aoa ‘conemy er OPTIMIZATION METHODS NANI Sa N.V.S. Raju 'OPTIMIZATION METHODS FOR ENGINEERS NV. Raju © 2014 by PHI Leaming Private Limited, Delhi. All rahts reserved. No part of this book may be reproduced in any form, by mimeograph or any other means, without permission in writing from the publisher. ISBN-978-81-203-4744-1 ‘The export rights of this book are vested solely with the publisher, Published by Asoke K. Ghosh, PHI Leaming Private Limited, Rimihim House, 111, Patparganj Industial Estate, Delhi-1 10002 and Printed by Rajkamal Electric Pross, Plot No. 2, Phase IV, HSIDC, Kundli-181028, Sonepat, Haryana.Contents Foreword xv Preface wit Acknowledgements xix 1. Optimization—An Overview Learning Objectives 1.1 Optimization J 1.2. Origin and Development of Optimization Methods 2 History and Development — 2 Chief Contributors 3 1.3 Mathematical Models 4 Based on Certainty 4 Based on Time Reference 4 Rased on Physical Appearance 5 Based on Purpose and Nature 5 Based on Method of Solution 5 1.4 Optimization Methods in Engineering—Modelling and Application 1.5. Characteristies of Optimization Models 6 1.6 Limitations of Optimization Models 7 1.7 Method of Optimization 8 Defining the Problem and Collecting the Relevant Data 8 Refining the Data and Pormulating the Problem 9 Icentilving and Fitting @ Suitsble Optimization Model 9 Analyzing and Deriving the Solutions from the Model 10 Validating the Model and Testing Its Solution 10 Sensitivity Analysis and Controlling the Solution 10 Decision-making and Implementing the Solution 10 1.8 Application in Engineering Areas Summary 12 Key Concepts 12 Review Exercises 13 Formulation of Optimization Problems Learning Objectives 15 2.1 Decision Variables or Design Vector 15 5 6 1-14 15-433. Contents 2.2 Objective Function 16 23 Design Constraints 17 ‘Types of Constraints 17 24 Boundary Conditions — 18 2.5 Formulation 19 2.6 Distinction between Linear Programming and Nonlinear Programming Problems 35 Summary 36 Key Concepts 36 Review Exercises. 36 Formulation and Computational Exercises. 37 Solutions by Graphical Methods for Optimization Problems 44-74 Learning Objectives 44 3.1 Strengths of Graphical Methods 44 3.2 Limitations (Weaknesses) of Graphical Methods 45 3.3. Graphical Solution Procedure (Algorithm) 45. Step I; Setting the Axes on a Graph Sheet 45 Step 2: Plotting the Constraint Set on a Graph Sheet 46 Step 3: Representing the Constraint Set 47 Step 4: Identifying the Feasible Region 47 Step 5: Searching for the Optimal Solution 48 Redundant Constraint 51 Types of Solutions 54 Solutions 55 Feasible Solutions 55 Basic Feasible Solutions Optimal Solutions — 50 3.6 ‘Types of Feasible Regions 56 3.7. Types of Optimal Solutions 58 Infeasible Solutions 58 Unique Solutions 58 Maltiple Optimal Solutions 59 Unbounded Solutions 60 Summary 69 Key Concepts 69 Review Exercises. 70 Formulation and Computational Exercises. 71 ae Nonlinear Programming Problems—Classical Optimization Techniques and Basie Concepts 75-92 Learning Objectives 75 4.1 Categorization of Classical Optimization 75 With Reference to the Methodology of Approach. 76 With Reference to the Number of Variables 76 With Reference to Constraints 76Contents Vi 4.2 Functions in Optimization Problems 77 Single Variable Function 7 Multivariable Function 78 4.3 Optimization Methods for Single Variable Functions 78 4.4 Maxima-Minima 78 Local and Global Maxima and Minima Values of Functions 4.5 Search Methods 83 4.6 Multivariable Functions 83 Some Important Definitions for Recapitulation 86 Summary 89 Key Concepts 90 Review Exercises 90 Formulation and Computational Exercises 91 ” Analytical One-dimensional (Single Variable) Unconstrained Optimization 93-113 Learning Objectives 93 3.1 The Concept 93 5.2. Necessary and Sufficient Conditions 94 5.3 Necessary and Sufficient Conditions for Single Variable Optimization a4 Necessary Conditions (Theorem 5.1) 95 Sufficiency Condition (Theorem 5.2) 97 54° Working Rule (Algorithm) for Single Variable Optimization 98 Summary 110 Key Concepts 110 Review Exercises 10 Formulation and Computational Exercises 111 Analytical Multidimensional (Multivariable) Unconstrained Optimization 114-131 Learning Objectives 114 6.1 Classification of Multivariable Optimization Problems 14 6.2 Optimization Techniques to Solve Unconstrained Multivariable Functions 1/5 Exhibit 6.1: Taylor's Theorem/Series Expansion 1/5 6.3 Necessary Condition (Theorem 6.1) 1!7 64 Sufficient Condition (Theorem 6.2) 117 6.5 Working Rule (Algorithm) for Unconstrained Multivariable Optimization Us Summary 128 Key Concepts. 128 Review Exercises 128 Formulation and Computational Exercises 129Vili Contents 7. Analytical Multidimensional Optimization with Equality Constraints 132-160 Learning Objectives 132 7.1 Multivariable Optimization with Equality Constraints 132 7.2. Solution Methods for Multidimensional Optimization with Equality Constraints 133 7.3. Direct Substitution Method 133 7A Constrained Variation Method 140 7.8 Lagrange Multiplicrs Method [49 Summary 156 Key Concepts. 157 Review Exercises 137 Formulation and Computational Exerteses. 158 8. Analytical Multidimensional Optimization with Inequality Constraints 161-180 Learning Objectives 161 8.1 Solution Method /6/ 8.2 Kuhn-Tucker Conditions for Solving Multivariable Inequality Constrained Problems 163 Summary 178 Key Concepts 178 Review Exercises 178 Formulation and Computational Exercises 179 9. Numerical Methods for One-dimensional Nonlinear Programming 181-214 Learning Objectives 181 9.1 The Philosophy — 181 9.2 General Method of Solution (The Working Rule: The Procedural Steps) 82 9.3 Classification 182 9.4 Distinction between Analytical and Numerical Methods 182 9.5 Unimodal Function 183 Unimodality Approach 184 9.6 Elimination Methods 184 The Concept — 184 9.7 Unrestricied Search 185 The Concept 185 ‘The Working Rule: The Procedural Steps 186 9. Exhaustive Search 188 The Concept 188 The Working Rule: The Procedural Steps 189 9.9 Dichotomous Search 190 The Concept 190 The Working Rule: Procedural Steps 192Contents ix 9.10 Interval Halving Method 194 The Concept ‘194 The Working Rule: Procedural Steps 195 9.11 Fibonacci Search 198 ‘The Concept 198 ‘The Working Rule; Procedural Steps 198 9.12 Golden Section Search Method 202 The Concept 203 ‘The Working Rule: The Procedural Steps 204 9.13 Interpolation Methods—Quadratic Interpolation 205 The Concept 205 The Working Rule: The Procedural Steps 206 9.14 Cubic Interpolation 208 The Concept 208 Working Rule: The Procedural Steps 209 Summary 211 Key Concepts 211 Review Exercises 212 Formulation and Computational Exercises. 213 10. Numerical Methods for Unconstrained Optimization of Multivariable Nonlinear Programming Problems 215-250 Learning Objectives 215 10.1 Classification — 216 10.2 Univariate Method 216 ‘The Concept 216 Working Rules: Procedural Steps 2/7 10.3 Hooke—leeves Pattern Search Method 221 The Concept 221 Working Rule: The Procedural Steps 222 10.4 Powell’s Pattern Search Method 225 Working Rules’ Procedural Steps (The Algorithm) 225 10.5 Indirect Search Methods 230 10.6 Cauchy’s Steepest Descent Method 232 Working Kules: Procedural Steps 252 10.7 Fletcher-Reeves Method 238 ‘The Working Rule: The Procedure 238 10.8 Newton's Method 241 The Working Rule: The Procedure 242 Summary 245 Key Concepts 247 Review Exercises. 247 Formulation and Computational Exercises. 248X Contents 11. Constrained Optimization Techniques for Nonlinear Programming Problems 251-279 Learning Objectives 251 11.1 Structure of Constrained Nonlinear Programming Problems 251 11.2 Factors Affecting the Solution of a Constrained Problem 252 Rifect of Constraints on Optimum Point 252 Infeasible Solution 253 Unique Solution — 253 ‘Multiple Solutions Induced by Constraints 253 Maltiple Solutions Induced by Objective Function 253 Unbounded Solution 253 11.3 Normalization of Constraints 253 11.4 Classification of Constrained Nonlinear Programming Problems 256 ‘Sequential Unconstrained Minimization (SUM) Techniques or ‘Transformation Metheds or Penalty Methods 257 11.5 Technique of Variable Transformation 259 11.6 The Penalty Function 262 Exterior Penalty Function Method Working Rule, The Procedural Steps 265 Interior Penaliy Function Method 270 Summary 274 Key Concepts. 274 Review Exercises. 27. Formulation and Computational Exercises 276 12. Pivotal Reduction Method for Linear Programming Problems — 280-302 Learning Objectives 280 12.1 Features of Linear Programming Problem 280 12.2 Structure of Linear Programming Problem 281 Scalar Form 281 Matrix Form or Vector Form 282 12.3 Classification of Solutions Methods for Linear Programming Problems 282 12.4 Pivot Reduction Method 283 Pivot Reduction of General Form to Obtain Canonical Form 283 12.5 Applications of Pivot Reduction Method 284 Consistency Check 285 Finding Solution of Consistent Square Set of Equations 285 Finding Solution of Consistent Non-square Set of Equations 285 12.6 Motivation to Simplex Method 292 12.7 Finding Optimal Solution by Examining All Possible Basic Solutions (With Equality and Inequality Constraints) 293 Summary 298 Key Concepts 298 Review Exercises 298 Formulation and Computational Exercises 299Contents Xi 13. Simplex Method for Linear Programming Problems 303-346 Learning Objectives 303 13.1 Simplex Algorithm (Working Rules and Procedure) 303 13.2 Use of Surplus and Artificial Variables in Simplex Method 3/4 Significance of Artificial Varicble 315 13.3 Big-M Method 315 Algorithm or Working Rules (The Procedure): For Big-M or Penalty Method 316 13.4 Two-Phase Method 321 Algorithm of Two-Phase Method 52/ Merits of the Two-phase Method over the Big-M Method 324 13.5 Variations in Simplex Solutions 329 Unbounded Solution 329 Alternate or Multiple Optimal Solutions 335 Unique Solution 337 Infeasible Solution 338 Summary 341 Key Concepts 341 Review Exercises 342 Formulation and Computational Exercises 343 14. Degeneracy and Duality in Simplex 347-386 Learning Objectives 347 14.1 Degeneracy in Simplex Method 3.47 14.2. Degeneracy due to Beale’s Cycling 348 Resolution of Degeneracy for Beale’s Cycling Problem 348 14.3 Degeneracy due to Tie between Basic Variables for Leaving the Basis 348 Resolution of Degeneracy in Case of Tie between Leaving Variables 348 14.4 Tie for Incoming Variable (Key Column) 354 14.5 Unrestricted Variables in LPPs 354 14.6 Equality Consuains in LPPs 355 14.7 Duality in Simplex 358 The Concept 358 Shadow Prices 364 14.8, Working Rules (Procedural Steps) for Conversion of Primal to Dual or Dual to Primal 364 14.9 Comparison between Primal and Dual 377 With Reference to Formulation 377 With Reference to Solution 372 14.10 Advantages and Applications of Duality 372 Summary 376 Key Concepts 376 Review Exercises. 377 Formulation and Computational Exercises. 377 Additional Exercises in LPP for Practice 381xii Contents 15. Dynamic Programming I 387-426 Learning Objectives 387 15,1 Meaning and Definition 387 15.2. Bellman’s Principle of Optimality 388 The Concept 388 Analysis and Explanation 388 15.3 Jargon of Dynamic Programming Problems and Their Definitions 389 15.4 Characteristics of Dynamic Programming Problem 392 15.5 General Algorithm: Working Rules 392 15.6 Applications of Dynamic Programming Approach 93 15,7. The Stagecoach Problem or Routing Problem 394 15.8 Load Distribution or Cargo Loading Problem 399 15.9 Allocation/Assignment Problem in Logistics and Material Handling 401 15.10 Capital Budgeting or Budget Allocation Problems in Financial Management — 406 15.11 Employment Smoothing Problem in Human Resource Management 411 15.12 Selection of Alternatives—Advertising Media in Marketing Management 414 15,13 System Reliability Problem 417 Summary 419 Key Concepts 419 Review Exercises 419 Formulation and Computational Exercises 420 16. Dynamic Programming I 427-448 Learning Objectives 427 16.1 Optimal Sub-division Problem 427 16.2 Linear Programming Problem by Dynamic Programming Approach 437 16.3. Nonlinear Programming Problem by Dynamic Programming Approach 443 Summary 446 Key Concepts 446 Review Exercises 446 Formulation and Computational Exercises. 447 17. Simulation 449-480 Learning Objectives 449 17.1 The Concept of Simulation 449 17.2. Definitions 450Contents 17.3. Merits and Demerits of Simulation 457 Merits of Simulation over Other Conventional Optimization Techniques 451 Demerits (Limitations) of Simulation 4/52 Situations Warranting the Application of Simulation 452 17.4 Engincering and Managerial Applications of Simulation 453 Risk Profile and Investment Decisions. 453 ‘nsuring the Project Completion within the Timeline 453 Shipping and Physical Disibution Systems 433 Service Organizations 454 17.5 The Methodology and General Procedure of Simulation 454 17.6 Validation of a Simulation Model 457 Ensure Accuracy and Volidity 457 Validation Process 457 Validation, Verification and Calibration 458 Validating the Existing Systems 459 Validating the New (First-Time) Model 459 17.7 Types of Simulation Models 460 Based on the Method or Process 460 Based on Complexity and the Means for Simulsting 460 Based on Certainty 460 Based on Change with Time 460 178 Discrete Event Simulation or Event Type Simulation 461 17.9 Simulation Languages 469 Selection of a Discrete System Simulation Language 46? 17.10 SIMSCRIPT 470 Background 477 17.11 General Purpose Simulation System (GPSS) 471 Salient Features 472 17.12 SIMULA (SIMUlation LAnguage) 472 Salicnt Features 473 17.13 Comparison between SIMSCRIPT, GPSS and SIMULA 473 17.14 Debate about Using Simulation Models as Optimization Tools 474 Summary 476 Key Concepts. 476 Review Exercises. 477 Formulation and Computational Exercises 478 18. Monte Carlo Simulation (Simulation II) 481-515 Learning Objectives 481 18.1 Random Variables and Random Numbers 48/ 18.2 Characteristics of Random Numbers 482 18.3 Methods of Generating Random Numbers 482 Manual Generation of Random Numbers 482 Generation of Random Numbers by Computers 483 Table of Random Numbers #85XIV Contents 184 Test for Randomness 486 Law of Statistical Regularity 486 Law of Inertia of Large Numbers 486 18.5 Mid-square Method of Generating Pseudorandom Numbers 487 ‘The Algorithm (Working Rule) 457 Limitations of the Mid-square Method 488 18.6 Multiplicative Congruency (or Power Residue) Generator 489 ‘The Algorithm (Working Rule) 489 Limitations of Power Residue Method 490 18.7 Algorithm for Monte Carlo Simulation 493 188 Strengths and Weakness of Monte Carlo Simulation 494 Summary 509 Key Concepts 509 Review Exercises 510 Formulation and Computational Exercises Su Quiz/Objective Type Questions 517-587 Index 589-596Foreword Optimization Methods for Engineers is the most useful subject to the students of all branches of Engineering. This subject is derived from the concepts of Optimization and Operations Researches during World War-II. This subject has become an integral part of the syllabi in almost all Indian Universities in some form or other with various titles such as Engineering Optimization, Optimization Techniques, and Decision Analysis in Engineering, Advanced Operations Research etc. This subject is so important that everyone should own one good copy for reference and application any time. T rate this book entitled Optimization Methods for Engineers authored by Dr. N.S. Raju as one among the best books in this field. His rich experience in teaching as well as in industry is embedded in every page of this book for the benefit of even beginners of this subject. It is not only useful 10 the students to learn thoroughly, but also for the lecturers to teach this subject effectively ‘The contents in this book are rightly blended with sufficient theory and lot of illustrations with practical examples, which create students’ interest in the subject. For instance, to quote the examples given in “The formulation’ (Chapter 3), the examples are very much live to explain the nature of constraints and are quite interesting. Similarly, “time finding’ ‘examples for explaining the concepts of the search methods of One Dimensional Optimization (Chapters 9 and 10). Such examples taken from real-life situations makes the subject memorable There are many such examples given in every chapter. I appreciate the author's effort in his keen observation of getting examples from day-to-day life and relating them to the subject At certain places, remarks, notes to students, footnotes, instructions, precautions etc are appropriately given in various chapters. Particularly, to explain the simplex problem (Chapters 12 to 14), Dynamic Programming (Chapter 15 and 16) and Simulation (Chapters 17 and 18), the efforts made by the author are innovative and interesting. Similarly, the remarks under this problem are noteworthy. Such remarks show the author's long and rich teaching experience. Tam sure that the new entrants to this field as teachers or practitioners will be highly benefited by his experiences, It seems there is a treasure of problems available with the author. A wide variety of collections of problems asked in last 2 years to last 20 years are also found. At some places,XVI Foreword the problems might have been designed by the author with his experiences, where a desired variety problems is not available in examination papers. This had filled the gap. covering all the types of problems and for the benefit of aspiring readers, little more extensions are also given. It is an appreciable work for the benefit of those who want more, ‘The conceptual framework done in this book gives a good understanding of such topics to the students and recapitulate to memorize or review, particularly for pre-final and final year students, since they appear shortly for the competitive examinations such as GATE exams Dr. N.V.S, Raju laid strong foundation stones for knowledge building of the followers of this book and removed the entire phobia about the subject. There are no sufficient textbooks available in the field of optimization, particularly in non- linear programming, This book is a welcoming edition in the field and fills the gep to cater to the needs of present day students of all branches. In view of this, I strongly recommend this book for not only PG level but also for UG level students Qacsiichnona Dns mines RO Prof. RAMESHWAR RAO Vice-Chancellor ‘esau emis Tease Uae HyderabadPreface Science is unbounded and is not an overnight development. From the antiquity, the use of scientific methods and their applications have been improving the living standards and casing the jobs and hence curtailed the cost and enhanced the profits. This is the core concept of optimization. ‘The Objective Authoring of this book Optimization Methods for Engineers started with muhi-objective. The first objective is to explore, compile and systematically present the applications of science and technology for optimization of engineering aspects of any disciplines. The second objective is to bring this subject close to the student community since most of the students, although having more feasible to scoring, arc scared to take this course as their clective, (This subject is offered as an elective in many Indian Universities at B. Tech. level.) On preliminary inquiry the students attributed the reasons to the difficulty in understanding and the application of the mathematical concepts, clumsy and confusing terms, no sufficient number of the titles in this field and even in the available literature, unclear explanations that are, in fact, necessary for understanding the subject. Thus my third objective of authoring this book is famed as ‘to minimize the confusion and to maximize the understanding by presenting the concepts through everyday life examples or incidents” so that the students can think of applying the concepts to optimize the works of their daily life also. Further, step-by-step explanation (algorithmic approach) can make the faculty leam the subject well on their own to teach well, which is my final objective The Constraints In the book, about 200 Illustrations and more than 300 Practice Problems, along with University examination questions/problems are provided. 1 am constrained to extract these questions/ problems from major Universities of South India, such as JNTUs, OU and so on. However, these questions have a large resemblance with those appeared in almost all the universities of xviiXVII Freface India if compared. While I was writing this book, some students inferred that the large and lengthy chaptering is another major constraint to study’ the subject. Further. they wanted a book to provide exhaustive number of step-by-step solved illustrations but book must be small and handy to carry, it should present difficult concepts but in simpler way. So, this is resolved by optimally sub-dividing the subject into smaller chapters having about 20 pages so that the learners can feel the ease, comfort and increased confidence of learning. ‘The text is organized into 18 chapters. Chapter 1 provides an overview of optimization techniques which discusses the history, development and jargon. Chapter 2 introduces and familiarizes the students with formulation of optimization problems, The graphical solutions are portrayed in Chapter 3, whereas the analytical methods of nonlinear optimization are covered in Chapters 5 to 9. Chapter 4 deals with the classical optimization techniques, and Chapter 5 gives the single variable (one-dimensional) unconstrained optimization and Chapter 6 paves the way for solving multidimensional problems, The constrained optimization is illusirated in Chapter 7 whereas Chapter 8 is on equality and inequality constraints. Chapters 9 to 11 provide the numerical methods. Chapter 9 illustrates the single variable, whereas Chapter 10 deals with multidimensional unconstrained optimization and Chapter II depicts solutions for constrained optimization problems. Motivation to the simplex method and pivoial reduction method is given in Chapter 12. The simplex method is discussed thoroughly in Chapters 13 and 14. In a similar fashion, the dynamic programming is explained through Chapters 15 and 16, Chapters 17 and 18 provide an introduction of simulation, Subject to the available and requirement constraints in all the chapters are sculpted to their best forms with surplus number of review questions, In addition, Recap and Key Concepts at the end of each chapter would help the student for revision and last-minute revision to the examinations The Conditions Utmost care has been taken to make this book error-free. However, we take the responsibility for any mistakes that might have crept inadvertenily, We would be glad to receive any mistakes, feedback and comments. Also, any suggestions to improve this book are always unrestricted and would be gratefully acknowledged without any boundary conditions. N.V.S. RAJU [email protected]Acknowledgements At the outset, I express my deep sense of gratitude for inspiration, encouragement and support extended by + Dr. K. Narayana Rao, Prof. and Head, Dept. of Mechanical Engincering, O.U. Hyderabad + Dr. CB. Krishna Murthy Prof. (Retd) in Mechanical Engg., JNTU + Prof. Rameshwar Rao, Prof. of ECE and Hon'ble Vice Chancellor, JNTU, Hyderabad + Prof G. Tulasi Ram Das, Prof. of EEE and Hon'ble Vice Chancellor, NTU, Kakinada + Prof. N.V. Ramana Rao, Prof. of CE and Registrar, INTU, Hyderabad + Dr. A, Vinaya Babu, Prof, of CSE and Principal, JNTUHCER, former Director, Admissions and School of Continuing and Distance Education, INTU, Hyderabad + Prof. N.V. Ramana, Prof. of EEE, INTUHCE, Manthani + Dr. Eshwara Prasad Koorapati, Prof. of ME and former Principal, JNTUHCE! and Director of Evaluation, JNTUH + Dr. (Retd, Col.) K. Prabhakar Rao, Director, Raja Mahendra Group of Institutions Though not given by each individual name since the list is big, I remember and thank all my present colleagues and former colleagues, friends and students who helped me while authoring this book. 1, from the bottom of my heart, acknowledge the DTP work by my students—Mr. Nagendra, Mr. P. Srikant and Mr. P. Yashasvi, graphical work by my daughter, Ms. Srija Nataraj, the help extended by Mr. K. Sudheer, Mr. SSR Krishna and active follow-up and processing of Mr, Malaya Ranjan Parida on behalf of PHI Learning, without which this book would not have seen any light While authoring this book, I had to go through numerous books on optimization techniques and met many professors/practitioners. I am grateful to all those authors and individuals who directly or indirectly helped me in authoring the book. | feel sorry if I miss anybody inadvertently Last but not the least, I thank the co-operation and patience of my wife, Mrs. Prasanna, and our children Ms. Srija Nataraj, Ms, Himaja Padmapriya and Ms. Sanja Karunamayi. Also,XX Acknowledgements am indebted to my Parents Smt. and Sri N.V. Murali Manohar Raju Telugu Pandit (Retd.). in-laws Mrs, and Mr. C. Bhaskar Raju and Brothers Dr. NVSN Raju, Neo-Natalogist. USA and Mr. N.Y, Srinivasa Raju, C&F, Nippon Paints and also my sisters for their encouragement, Finally, I thank entire Kasu and Chennamadhavuni families for showering their good wishes on me. N.V.S. RAJU [email protected]Optimization An Overview Learning Objectives After studying thi s chapter, you should be able to: Explain the meaning of the term optimization. Understand the origin and development of optimization methods. Appreciate the structure of an optimization problem. Describe the various optimization model Apply optimization methods lo engineering problems. Understand the limitations of optimization methods. aaeene Life is riddled with problems. A successful person is the one who succeeds in finding the best solution and implements it at the right time. Among the several decisions that we take in our daily life, some decisions are based on logical reasoning, but most are merely intuitive. Intuition-based decisions sometimes mislead or confuse us. Also, such decisions may neither have any evidential support and nor have any logic or scientific backup. Hence, it is necessary for managers and engineers to resort to scientific methods and techniques that provide the evidential support and logical base. Optimization is one such applied science that provides managers. engineers, researchers, and practitioners with better solutions to problems. 1.1 Optimization The term optimization means obtaining the best result (or output) under the given circumstances (or inputs). It is the most significant act of decision-making during design. manufacturing, maintenance. construction and/or any engineering and management activity. We come across ‘two major requirements of optimization methods. (1) how to minimize the effort required and (2) how to maximize the benefit desired?2 optimization Methods for Engineers If the effort required or the benefit desired for a practical situation can be quantized and expressed as a function of a certain variable based upon which the decision has to be taken, then the problem is said to be that of optimization, Hence, optimization may be defined as the process of finding the best result in the form of minimizing the effort required (cost function) ‘or maximizing the benefit desired (profit function), expressed in the form of a function of decision variables under certain constraints and/or under given conditions. If the benefit desired or the effort required is represented as the function /ix), then the optimization problem is represented as ma ize f(x), ie. to find the point x = where /*() shows the highest value of /(x), and minimize (2), ie. to find the point x = x* where /*(x) shows the lowest value of f(x). 1.2 0 ation Methods ‘in and Development of Opti The optimization methods seeking the optimum values need mathematical modelling or programming, The branch of science concerned with the application of models, methods and techniques to aid decision-making in establishing the best result is called operations research (OR). Various managerial and simple engineering problems comprising the linear relationships between variables, logical methods and mathematical models are dealt within OR. But the problems in this ficld with nonlincar relationships require higher mathematical models and hence are studied as part of optimization techniques. History and Development ‘The first form of optimization methods is operations research. The term operations research was first coined by McCloskey and Trefethen in 1940 in Bowdsey, a small town in the United Kingdom. This science came into existence during the Second World War. ‘The subject “operations research’ was so called because it started with research of (military) operations. During the Second World War, the military commands of the United Kingdom and the United States of America employed several teams of scientists to uncover logical, tactical and strategic military operations, using mathematical, statistical and scientific concepts. Their objective was to formulate specific proposals to arrive at the decisions that could optimize the utilization of scarce resources to acquire maximum possible level of effective results, In other words, it was to find the methods that could yield the best results with minimum efforts. Thus, OR gained popularity and was called “an art of winning the war without actually fighting it After the end of the Second World War, the encouraging results of the UK teams attracted industrial managers to apply these methods to solve their managerial problems. The most significant work in this direction was perhaps done by G.B. Dantzig of USA, in 1947 through his simplex method applied to linear programming problems (LPPs). Since then many scientists have been expanding the bounds of OR and developing new dimensions to this science in theOptimization—An Overview 3 interest of making operations yield high profits at lowest costs. Today, the optimization tools have become universally applicable in many areas such as construction, transportation, medicine, agriculture, library, town planning, financial planing, and so forth Chief Contributors ‘The term optimization can be traced to the works of Ramanujan, Newton, Lagrange, Cauchy Leibnitz, and so on. It came into being during the development of the differential calculus methods for optimization. The foundations of calculus of variations, which deals with minimization of functions, were laid with the contributions of Bernoulli, Euler, Lagrange, and Weirstrass. The method of additions of unknown multipliers for optimizing constrained problems by the name of its inventor, Lagrange, and Cauchy's application of the steepest descent method to solve unconstrained minimization problems are noteworthy. However, progress was very little until the discovery of high speed digital computers during the mid-twentieth century, The developments in this field resulted in the emergence of several well-defined new areas of optimization methods. ‘The simplex method developed by Dantzig in 1947 for lincar programming problems (LPPs) and the principle of optimality developed by Richard Bellman in 1957 for dynamic programming problems (DPPs), are the two important milestones in the development of methods of constrained optimization. Contemporarily, the concept of necessary and sufficient conditions for the optimal solutions of programming problems by Kuhn and Tucker in 1951 is another landmark in the development of nonlinear programming problems (NLPPs). The major developments in the area of numerical methods of unconstrained optimization were made in the United Kingdom during the 1960s, The contributions of Zoutendijk, Rosen, Carroli and Fiacco and MeCormick are particularly important, In the 1960s, Duffin, Zener and Peterson developed geometric programming (GP). Gomory did a pioneering work in integer programming. Dantzig, and Chames and Cooper developed stochastic programming techniques, assuming the design parameters to be independent and normally distributed. The work of Charnes and Cooper in 1961 to optimize linear problems with more than one objective (or goal) while satisfying the physical constraints, paved the way for developing goal programming (or multi-objective programming) methods. The game theory proposed by von Neumann in 1928 has been applied to solve several mathematical, economic and military problems. Although no single technique is universally applicable to solve NLPPs, various optimization methods mentioned above have allowed solving many difficult problems During the last decade of the twentieth century, simulated annealing (SA), genetic algorithm (GA) and neural network (NN) methods played a key role and created a new class of mathematical programming techniques, Simulated annealing (SA) is analogous to the physical process of annealing of solids, whereas genetic algorithm (GA) is a search technique based on the mechanics of natural genetics. Neural networks (NNs) are techniques with efficient computing power of the network of interconnected “NEURON” processors. Some contributions related to this subject are summarized in Table 1.14 Optimization Methods for Engineers ‘Table 1.1 Chief Contributors of Optimization Methods Period ‘Optimization Method Chief Contributors 1928 Game theory yon Neumann 1940s Mean value theorems Lagrange. Taylor. Cauchy 1940s Differential caleulus methods Newton, Lagrange, Cauchy 1940 Operations research (basics) McCloskey and Trefethen 1947 Simplex method and linear programming GB. Danizig 1960s Big-M (penalty) simplex method Dantzig, Chames 1960s Variations in calculus Bemoulli, Buler, Lagrange, and Weirstrass 1951 Necessary and sufficient conditions Kuhn and Tucker 1960s Unconstrained optimization (numerical) Zoutendijk, Rosen, Carroli and MeCormick 1957 Principle of optimality and dynamic programming problems Richard Bellman 1960s Geometric programming Duffin, Zener and Peterson 1960s Integer programming Gomory 1961 Goal (multi-objective) programming Charnes and Cooper 1990s Simulated annealing (SA), genetic algorithm (GA), neural — networks 1.3. Mathematical Models A model is a representation of a real-life situation. A model can be studied by classifying it in many ways. Various types of mathematical models are discussed and distinguished here as follows. Based on Certainty Deterministic model: A model in which all the parameters of decision variables and their functional relationships are known with certainty, is said to be a determintstic model, for example, certain inventory models and games with saddle points. Probabilistic (stochastic) model: A model in which at least one parameter or decision variable is a random variable, is said to be a probabilistic or stochastic model, for example, probabilistic inventory models, games without saddle points, and queuing models Based on e Reference Static model: A static model presents a system at a specific time, which does not change over a certain period of time, for example, the replacement of machines when the money value does not change with time Dynamic model: In the dynamic model. time is considered one of the variables and the impact of changes generated by time is accounted while selecting the optimal course of action. for example, the machine replacement model when the money value changes with time.Optimization—An Overview 5 Based on Physical Appearance Physical models: ‘These models represent a physical appearance of a real object, direcily or indirectly. These are further divided into two categories. (a) Iconic models: These are scaled versions (enlarged or reduced size) of real objects, for example, blueprints, globe, photographs, drawings. and templates. (b) Analogue models: These models represent a system by a set of properties differing from that of the original system and physically not resembling the original system, but the solution is reinterpreted in terms of the original system, for example, frequency curves, flow charts, and organization charts Symbolic models: ‘These models use symbols in the form of letters or mathematical operators to represent the properties of the system. These are: (a) Verbal models: These models are used to describe a situation in written or spoken language in the form of letters, words, and sentences. () Mathematical models: These models are used to describe a situation through mathematical symbols, letters, numbers and mathematical operators by representing the relationship among the various variables of the system in order fo explain its behaviour and properties Based on Purpose and Nature Descriptive models: These models describe the situation in the form of a narrative, reports of surveys, questionnaire results, inferences of the observations, etc. These also include the models such as plant layout diagram and flow chart of an algorithm. Predictive models: These models use the query such as “if this happens, what will follow for example, preventive maintenance schedules. Based on Method of Solution Analytical models: These models have a specific mathematical structure and hence can be solved by analytical or mathematical techniques, for example, any optimization model, such as inventory models and waiting lines. Iterative models: In these models, the solution is obtained from the conclusion of the previous step, for example, the simplex method and dynamic programming. Simulation models: Though these models have a mathematical structure. they cannot be solved by applying mathematical techniques. Instead, a simulation model is essentially a computer-assisted experimentation on a mathematical structure of a real-life problem under certain assumptions over a period of time, for example, Monte Carlo simulation, the use of random numbers, and forecasting models,6 Optimization Methods for Engineers 1.4 Optimization Methods in Engineering—Modelling and Application There are numerous methods developed in this field and many more are being added to the list. However, some common methods and approaches which we come across are as follows: 1 2 10. ul 14. 15 We have seen a variety of application mod Linear programming model: A model in which the variables are linearly related. Nonlinear programming model: A model in which variables are nonlinearly related. ic, the variables assume higher degrees Allocation model: A model which is used to allocate resources to activities in such a way that the sum measure of effectiveness (objective function) is optimized, for example, transportation, assignment, sequencing, line balancing, and line of balance. One-dimensional programming: A model in which only one variable exisis. Multi-dimensional programming: A model in which more than one variable exists. Integer programming: A model in which the solution value of a decision variable is necessarily an integer. Dynamic programming model: A model with multistage. inter-related decision processes. Inventory model: A model which is used to determine the economic order quantity and (o minimize the inventory cosis such as order cost, carrying cost, and shortage cost. Game theory model: A model which is used to characterize the behaviour of two or more opponents (called plavers) who compete for achieving the conflicting goals Goal programming model: A model with more than one objective or goal Network model: A model which is applied to plan and schedule large-scale projects for example, PERT/CPM which help in identifying trouble spots in a project by critical path and in determining time—cost trade-off. resource allocation, and activity times. Waiting line or queuing model: A model which establishes a trade-off between the cost of providing a service and the waiting time of a customer in a queueing system by using probabilities Replacement model: A model which is used to decide the optimal time to replace ‘an equipment, for instance, when that equipment deteriorates or fails. Monte Carlo simulation method: A model for conducting a series of organized experiments to predict the behaviour of a process over time. that is. to find how the real process would react to certain changes. Characteristics of Optimization Models . their distinguishing features and the approach used in each, Although each individual optimization model differs from the other in the approach and method of application, all these models together have some common features. Some important characteristics are as follows, (Remember the acronym OPTIMALS.) n (O): An optimization model, whether constrained or not, will definitely have an objective or goal. The goal orientation keeps the optimization models on the top among all the problem-solving techniques.Optimization—An Overview 7 o Object Orientation P Practical and Practicable T Tangible and Time bound 1 Interdisciplimary M|__ Methodical and Mathematical A | Accurate/Approximate and Analytical L Logical s Scientific and Systematic Figure 1.1 Characteristics of optimization models. Practical and practicable (P): An optimization model is formulated to solve the real-time problems and hence it is realistic in nature and practicable. ‘Tangibility and time bound (1): A model is normally tangible and time bound. Since ‘an optimization problem is able to be expressed quantitatively. the results are measurable (‘angible). Further, the models are time sensitive, thanks to changing environment, The optimal solution may not be true every time and any time. However, in a given solution it behaves statically in a defined time span. 4, Interdisciplinary (I): Optimization is the science that is embedded with suitable blends from all the features of various fields. Thus. optimization models can play a significant role in bringing a balance, among different disciplinary people, to managerial problems. 5, Methodical and mathematical (M): The optimization models involve mathematics and its principles. They have a systematic methodology with universality of application. 6. Accurate as well as approximate and analytical (A): The optimization models are derived with a great deal of analysis. They have the ability 10 provide accurate answers as well as approximations to a given problem 7. Logical base (L): Since every optimization model is backed up and framed by using mathematically proven theorems and concepts, the optimization models stand upon a strong foundation with logical base. 8, Scientific and System approach (S): On close observation, we understand that there is a system orientation in every optimization model, because once a model is formulated. ithas a defined set of design variables and an objective function upon which the output is computed and a decision is made. Since, the model will have clear-cut input and output, it will have system orientation 1.6 Limitations of Optimization Models Optimization methods. though widely used, have got certain limitations. These are given below Impossibility and infeasibility Optimization models take care of all the factors in choosing the best alternative. In modem8 Optimization Methods for Engineers society. and in real life, factors are numerous and establishing a relation among these is either impossible or infeasible in some cases. Hence. all such problems are left unsolved. Intangible situations Optimization models can provide solutions to only those problems in which all the factors and parameters are quantifiable, But they cannot provide any solution to intangible variables or attributes that are qualitative in nature, such as a human factor. Gap between the implementer, the decision-maker and the optimizer Optimization is a specialist's job and requires the knowledge of mathematics, statistics, ete Hence an optimizer may not be familiar with the business problem and its inner aspects of flexibility and external aspects such as sensitivity to changes. Since the optimization specialist is not the decision-maker, even the best solution obtained by optimization may not actually be fruitful unless it is fully believed to be so by the decision- maker, Thus, this large gap built up between the ones who provide solutions and the ones who use them, leads to a lot of issues during implementation, So, the implementers should have clarity on the usefulness of optimization modelling, otherwise, a successful formula may result in utter failure if there is no coordination and understanding. 1.7. Method of Optimization It was difficult to describe a procedure for optimization about four or five decades ago, But today, every engineer, manager, researcher, practitioner or student is able to present the methodology or procedure in different ways to suit their own situations. However, the most acceptable general procedure for modclling and solving an optimization problem involves the following common steps: Step 1: Defining the problem and collecting the relevant data Step 2: Refining the data and formulating the problem Step 3: Identifying and fitting a suitable optimization model Step 4: Analyzing and deriving the solutions from the model Step 5: Validating the model and testing its solution Step 6: Sensitivity analysis and controlling the solution Step 7: Decision-making and implementing the solution Now let us analyze these steps in detail. Defi ing the Problem and Collecting the Relevant Data ‘The very first step of optimization process is to identify the problem, Initially, the problem may appear vague. Hence the relevant data should be collected in order to formulate the problem in the form of an appropriate model. For example, when a patient visits a doctor, he may not be able to explain what his disease is, but can only tell what the difficulty is. Only a well- experienced, qualified doctor can define the problem of the patient. So also in engineering, the problem posed by workmen is often not clear. It has to be defined clearly.Optimization—An Overview 9 Further, while defining a problem, it should not be forgotien to declare and identify the boundaries of the problem. For example, a production problem may have the variables that cannot be negative. Here, it should be stated that the boundary of the variable is defined as x > 0. Once a problem is defined and understood clearly, the relevant data are collected to establish the facts, goals and bounds. In the process, we require the following information. What are the objectives or goals? Who has to take the decision’ What are the ranges of controllable variables? What are the uncontrolled variables that may affect the possible solutions? What are the restrictions or constraints on the variables? What are the other conditions or the types or the nature of variables? Refining the Data and Formulating the Problem Often the data collected will be in the descriptive form. The descriptive problem has to be transformed into a mathematical expression or relation, Suitable variables have to chosen to institute a mathematical relationship. Suppose, Ram and Laxman go to a garden daily and bring a thousand mangoes. Here. we can choose the variables x and y, where x is the number of mangoes brought by Ram and y is the number of mangoes brought by Laxman, Now, the descriptive information can be formulated mathematically as x + = 1000. There is no guarantee that the information collected would not have some unnecessary or spurious content or misinterpretations. Therefore, the data collected should be refined and validated. The formulation of the problem should be done carefully because a wrong formulation cannot give a right decision (solution), and may even be disastrous in some cases. Identifying and Fitting a Suitable Optimization Model The next step of the research is concemed with searching a suitable optimization model in an appropriate form which is convenient for analysis. It requires identification of both static and dynamic structural elements, Also, it requires judicious estimation of the qualitative and quantitative factors, determination of their deterministic and probabilistic nature, and finding the linear and nonlinear relationships of the variables. The model consists of the following four important basic factors. 1, Decision variables and parameters 2. Objective function 3. Constraints or restrictions 4. Boundary conditions or the range of values of variables. Based on the above factors, the problem can be fitted in a suitable optimization model depending on one of the following structures of the problem. Based on the number of variables present in the problem: One-dimensional and multi- dimensional problems10 Optimization Methods for Engineers Based on the number of objectives: Single objective and multi-objective problems (goal programming). Based on constr: constraints), Unconstrained and constrained (equality constraints or inequality Analyzing and Deriving the Solutions from the Model In fact, this stage is pivotal to an optimization problem, This stage involves the computation of those values of decision variables that maximize or minimize the objective function. Such a solution is called the optimal solution, which is always in the best interest of the problem under consideration. ‘The analysis may be done manually or may require certain aids such as graph sheets and drawings. The use of computers may also become necessary at times when the calculations are huge and complex or require more number of runs or demand more accuracy Validating the Model and Testing Its Solution After computing and deriving the solution from the model, it is once again tested as a whole for errors, if any, There may be doubts about the correctness of the model and about its suitability to the situation, Therefore, it is often advised to confirm the appropriatness of the model, which is referred to as vaiidaiion A model may be said to be valid if it can provide a reliable prediction of the system's performance. A good optimization method should stay applicable for a longer time. Hence, the model has to be updated from time to time by considering the past, present and the future specifications of the problem Sensitivity Analysis and Controlling the Solution In this stage, control over the solution with the desired degree of satisfaction has to be established. ‘The model needs immediate modification as soon as the controllable variables (one or more) change significantly, otherwise the model goes out of control. In other words, it is the analysis of the solution by computing the change in the output for the given change in the input. This action is often referred to as sensitivity analysis As the conditions are constanily changing in the world, the model and the solution may not remain valid for a long time. Therefore. it leaves a large scope for further improvement in controlling the solution, and the sensitivity analysis paves the way for improvement in the solution, Decision-making and Implementing the Solution ‘The final step in the optimization cycle is ‘decision-making’ about the solution and its implementation. ‘The solution obtained from the analysis has to be approved by the decision-maker by proper testing. The best solution thus obtained by the tested results of the model has to be implemented and put to work. This phase is primarily executed with the cooperation of the optimizationOptimization—An Overview 11 experts and those who are responsible for managing and operating the systems. This phase essentially requires taking into confidence those people who implement the model or are practically involved in it, ‘The right feedback on the implementation and its results may create a new need of evaluating another solution if the first implementation fails. The success of implementation on. the other hand, may generate a need for an extension or a need for a new problem for further optimization when the new cycle begins 1.8 Application in Engineering Areas Optimization methods are employed in engineering as well as non-engineering areas. The following are some of the important areas of application of optimization techniques. 1. Civil engineering: Implementing construction projects such as bridges, roadways, frames, foundations, towers. chimneys and dams at minimum cost; and evaluating flow patterns in hydraulics and fluid mechanics, water works, drainage planning, etc Mechanical en Minimizing the effort, maximizing the output of machinery and equipment, weight reduction, efficiency enhancement, optimum design of linkages. cams, gears, machine tools and other mechanical components, design of pumps, turbines and heat transfer equipment for maximum efficiency 3. Production engineering: Selection of machining conditions in metal cutting processes for achieving minimum production cost, design of material handling equipment like conveyors, trucks and cranes at minimum cost, planning the best strategy in competition in marketing 4. Industrial engineering: Selection of a site for an industry, inventory control to minimize inventory costs such as ordering cost, carrying costs, and shortage costs, controlling the waiting and idle items and queueing in production lines to reduce the costs: shortest route taken by salesmen visiting different cities: optimum production planning, controlling and scheduling 5. Electrical engineering: Optimum design of electrical machinery like motors, generators and wansformers: optimum design of electrical networks. distribution planning, stre light pole laying, ete. 6. Electronics engine Miniaturization and cost reduction programmes. 7. Mining, marine, shipbuilding: Designing, planning and production, allocation of resources of services among several activities to maximize the benefit; planning maintenance and replacement of equipment to reduce the operating costs, designing heavy engineering, manufacturing plants of equipment, ete. 8. Aviation, aeronautical and aerospace engineering: Designing aircraft and aerospace structures for minimum weight, finding the optimal trajectories of space vehicles, space programming, rocket launching, etc 9. Computer science and engineering: Computerization and modernization, analysis of statistical data and building empirical models from experimental results to obtain the accurate representation of the physical phenomena, simulations and trial runs12. Optimization Methods for Engineers 10, Chemical engineering: Optimum design of chemical processing equipment and plants, design of optimum pipeline networks for process industries. In addition to the above areas of engineering, the optimization models are widely employed in imterdisciplinary areas. A few such areas are listed below + Railway route planning and transport system planning + Research and development (R & D) + Customizing engineering products + Designing new equipment + Town planning, + Modification of existing plants/equipment for improvement. + Expansions and extensions. + Weapons system development and warfare activities. Summary In this chapter, we learnt the meaning and concept of optimization, its history and development. Contributions to the development of optimization made by Newton, Cauchy, Lagrange, Kuhn. and Tucker, are briefly described. The various types of mathematical methods are discussed and distinguished. Numerous optimization methods with different models, such as linear and nonlinear, are briefly explained. A detailed procedure for modelling and solving optimization problems is narrated in neat steps such as formulating, fitting optimization model, anal; sensitivity analysis. The important areas of application of optimization methods in en; and non-engineering areas are listed Key Concepts Static model: A system at a specific time, which does not change over a certain period of time Dynamie model: A model in which the impact of changes generated by time is accounted while selecting the optimal course of action Ieonie modek: A scaled version (enlarged or reduced size) of real objects Analogue model: A model representing a system by a set of properties differing from that of the original system but the solution is reinterpreted in terms of the original system. Verbal model: A model used to describe a situation in written or spoken language in the form of letters, words and sentences. Mathematical model: be @ situation through mathematical symbols, letter numbers and mathematical operators by representing the relationship among the various variables of the system in order to explain its behaviour and properties, Deseriptive model: A model which describes the situation in the form of a narration or reports of surveys, questionnaire results, inferences of the observations, etc Predictive model: A model which uses the query: “If this happens, what will follow’ ‘A model used to de: Analytical model: ‘The specific mathematical structures which ean be solved by analytical or mathematical techmiques.Optimization—An Overview 13 Herative model: A model in which the solution is obtained from the conelusion of the previous step Simulation model: A model which uses a series of organized experiments to predict the behaviour of a process over time with a view to finding how the real process would react to certain changes. Linear programming model A model in which the variables are linearly related. Nonlinear programming model: A model in which the Variables are nonlinearly related, ie the variables assume higher degrees, Allocation model: A mode} which is used to allocate resoure sum measure of effectiveness (objective function) is optimized One-dimensional programming: A model in which only one variable exists Multidimensional programming: A model in which more than one variable exist Integer programming: A model in which the solution value of decision variable is necessarily an integer Dynamic programming model: A model with multistage and interrelated decision processes, s to activities in such a way that the Inventory model: A model which is used to determine the economic order quantity and to minimize the inventory costs such as order cost, carrying cast, and shortage cost Game theory model: A model which is used to characterize the behaviour of two or more opponents (called players) who compete for achieving conflicting goals Goal programming model: A model with more than one ebjective or goal Network model: A model which is applied to plan and schedule large time-cost trade-off, resource allocation and updating of activity times Waiting line or queueing model: A mode! to establish a trade-off between the cost of providing a service and the waiting time of a cusiomer. ale projects to determine Replacement model: A model to decide the optimal time to replace equipment, when that equipment deteriorates of fails. Object orientation: An optimization model, whether constrained or not, will definitely have an objective or goal Practical and practicable: An optimization model is formulated to solve real-time problems which are realistic in nature and are able to be put in practice ‘Tangibility of optimization: ‘The results which can be measured. Time-bound optimization: ‘The given solution behaves statically in a defined time span. Interdisciplinary optimization: _ An optimization method which suitably blends the features of vanious fields Methodical optimization: An optimization model involves mathematics and its principles. It has a systematic methodology, which has universality of application Accurate and approximate optimization: An optimization model having the ability to provide accurate answers (rigidity) and approximations (flexibility) Logical and scientific base of optimization: A logical and scientific base of optimization model is backed up by mathematical theorems and concepts. System approach: An optimization model having an approach of an input-output model. Review Exercises 1. What do you understand by optimization? How would you represent an optimization problem? 2. Write a note on the origin and development of optimization methods14 Optimization Methods for Engineers Discuss the development of various optimization methods after the Second World War. Briefly explain the developments in optimization techniques during the 1960s. Write the names of the various contributors who played a key role in the development of optimization methods and discuss their contributions. 6. Write a note on the contributions of the following scientists towards optimization techniques: (a) Newton (b) Cauchy (¢) Lagrange (a) Kuhn and Tucker 7. Describe the various optimization models based on the physical appearance of the model 8. Discuss the different optimization models with reference to the following (a) Time (b) Certainty or risk (c) Method of solution 9. What are the application models of optimization used in engineering? 10. Name some of the areas of applications of optimization in various fields of engincering 11. Enumerate the merits of optimization methods. What are their limitations? 12. Discuss the method (procedure) of optimization. 13. Describe the procedure of mathematical modelling 14, Explain the terms ‘validation’ and ‘sensitivity’ in the context of optimization modelling, 15. Discuss the demerits of optimization method when a gap exists between the optimizer and the implementer. 16. What are the different types of models of optimization? Explain, 17, Enumerate the strengths and weaknesses of optimization approaches. 18. Explain the characteristic features of optimization models. 19. Write a short note on ‘general optimization models” 20. Write a detailed note on the history and development of optimization.Formulation of Optimization Problems Learning Objectives After studying this chapter, you should be able to: 1. Understand the structure of an optimization problem, Identify the different decision variables, Describe the objective functions. Explain and identify the various types of constraints, Appreciate the significance of the conditions of variables. Formulate the given information into a mathematical model. When a problem is identified, the relevant data are collected as part of the process of solving the problem. Often the data are in descriptive model, and are required to be translated into an appropriate mathematical or optimization model. Depending on the nature and relationships among the decision variables, there are two major models befitting the real-life problems, namely the linear programming problems (LPPs) and the nonlinear programming problems (NLPPs) As the name suggests, the variables arc linearly related in LPPs while nonlinearly related in NLPPs. Whatever the problem (LPP or NLPP), the process of translation of the descriptive model into a mathematical model is called formulation. In this chapter, we will leam how a given problem in descriptive form can be formulated into a mathematical form. Before learning the process of formulation, we will first understand the terms that we often come across 2.1 Decision Variables or Design Vector A variable in the context of optimization problems is a quantified parameter which can be altered while designing the problem as illustrated in the following cases 1