
6.4.2 Second-Moment Criteria 414

Problems for Chapter 6 434

7. Systems Reliability 448

7.1 Introduction 448

7.2 Multiple Failure Modes 448

7.2.1 Probability Bounds 450

7.3 Redundant and Nonredundant Systems 471

7.3.1 Series Systems 473
7.3.2 Parallel Systems 474
7.3.3 Combined Series-Parallel Systems 478

7.4 Fault Tree, Event Tree Analyses 485

7.4.1 The Fault Tree Diagram 486
7.4.2 Probability Evaluation 495
7.4.3 Event Tree Analysis 498

7.5 Approximate Methods 504

7.5.1 The PNET Method 505

Problems for Chapter 7 517

 CONTENTS xi

Appendix A—Tables 531

Table A.I Values of the Gamma Function 532

Table A.2 CDF of the Standard Extremal Variate S 533

Table A.3 The Three Types of Asymptotic Extremal Distributions 537

Table A.4 Weights a, and bt for Leiblein's Order Statistics Estimator 538

Table A.5 Relations Among Parameters of the Type III Asymptotic

Distribution 539

Appendix B—Transformation of Nonnormal Variates

to Independent Normal Variates 540

B.I The Rosenblatt Transformation 540

B.2 Determination of Safety Index 544

References 549

Index 559