Chapter 1: Introduct 24 Operations Research fy Problems(L.PP) Meining abe Definon” ft) at [Grete ane 2s Origin of Operation Research(OR), 11 23.1 ce rare 25 Sein eee of Operations 12! 23:3 Se Oe ai raee 6 Research 2.3.3. Maximisation Linear Programming 26 14, Nature of Operations Research 13 Problems 1'LS. Scope of Operations Research 13. 23.4. Minimisation Linear Programming 29 1% Role of OR in, Managerial Decision 14 _. Problem . % Nang = 2495." “Mixed, Constraints Linear LI]. © Methodology”, “ of “Operations, 14 ‘Programming Problem a Research . ee 2.316, Liilutions of Graphical Method 32 OR Models 14 237: special Cases in Graphical Solution 3; 2 Classification based on Structure 15 to LPP Classification based on Purpose 16, 23. ‘Alternative Optimum Solutions 32. * Gesntesbon ised on fitie'’16 23:72. ”"'Onbounded Solutions 3 ference 23.73. Infeasible Solution 1.184, Clasifiation based on Degree of 17 24. Simplex Method 2 ertainty 24.1, Introduction 34 Classification based on Method of 17 242. Pe of Simplex Method 34 _. 4.3. .f Simpl 1:1:9. ~ Reasons of Using OR/Advantiges of 17 2.4.4. Slack, Susplos hry Artifici Z Operations Research Be enue be “4 at ables Vo, Dissdvanages, of Operations 17 245. Simplex Algorithm (Maximisation 35 1.1.11. Applications 6f Operation! ae? 2 oe Operatiofis Research 18 2.4.6. Sip Algorithm (Minimisation 39 12 Exercise . xercl 15 247 _Aivaniages of Simplex Metiod b Chapter 2: Linear 9. Anite Variables Wecholque ay Programming Simplex Method =“ Ls idear egrenanian) Big-M —_ Method/Primal-Penalty an Meaaing aod Defiaiisn ‘Method: bi 1.2. Sete ae Model/General Be Phase Simplex Method 42 213, Componeats Bice un Bene ee between Two Phase a 45 re ear 20 - Progamming 24.10. ial Cases in Si 204, Aepatgrione) Fthe : 2.4.10.1 Hee se So Prograinming Probt of Linear 20 DB ai0g lative Optimum Solutions 45 es ing Probeen 4.10.2. Unbounded Solution 46 Sundicd sod. Canoniea forms of Infeasible Solution ar Fee, Programming Problem Degeneracy 47 Duality in Linear Programming 48 “formulation of Industrial and 22 Business Problems as Linear Programming Problems Solution of Linear Programming 25 UNIT I idn'to Reduction of LPP to Canonical a Redluction of LEP to Standard Form 21 j1. Introduction2.6. Exercise UNIT IT 82 Chapter 3: Transportation RES 3d. 3.5. 31.6. 3.17. 3.8. 3.1.10. 3.10.1. 3.1102, 3.1.10.3. 3.1.10.4, 3.1.10.5. 3.1106. 3.1.10,7. 3.2, Problems ‘Transportation Problems. Introduction Mathematical Model gp ‘Transportation Problem Assumptions in ‘Transportation Problem Types of Transportation Problems Some Basic Solutions for Transportation Problem, ‘Transportation Algorithm Phases of Solution of Transportation Problems Initial Basic Solution(BFS) of Problems North-West Comer(NWC) Rule Least. Cost Method 7 Matrix Minimum Method Vogel's Approximation Method (VAM) Optimal. Solution of Transporation Problems Stepping-Stone Method Modified Distribution (MODI or UV Method Cases in Transportation Problems Balanced Transportation Problems “Feasible Transportation Method Unbalanced ‘Transportation Problems ‘ Minimising Transportation Problems . Maximising ‘Transportation Problems. ~ Degeneracy. in Transportation Problems : Multiple Optimal Solutions ‘Transshipment Models Exercise Chapter 4: Assignment Problems Assignment Problems Introduction Mathematical Model of Assignment Problem Applications of Assignment Model Solution of Assigament Problem Hungarian Method/Flood's Technique 54 54 54 55 55 56 56 56 ST 37 58 60 62 62 67 16 16 78 80 82 86 88 89 B25 92 92 92 93 93 93 Cha Sa. Special Problems Ralanced Problems Unbalanced (Non .Square Matrix) Problems nisation Problems Maximisation Problems Mittiple Optimal Solutions Traveling Salesman’s Problem Comparison between Transportation and Assignment Problem Exercise Cases in UNIT 111 pter 5: PERT and CPM Network Analysis Introduction Objectives of Network Analysis Network Terminology Precedence Relationship Network Construction Rules for Drawing Network Diagram Common Errors & Dummies in Network Advantages of Network Analysis Disadvantages of Network Analysis ‘Techniques for Network Analysis Critical Path Method (CPM) Introduction Principles of CPM Analysis _ ‘Advantages of CPM Disadvantages of CPM Critical Path Finding Critical ~—PatlyTime Estimation in Network Analysis Calculating EST and EFT(Forward Pass Computations) Calculating LST and LET(Backward Pass Computations) Floats and Slacks Project Crashing Objectives of Project Crashing Crash Time and Crash Cost we-Cost Trade-off Steps in Project Crashing Program Evaluation and Review ‘Technique (PERT) Introduction Features of PERT Steps Involved in PERT. Advantages of PERT Disadvantages of PERT 7 98 100 102 105 108 M1 4. ng 6 U6 116 Ww 417 19 119 119 121 121 121 122 12 122 12 13 123 123 126 132 132 132 133 134 139 139 139 140 140 140,5.3.6, 5.3.7. 5.3.8. 5.3.9. 5.4, Probability Consideration in PERT Probability of Project CompletiowEstimation of Project Completion Tine Api ns of CPM and PERT Techniques in Project Planning and Control Difference between PERT and CPM Exercise 140 Mt 146 146 146 Chapter 6: Decision’ Theory 6.1. 6.1. 6.1 6.1 Decision Theory Introduction Elements in Decision Making Steps in Decision Making s Making Environments Decision-Making Under Certainty Decision Making Under Uncertainty . Maximax or Minimin (Criterion of Optimism) Maximin or Minimax, (Criterion of Pessimism) Minimax-Regret Criteria (Savage Principle) Equally Likely ‘Decision: (Laplace Criterion) Criterion of Realism (Hurwicz Criterion), Decision Making Under Risk Maximum Likelihood Rule Expected Pay-off Criterion Expected Monetary Value(EMV) Expected Opportunity Loss (EOL) Expected Regret Expected - Value... of Information (EVP) Bayesian Analysis Decision Trees Introduction . Steps in Decision Treé Analysis Advantages of Decision Tree Disadvantages of Decision Tree Applications of Decision Tree Exercise UNIT IV Perfect Chapter 7: Game Theory Game Theory Meaning 148 148 148, 150 150 151 151 152 152 152 153 153 155 156 is6 157 158, 161 162 165 165 166 166 167 167 169 1 71 i 7 114 TLS. 116. TA, TB 718d. 1182. 71.83. 71.9. 7.1.10, TALL TAZ, TAB, 12, Ta. 122. 723, 124, 725. 725.1. 7252. 1253. 125.4. 7255., 73. ‘Terminology of Game m2 Assumptions of Game In Saddle Point m Value of Game 173 Strategies of Game 173 Pare Strategy» 173 Mixed Strategy . 173 Difference between a Pure Stategy 174 and a Mixed Strategy” Solutions of Game 14 Types of Game 174 Advantages of Game Theory 175 Disadvantages of Game Theory 175 Applications of Game Theory 175 ‘Two Person Zero-Sum Game 176 Introduction —- 116 Payotf Matrix 176 Assumptions of Two Person Zero 177 Sum Game Solution of Games with Saddle 177 PoinvPure. Strategy Game with Saddle Point: Minimax & Maximin Principles Solution of Games without Sadle 178 Points/Mixed Strategy Game without Saddle Points Algebraic Method-2x2Game, 178 Principle of Dominance/ 180 Dominance Rule—m x n Game Graphical Method —2 xn and mx 182 2 Game Odds Method/Arithmetic Method- 187 2x2Game ~ Method of Matrices-nxnGame 189 Exercise 192 Chapter 8: Queuing Theory Queuing Theory 192 Concept 192 Queue/ Waiting Line 192 Operating Characteristics of | 192 Queuing System ements of Queuing System 193 Advantages of Queuing Theory 198 Disadvantages of Queuing Theory 196 Applications of Queuing Theory 196 symbols and Notations 197 Terminology of Queuing Model 197 Queuing Models 197 Tatcoduction 7 Assumptions of Queuing Model Kendall's Notation for 198 Representing Queuing Models84.1. 8.4.2, Classification of Queuing Model Model f (Erlang Model): Poisson Distributed Arrivals and Distributed Exponentially Service Time Model(M/M/) Introduction Assumptions of Single Channel Queuing Model Steady State Solution of MAM/I Single Channel and Infinite Population (M/M/1): (co/FIFO) Characteristics of M/M/1 Queuing System with Infinite Population Relations between Average Queue Length & Average Waiting Time — Little's Formula Sing! Channel and Finite Populattiou «M/M/1): (FIFO/.NIM) Model i (General Erlang Model): Poisson Distributed Arriesig Exponentially Distriouted Service ‘Fime Model AUMIK) of Multi-Channel "Steady State Solution af (M/M/K) Multi-Channel = with Infinite AMIK): (eofFIFO) -10- 198 199 206° 206 206 207 8.4.4.1. Assumptions of Multi-Channel 207 with Infinite Population (M/M/K) 8.44.2. Operating Characteristics of 207 M/M/K Queuing Model with Infinite Population 8.5. Exercise 210 Chapter 9: Simulation Simulation 2u Meaning and Defintion - 2 Scope of Simulation Techniques Sa <, Simulation Process “Advantages of ~—- Simulation 212 Technique 9.15. sLimitations . of. Simulation 213 Technique 9.1.6. Applications of Simulation 213 9.1.7. Simulation Models 214, 9.1.8. Monte Carlo Simulation 214 Monte Carlo Simulation Procedure 214 Random Digits and Methods of 215 918. Generating Probability Distribution 9.2. Exercise 217, Solved Paper (2017) 219 Solved Paper (2018) 222 Solved Paper (2019) , 22512 (Usit) control charts in 1924 by Shewhart and was named as Shewhart control charts. Many terminologies related to the acceptance sampling of quality control were described by Westem Electric Company in the time period of 1925-26. The attributes of raw materials/components/. finished products were controlled using these terminologies. The developed terminologies were _ like consumer's risk, producer’s risk, probability of acceptance, Operating Characteristics (OC) curve, Lot Tolerance Percent Defective (LTPD), double sampling plan, type-I error, type-II error and many more. ‘The fundamental concept of sampling inspection was introduced in 1925 by Dodge and after a decade, the British Standard Institution Number 600 was developed by Pearson with title ‘Application of statistical methods (o international standardisation and quality control’. 2) Developments during World War I: At the time of World War Il, the Britain military was maily concerned for the effective exploitation of infrequent resources. Hence, a research was conducted by the native scientists. They discovered the various ways to make the fullest use of the resources to improve the efficiency of the military operations. ‘This approach was incorporated in operations -—_research methodologies for problem-solving aspects. Post-World War II Developments: Post Second World War, the main emphasis of the ‘American and British companies was to ‘maximise the profit from limited resources, so, these companies focused on application of ‘operation research methodologies. ‘The simplex method, for solving linear programming problem, was developed by Dantzing in 1947. Thereafter, in 1952, + Operations Research Society of America and in 1953, the Institute of Management Sciences was established. 4) Computer Bra: There were various complex computations in the: operations research which consumed much time; hence, the computers were eloped to solve such problems easily. The ent of the recent interactive computers 3) MBA Third Semester (Operations Research) MDU 11 Characteristics of Operations Research Some significant characteristics of Operations Research are highlighted below: (Characteristics of Operations Research Decision Making, Sciemific Approach Objective Oriented ‘Approsch Tsien Dislplinary Approach 1) Decision-Making: Decision-making or problem solving constitutes the major working of ‘operations - research. Managerial decision- making is considered to be a general systematic process of operations research (OR). 2) Scientific Approach: Like any other research, operations research also emphasises on the overall approach and takes into account all the significant effects of the system. It understands and evaluates them as a whole. It takes 2 scientific approach. towards reasoning. Kt involves the methods defining the problem, its formulation, testing and analysing of the results obtained. 3) Objective-Oriented Approach: Operations Research not only takes the overall view of the problem, but also endeavours to arrive at the best possible (say optimal) solution to the problem in hand, It takes an objective-oriented approach. ‘To achieve this, itis necessary to have a defined measure of effectiveness which ‘is based on the goals of the organisation. This measure is then used to make a comparison between altemative solutions to the problem and adopt the best one. 4) Inter-Disciplinary Approach; No approach can be effective, if epplied individually. OR is also inter-disciplinary in nature. Problems are ‘multi-dimensional and approsch needs a\team work. For example, mansigerial problems are affected by economic, _ sociological, iological, ological, physical and cnginesring aspect, that plans to arrive at a solution, to sach needs people § specialists in uch as mathomat aay engineering, ete,Introduction to Operations Research (Chapter 1) 114, Nature of Research ‘The nature of Operations research is as follows: Operations Natore of Operations Research Goal Oriented Optinum Quatiryorsoiuice FL] ‘Soltion Use of sodels Require Wikiog Bxecuives Reduces Complenity 1), Quality of Solution: Operation research simply helps in improving the quality of the solution but does not result inta a perfect solution. 2). Goal Oriented Optimum Solution: Operations Research tries to optimize a well-defined function subject to given constraints and as such is concerned with the optimization theory. 3), Use of Models: Operations Research uses models built by quantitative measurement of the . Variables conceming a given problem and also derives a solution from the model using one or more of. the diversified solution techniques. A solution maybe extracted from a model either by conducting experiments on it or by mathematical analysis. The purpose is (o help the management to determine its policy and ections scient 4) Require Willing Executives: Operations Rescarch does require the willingness on the part of the executive for expcrimentatioarté evaluate the cots and the consequences of the alternative solutions of the problem. It cnables the decision- maker to be objective in choosing an alternative from among many possible alternatives. 5) Reduces Complexity: Operations Research tries to reduce the complexity of business operations and docs help the executive in conecting a difficult function and to consider innovations which are too costly and complicsted to experiment with the actual practice;, = 1.4.5. Scope of Operations Research Operation Research can be applied (o different areas of business such as: = Scope of Operations Research Tecusy Developing Economies ‘Apical tnausy (Tieton snes and Socery » y 3 4 R Industry: Industrial management deals with a series of problems, starting sight from the purchase of raw materials till the dispatch of final Products. The management is ultimately interested in overall understanding of the methods, of optimising profits. Therefore, to take decision on sciewtific basis, operations research team has to think about various alternative methods, to produce goods and obtaining retums in each case. ‘Not only this, the operations research study should also suggest possible changes in’ the overall structure like installation of a new machine or introduction to automation, etc, for optimising the résults. Many industries have gained immensely by applying operations research in various tasks. For example, operations research can be used in the fields of manufacturing and production, blending and product mix, inventory management, for forecasting demand, sale and purchase, for repair and maintenance jobs, for scheduling and sequeticing planning, and also for scheduling and control of projects. Developing Economies: OR is applicable to both developing and developed economies, But 2 lot of scope exists in developing economies, for building up an operations research approach towards planning. The basic idea is to orient the planning, to achieve maximum growth per éapita income in minimum time: considering the goals and restrictious of the country, Poverty ‘and hunger are the core. problems faced by many countries of Asia and Africa, Therefore, people like’ statisticians, economists, technicians, administrators, politicians and agriculture experts can work in conjunction, 10 solve this problem with an operations research approach. Agriculture Industry: Operations research. ‘approach has a huge scope in agriculture sector. Population explosion has led to scargity of food. ‘Optimism allocation of land for various crops in accordance with slimatic conditions is a challenge for many countries. Also, each developing country is facing the problem of optimal distribution of water from several water bodies. These areas of concern hold a great scope for scientific research. Organisation: Organisation, big or small, can adopt operations research approach effectively. Operstional productivity of organisations has improved by using quantitative techniques. Techniques of operations research can be applied o minimise cost, and maximise ‘benefit for decisions. Kor example, a departmental store faces roblem like, employing additional sales ‘girls, or purchasing an additional van, etc.1 (Ua 5) Business and Society: Businesses and society can directly be benefited from operations research, For example, hospitals, clinics ete» Operations research methods can be applied directly to solve administrative problems such as minimising the waiting time of outdoor patients. Similarly, the business of transport can also be benefited by applying simulation methods. Such methods can help to regulate train arrivals and their runing timings. Queuing theory can be applied to minimise Congestion and passengers waiting time. These methods are increasingly being applied in LLC. workplaces. {¢ helps in deciding the premium rates of various policies. Industries such as petroleum, paper, chemical, metal processing, aircraft, cubber, mining and textile have been extremely benefited by its use. 1.1.6. Role of OR in Managerial Decision Making In managerial decision-making, the management chooses a specific course of action, from a sct of possible options, when face a problem. The manager tries to choose the most effective course of action in the given circumstances keeping the targets of the organisation in mind. x Some decision-making situations are simple in real life. Complexities in decision-making situations arise due to several factors. These factors include the interaction of the economic, political, technological, environmental and competitive forces in society in 2 complicated manner, the limited resources of an organisation, the values, the risk attitudes and the knowledge of the decision-makers. For example, the factors such as market conditions, labour rates and its availability, investment fequirements and availability of funds, influence = ‘company's decision to introduce a new product. The decision for production methodology, cost and quality of the product, price, package design, marketing and advertising strategy will be of multidimeosional response. Every segment of the organisation will be affected conceivably by such decisions The major roles of Operations Research in decision making are as below: pa 1) It provides 2 tool for scientific analysis. 2) Itoffers solutions for various business problems 3) enables proper deployment of resources. tions Research) MOU MDA Third Semester (OPE sninimising waiting and servicin management to decide when (0 bY the procedure of buying. It helps in reducing the total proces eceseary for performing a set of Jobs. 5) Whelps the time 6 of Operations 1.1.7. Methodology Research Following are the six steps towards problem at the Problem: First of all, Step 1: Formulation of ‘a manager should be abl f model of the problem, so as to arrive a Mathematical Model ‘a mathematical mods}, dor study using Je to form an appropriate 1a solution. Step 2: Coastruction of The second step is to build which represents the system Un variables. Step 3:Deriving the Solution from the Model: ‘An effective computation of all the decision variables that constitutes the problem is needed to maximise of minimise an objective function is required. Such solution is called an optimal solution. Step 4: Validity of the Model: Every mode] neods a validation for accuracy. A model can be valid or accurate if: 1) All. the objectives, constraints and decision variables included in the model and are relevant to the problem or area part of it, and ships. Step 5: Establishing Control over Solution: The immediate next step after arriving at a solution, is to exercise and establish control ovér it. It requires enforcing feedback on those variables, which actually have tendency to deviate from the acceptable regime considerably. 2). Ithas valid functional rel Step 6: 1mplementat of the Final Results: In the end, the final results of the model are put to work, Crefl explanation ofthe adopted solution, and its telatioaship with the functional realities should be considered, hn OR Models eration rezearch ae he: by which any unique set at, aes teats. ‘There are various methods for classification models. By the help of such modets decision ak can easily recognise Which type of mod suitable fo his problem.is are useful in defining the system in place of explaining the system. Therefore we can sy that iconic models are usually characterise of a static event and characteristics which are not useful in defining or forecasting effect, ue to certain changes in the original sysien, For example, for scientific study of the structure of an atom colour does not play an vital role. y tmnt to Operations Reetch (HRP » Types of Operations Research Models Several classifications are as follows: [Frasitiaton based on Sires Gasca vocton | Purpose Reference cf certainty cation bared on Method of Solonen Chassifi 1.1.8.1. Classification based on Structure ‘The model based on the structure includes: | i ! Chinon ted 08 Physical Medes Symbolic Models . i Ieovie Mes E ¥epatsete ‘slopue Movels Mathematical Medels cal Models: The model which provides a jn areal form, for the problem is called physical model. Such models works cither by reduction in size or by mounted up but t past from that these models are helpful only for designing the problems as they arc easy to observe, build and describe. 1) Phys physical presence, For example, in the aircraft industry, scale models of the proposed new aircraft are build and tested in wind tunnels to record the stresses experienced by the air frame. Since these models cannot be manipulated and not very useful for prediction, therefore, problems such as portfolio selection, media selection, —_production scheduling, etc. cannot be analysed with physical model. Physical models are classified into the following two categories: i) Teonle Models: An iconic model is an unrealistic or topped variety of the system. This iconic model basically holds few physical properties and features of the system which they represent. Moreover this \——_ [Gavaeaton basnt on Deere] ji) Analogue Models: Analogue mode) signifies a system with the set of propery which is apart from the actual system and not similar substantially. But at the same time, when the problem is solved the soh : again interpreted with reference to origina, system. For example, the organisational chon represents the stale of format elatlonships existing between members of the organisation. Maps in different’ colours may represent ‘water, desert and other geographical features, Graphs of time series, stock-inarket changes, frequency curves, ele.. may be used to tepresent quantitative relationship between any two properties, These models are less specific “and concrete but easier (0 maniptlate and more general than iconic models. 2) Symbolic Models: The models which use symbols, lelters and numbers for representing the properties of the system is called symbolic model. Such models are also useful in characterising relationship which can also be denoted in physical form. Symbolic models ean be classified into two categories: i) Verbal Models: Verbal models are defined as the model which refers the condition in 2 written or spoken form written sentences, books etc For example, Mathematical Models: The models. in which use of mathematical symbols, Jetters, nurnbers and mathematical operators (+, ~» * %) for Teptesenting relationships bewveen_ various variables of the system for describing. it properties or behaviour is called mathe model. By applying appropiate inathematcal Techniques ro th the help of in the solution of the160 D 1.1.8.2. Classification based on Purpose The models based on the purpose of their utility include Descriptive Masts = H. Predictive Modes tion) Modets Normative (Opti 1) Descriptive Models: The model which describes the various facet of a situation, which is based on observation survey, questionnaire results or other offered data related with situation and do not forecasted or appraised For example, i) Organisation chart, ii) Plant layout diagram, iii) Block diagram representing an algorithm or method for solving a problem, ete 2) Predietive Models: Predictive model works on mainly one strateg: this occurs, then that will follow”. Predictive model communicates dependent and independent variables and wies to find out, ‘what if” questions. Basically predictive model can also be defined as that model which is useful for forecasting the consequences because of an available set of alternatives for the ation. But on the contrary, predictive models do not have an objective function as+a part of the model for evaluating decision alternatives. For example, S = a + DA + cl is a’model that describes how the sale (S) a product changes with a change in advertising expenditure (A) and disposable personal income (1), Here a, b and ¢ are parimetérs whose values must be estimated, ‘Thus, having estimated the values of a, b and-c, the value of advertising expenditure (A) can be adjusted for a given value of Ito study impact of advertising on sales. For predictive models, no fone can try to select the best available altematives for decision, but they have only hint for each available alternative to him. 3X) Optimisation (Normative) Models: Optimisation or ‘we can say normative models provide the best or optimal solution for the problems within the given limits on the use of resources. Optimisation models provide acelsimed course of action For example, for mathematical programming, generally models are framed for improving the given objective MBA Third Semester (Operations Resezrch) MDU function, subject to restrictions on resources in the reference of the problem under consideration and non-negativity of variables. Such models are also known as prescriptive model. It isso ‘cause they proposed only what the decision- maker should do. In optimisation model, when presentation of altemative occurs, an intensive effort is made for reaching an optimal solution. According to input criteria, when optimisation is solved, it produces the best alternatives. ‘Therefore if we talk about optimisation model, it is basically concerned in arriving at an optimal solution, which is referred as the best solution, Optimisation model works very simply by maxfnizing or minimizing a real funetion by choosing input values analytically from within an allowed set and figuring the value of the function. If we look a broad view of optimization theory and techniques to other originations, it encompasses a vast area of applied mathematics. Commonly optimization theory comprises of finding best available values of few objective function within a given defined domain, comprises of different types of objective functions with different types. of domain, These models are basically used in linear programming problem, transportation problems etc. 1.1.8.3. Reference ‘The models based on the time reference include: Classification based on Time Classification based on Time Reference ‘Static Modets | | Dynamic Models 1) Static Models: Static modeis present a system at some specified time and do not account for changes over time. For example, an inventory model ean be:developed and solved to determine economic order quantity for the next period assuming that the demand in planning period would remain the same as that for today. 2) Dynamic Models: The models, where time is measured ‘as one of the variable and also eae ges that the effect of changes are produced only by the time for selection of the optimal courses of the action. Therefore, we ca say that the arrangements of inter decisi sing the optimal course the given objective. Dynamic ming is an example of aInuodvetion to Operations Research (Chapter 1) 1.1.8.4, Certainty ‘The model based on the degree of certainty includes: Classification based on Degree of (Classification based on Degree of tainty Probaitisie (Sechasue) Mote 1) Deterministic Models: A model is said to be deterministic if it is assumed that the all the Parameters, constants and functional relationships are known with certainty when the decision is made. Thus in this case, the outcome associated with a particular course of action is known, There is @ uniquedetermined output for 2 specific set of ‘input values. This output Tepresents the solution of the model under the conditions of certainty. The results of the models assume single value. For example, linear Programming and inventory problem model, 2) Probabilistic (Stochastic) Models: A model is called probabilistic (or stochastic) model least one parameter or decision variable is a random variable, An independent variable which is the function of dependent variable(s) will also be random because at least one decision variable is random. Determine outs LY This means consequences or pay-off due to certain changes in the independent variable cannot be predicted with certainty. However, it is possible to predict a pattein of values of both the variables by their probability distribution. For example, insurance against risk of fire, accidents, sickness, etc. where the pattern of events is studied in the form of a probability distribution. 1.1.8.5. Solution ‘The models based on the solution methods include: Classification based on Method of Classifcallon based on Method of Solution Analytical Medels mulation Models 1) Analytical Models: ‘The models, which are having a precise mathematical structure and that's why they can be known as analytical or ‘mathematical technique, are known as analytical model. In other words, any type of optimisation ‘model, in which maximisation or sninimisation of an objective function is required, is known as analytical model. 0 2) Simulation Models: Simulation model. also having a property of mathematical structure but such model cannot be solved by applying mathematical techniques for getting solution, As a substitute, simulation model is necessarily computer assisted experimentation on any mathematical structure of a real life probler, for describing and evaluating its behaviour, under certain assumptions over a period of time. Moreover we can say that simulation models are more flexible in comparison to any other mathematical model so, thetefore, they are easy 10 represent a complex system, which otherwise cannot be -—srepresented mathematically. But at the same time simulation model cannot provide general solutions like other mathematical models. 1.1.9. Reasons of Studying OR/Advantages of Operations Research 1) Better Control: For large organisations, it is practically impossible to continuously supervise every routine work. OR approach comes handy and gives an analytical and quantitative basis to ideotify the problem area. OR approach is most frequently adopted with production scheduling and inventory replenishment. 2) Better Systems: Problems identifying the best location for factories or decision on whether to open a new warchouse, etc., are often been studied and analysed by OR approach. ‘This approach helps to improve the existing system such as, selecting econorhical means of ‘ransportation, production scheduling, job sequencing, or replacing old machinery. 3) Better Decisions: OR models help in improved decision-making and thereby reduce the risk of wrong decisions, OR approach gives the executive an improved insight into the problem and thereby improves decision-making, 4) Better Coordination: OR models help in coordination of different or various divisions of ‘an organisation. ii 11.10. Disadvantages of Operations Research approach) 18 mathematical in nature. OR techniques tfY (0 find out an optimal solution toa problem, “By taking all the factors intoGomiderstion. The need of computers: become (ng ecidable because these factors are enormous (RFE). tequites bur calculations to express hem in quantity and to establish relations among then “eeshns 2) Non-Quantifiable Factors: One of the drawbacks of OR techniques is that they provide 8 selution onty when all the elements related to Problem are quantified. Since all relevant Variables may not be quantified, they do not find a place in OR models, 3 Wrong Estimation: Certain assuniptions and estimates are made for assigning quantitative values to factors involved in Ol, so that a quantitative analysis can be dose. If such estimates are wrong, the result can be misleading, 4) Involves Time anid Cost: Operations research is 8 costly affair. An organisation needs to invest fime, money and effort into OR to make it effective. Professionals need to be hited to eonduct constant research, For better research outcomes, these professionals must constantly review the rapidly changing business scenarios, 5) Implementation: The complexities of human selations and behaviour must be taken into azcount implementing, OR decisions, as it is a very ate task. 1.1.11. Applications of Operations Research in Business Applications of operations research in business are as follows: 1) Finance, Budgeting and Investment i) Long range capital requirements, cashflow sumalysis, investment portfolios and dividend polices. Credit policies, credit risks and procedures for delinquent account. iii) Procedures to deal with complaints and claim, 2) Marketing i) Selection of product, its timing and competitive actions, ii) Cost and time-based decision for advertising media, .‘ ii) Rate of calling an “of rumber of salesmen, eee MBA Thin Semester (Operations Research) MOU 3) Physical Distribution i) Size of warehouses, distribution centre, retail outlets, etc, and theis location, and i) Policy for distnbution. 4) Purchasing, Procurement and Exploration 1) Buying | 3) Determining purchase timing and its quanti lu) Policies for bidding and analysis of vendor, and iy) Replacement policies of equipment. 3) Personnel : i) Manpower requirement faree: recruitment policies and assignment of job 4i) Suitable personnel selection considering age and skills, ete., and fii) For cach service centre determining the optimum number of persons. 6) Production ; 4) Proper allocation of machines for scheduling. and sequencing the production. ii) Optinium product mix calculation, and iii) Selecting production plant sites alongwith its location and design 7) Research and Development i) Altemative designs evaluation reliability. fi) Developed projects" control. iii) Multiple research projects’ coordination, and iv) Required determination of time and cost. 1.2.1. Short Answer Types Questions 1) Whatis operations research? 2) What are the applications of operatons research in industries? 3)_ List out the main operations research models? 4) What are the advantages of operations research? 5) Write any four limitations of operations research? ©) Define Iconic and Analogue model of OR. 1.2.2, Long Answer Types Questions 1) Define operations research. Explain the nature of OR. 2) Explain the origin of operations research. 3) What are the characteristics of operation research? 4) Brelly explain the operations research models. 5) Explain the scope of operations research in different areas. 7 6 ey the role of OR in managerial decision i and itsLinear Programming (Chante?) Linear Program); Min, SS 21.1. Meaning and Linear programming is a (Wor He refers to any linear word phrase. It 1) The word tinear , relationship among variables in a model, means any change in one variable will-result into woportional change in other variables. Programming refers to any problem that can be modelled and solved niathematically. There are various altemate strategies 10 achieve the desired objective. But in programming, economic allocation of limited resources and choosing a particular course of strategy helps in solving the problem. 2 ‘An analysis of problems represented in linear function with a number of variables that is to be -oplimised, when subjected to a number of restraints in the form of inequalities, is called Linear Programming. A problem can be approached and solved with different strategies. Linear Programming selects the best possible strategy from the available alternatives. Strategy is selected on the basis of maximisation or minimisation of some required output, ic, it has to be optimal. For example, maximisation of outputprofit or minimisation of production cost. According to William =M_ Fox, “Linear programming is a planning technique that permits some objeclive functions to be minimized or ised within the framework of given situational restrictions.” ar programming is 2 technique that works on a mathematical basis’ for determining the optimal solution. It takes into cansideration the alternative uses of resources like man, machine, money, material, etc. to attain a panicular objective. Lincar programming can be applied on varior mnoblems. However, tis technigee ean be Meets used for solving maximisation andior mini, problems, subject 10 some assumpiq Stn example, maximisation of profi = Poy minimisation of cost es 2.1.2. Mathematical ModelGe, Structure of LPP Let 1, Ar. Xq be n decision variables. They mathematical form of linear programmin is as follows: 8 Probles Optimize (Maximize or Minimize) Z = eyx, +c, sect CaKe (Objective function) Neral Subject to the constraints; yk) yyy teow ay (SEI, =.2)b, yh tga Ra toe ba (Constraints) BeiR Hg i Rs tot Oph (SH 2)Py and X), Xv: %o20 — (Non-negativity condition) Where aj, B, ¢, are constants and x; is decision variable. By using the symbol “E’,i.., the “sum’ of rotation, the above formulation may be put in the following compact form: Optimise (Max. or Min.) Z= 3° Soya otf (Objective Fonction). Subject to the linear constraints: Days Ss2ybyi=1,2, x (Constraints) -..@) (Non-negativity) OG) Some Useful Definitions snc Definition 1: A Feasible Solution to we Une programming problem is a vector X = (ie Se — X5,...,) which satisfies conditions (2) and i A feasible solution to a linear program is a solution that satisfies alt constraints.20 (Usit MBA Thind Semester (Operations Research) MDU Definition 2: A basic solution to equation (2) is a i + clulun oie ty sting (om) saab enal ag ganageaeaians | Of Linear to zero and solving for the remaining, m variables, gramming Problem provided that the determinant of the coefficients of Four basic assumptions that are important for all these m variables is non-zero, The m non-zero fears Broblers re af follows variables are called basie variables and (n — m) rtainty: the variables or parameters in ‘zero variables are called non-basic variables. an LP model should be assumed to be constant or known, For the best optimal solution, parameters such as availability and consumption of resources or the profiveost contribution variable should be assumed to be For examples, 1) Maximization Case Maximise Z=40xi+35x2 (Profit) Subject to . . certain. Certainty in an LP model is concerned 2x +3x2 £ 60 (First constraint) with the coefficients in the objective function 4x, + 3x2 $96 (Second constraint) g, the coefficients in the functional coefficients xyx220 (Non-negativity condition) 44; and the sight hand sides of the functional 2) Minimization Case constraints b,, The values of the coefficients of Minimize Z=40x)+24x2 (Cost) _ each variable are to be constant or certain. This Subject 10 means that values of Gy aj by are fixed and 20x, + 50x: 2 4800 (First constraint) known with certainty. 80x: +508227200 (Second constraint) 2) Divisibility (or Continuity): It is assumed x20 (Non-negativity condition) that decision variables and resources have solution values which are either whole 2.1.3. Components of Linear numbers (integers) or , mixed numbers Programming Problem integer and fractional). ‘The integer Following are the components of a lincar programming method may be applied to get programming problem (LPP): the desired values if only integer variables 1) Objective Function: LPP has a component are desired, €.g.. machines,, employees, ct, called Objective Function or Criterion function. However, in an LPP iodel, decision It is a linear function and can work either by variables can have any value that satisfy the maximisation or minimisation process. i functional and non-negative constraints. It includes all possible components of a problem tan also tale a net-lnteger, vatne. to optimise the solution. 3) Additive: This means, the function value is the 2) Decision Variables: The objective function uses sum of she cootributlons of each term, Le, variables to reach ata solution. The variables that when (wo of more activities are used, the total help to decide the outcome are called ‘Decision + product is equel to the sum of the individual Variables? or ‘Activity Variables’. The level of products where there is no interaction effect activity for éach variable is specified. However, between the uciivities. it can specify a zero value of some variable but not a negative value. Decision (or choice) separately from A and B must be equal to the ‘variables can be X1, Kare. -oKae total profit earned by the sale of avo products A. 3) Constraints: Constraints are a pant of every and B, Similarly, the sum of resources used for practical situation, The mathematical program A and B individually must be equal to the slated in an algebraic form will also specify amount of a resource consumed by A and B. constraints, i.c., the limited availability of Linearity: Linear programming requires resources, if itis a maximisation problem or a lineasity in the equations, ie, the relationship minimum quality or composition, if it is a in both objective function and constraints must minimisation problem, Thus, every problem be linese. Linearity requires that a change in a will come with its limitations and accordingly, variable stroutd result in proportionate change a strategy or an approach (o a problem is to be in that variable's connbution 7 the value of de 1. This is called a Constraint, i: function. For example, the resource i ee 5 Hines Sty if decision variable x= 5 hut the jon will be 10a, if x = 10, and so on, Sfepresenis the amount of resource activity j (decision variable), For example, the sum of thé profits eamedLinear Programing (Chapter 2) 2.5. Standard and — Canonical forms of Linear Programming Problem Canonical Form Canonical form of a Linear Programming Problem is expressed as: nise Z= ej; Mi ‘And the Constraints 2s: Days Sb. i and, x52 055 “There are a few characteristics of this form. Which are as follows: 1) Maximisation type of objective function, 2) If the conditions are not, non-negative, then all the constraints should be of < type. If we multiply both sides by ~1, then the inequality of > type can be converted to the inequality of type. 3). For a canonical form to work, all the variables vst have non-negative values, Any variable, say x, which is unrestricted in sign (ic, positive, negative or zero), can be replaced by: xy=X)-% Where xj and X, are both non-negative, 44) The right-hand side of each constraint must be positive. Standard Form Standard form of LPP is expressed a3: Maximise Z= Y1ojx, a ‘And the constraints ayy sds a andyj20; 0 j= ‘There are a few characteristics of this form. Which are as follows: 1) Maximisation or minimisation type of objective a 2.1.5.1. Reduction of LPP to Canonical Form ‘The following transformations reduce the general Iinear programming problem into canonical for 1) The canonical form requires the objective function to be in maximisation form. So, if the objective function is in “the form of minimisation, it has to be first converted into maximisation. The following relation converts the objective function from minimisation to maximisation: Minimise {(x) = -Maximise (-f()) 2) The constraint condition ean be changed from 2 to < by multiplying both sides of the inequality by-1. For example, the linear constraint ‘a + aga + sky eect Bink 2 Bie is equivalent to ight = Bgky ~Bg%I— +o isk S—B, 3) The constraint that is presented in the form of an equation can be replaced by two weak constraints in opposite directions. For example, uXt+ BAK AyKSt ooceeet MaKe = By is equivalent to ajxt + OaX2 + 2a8yt sweet aKa S by . is andl 3X14 9aX0 + AgKst -voeeot GR ZB, 4) For any variable, say x), which is unrestricted in sign can be replaced by yay Where u, and vj are both positive. If any variable x; $0, then substitute x) =~}. 21.8.2. Reduction of LPP to Standard Form ‘The following tansformations reduce the general linear programming problem to standard form: 1) Inequalities can be converted to equalities by adding or subtrscting non-negative slack variables. 2) For any variable, say xj, which is unrestricted in sign can be replaced by N=U-% A where and yj are both positive. 3 Many variable x; 0, then substitute x 2.1.6. ‘ Benefits x of Linear Decision: LP improves the isions. There might be other ig outside the problem. amming gives possible and22 (Unity 2) Cost-Benefit Analysis: LP is a versatile fechnique that helps in representing. real-life business situations. It is a very cost-effective technique and is helpful in planning and ‘executing the policies of the top management in cost effective manner. 3) Flexibility: LP provides better tools for meeting “the changing conditions. Re- evaluation for changing conditions can be done, even after the plans are prepared. LDP is a very effective technique under such changing circumstances. 4) Number of Possible Solutions: Management problenis are complex, but with LPP technique, Managers can arrive at the best allemative solution to the problem. LPP helps in assuring that the manager is considering the best optimal solution, 5) Use of Productive Factors: The Lincar Programming technique belps in making. the best possible use of the available productive resources such as time, labour, machine etc. A decision-maker can employ his productive factors effectively by selecting and distributing these elements. 6) Scientific Approach: LP is effective as it highlights the bottlenecks in the production process. This means that an LP presents a clear picture of the problem. Hence, it becomes easy to deal with the problem. 2.1.7. Limitations of Linear Programming 1) Linear Relationship: A primary requitement of an LP is that the objective function and every constraint must_be linear. However, in real life situations, many business problems can only be expressed in a non-linear form. In such situations, LP technique is not epplicable. 2) Coefficients are Constraints: LP assumes that all values of coefficients of decision variables are stated with certainty. Due to this restriction, LP cannot be applied (o a wide variety of problems. 3) Fractional Solutions: Many times the solution to a problem may not be an integer but a fraction. Solution in fractions may not remain optional in rounding off. 4) Complexity: There are computational difficulties when it comes to large problems. LP model is a mathematical formulation which becomes complex when there are large number of variables and constraints. MBA Thind Semester (Operations Research) MOU 5) Possibility of More than One Objective: LP deals with the problems with single objective But in real life situations there are more than fone objective. LP faces limitations in such situations, 6) Time Effect: The effect of time is not considered in linear programming model. 2. Scope of Linear Programming ; Linear programming models can’be applied to a variety of business problems. Following areas make use of this tectinique: 1) Fingnee: LP techniques are very important for the finance sector. It is very useful in profit planning and control. LP can be used to maximise the profit margins by making an optimum use of financial resources. 2) Industrial: Production costs can be put under check with the help of linear programuning. LP finds the best solution under prevailing constraints and helps maximising output and tminiznising costs. i 3) Administrative: The administrative tasks become easier and efficient with the use of LP techniques, It helps the manager to choose and decide the best method out of various managerial methods for achieving the desired output, 4) Defence: LP is extensively used in military operations. In the defence sector, LP can be used to utilise the optimal limited defence resources to achieve the desired goal, 5) Trade: Linear programming can be used very widely in this sector. It helps in the estimation of demand and supply, price anc:cast ete. 6) Transport: In the tansport sector, LP models are employed to determine the optimal distribution system. It also helps in ascertaining Gat minimum cost is incurred on transportation, 2.1.9, Formulation of Industrial and Business Problems as Linear Programming Problems Mathematical model use for LP is known as Formulation of Linear Programming Problem, Steps of LPP Formulation 3 LP requites the formutation of a ‘model mathematical form using various symbols, basic steps in formulating a linear proge model are as follows: }Linear Programming (Chopiet2) Stop 1: Idsatifying the decision variables (values of which are found by solving the LPP.) and assigning the symbols Xi, Ris. OFX, Ys» t ther Step 2: Identify the objective function to be ‘optimised (minimised or maximised) and expressed interms of pre-defined decision variables as below: TECK + CDI on Cay Step 3: Identify all the constraints in «given problem, which restrict the operation at a given point of time and express them as linear equations and/or inequalities in terms of pre-defined decision variables a5 below: ayy + ake +. (=, 2)b,, where i= 1, 2, Step 4: Since the negative values of decision variables do not have any valid physical interpretation, write the non-negativity conditions as: 4120, %220, ... orx,2 0, where j= 1,2, ...0 ‘Step 5: In the last stage, form a linear programming problem model by putting the objective function, linear constraints and non-negativity conditions together in the form of an equation, Example 1: A firm uses lathes, milling machines ‘and grinding machines to produce two machine pats, Table given below represents the machining, times required for each part, the machining times available on different machines and the profit on ‘each machine part. Typrot | Machining Time] Maximum Time machine | Required forthe | Available per Week Machine Part (minntes)| —(inutes) M, M, [rates 2 6 3.000, [Mining 4 10 2,000 lmachines [Grinding | 2 3 300 machines Profit per) 40 | 100 unit Formulate the problem so that the number of parts 1 and Il to be manufactured per week t0 maximize the profit. Solution: Formulation of L.P. Model 1) Objective Functio ‘The first major requirement of an LPP that we should be able to identify the objective functions. It will be maximised of minimised. Mathematically the ‘objective function relates the variables which We are dealing in the problem, In. this problem we could be obtained by producing and selling two machine parts (M,) and (My) Let x) and xy cepresent the nurnber of time required for producing and selling M, respectively. me Nand Mi 2 Now, Machine part (M;) is obtained profit 40 and Machine part (M2) is obtained profit 100 respectively, So, objective Max @) funetion, (0 maximise the profit. 40x + 100K is the objective ‘onstraints/Conditions (Constraints are on > Time Avatle on Each Meching The mathematical relationship which is used 10 ‘explain the inequality in the variables. ‘The inequalities can be expressed in terms of less than or equal (S) and in the terms of greater than or equal (2). Each part of machine (My) required 12 minutes for Lathes and Machine (My) require 6 minutes for Lathes. The total maximum time is available of Lathes machine {3,000 minutes. ‘We can express the constraints for Lathes as: 12x + Gx2 $3,000 Each of machine part (Mi) require 4 minutes and machine part (M3) requite 10 minutes for milling machines and total maximum time available is 2,000 minutes. We ean express the ‘constraints for milling machines as: xy + 10x $2,000 Bach of machine part (M;) will require 2 minutes and machine part (M2) will require 3 minutes for grinding machine and total maximum time available is 900 minutes. We can express the constraints for grinding machine as follows: 2x; +3x2 $900 ‘Therefore, the subjective function is: For Lathes, 12x, + 6x, $3,000 For milling machines 4x, + 10x2 £2,000 For grinding machine 2x, + 3x: $900 3) Non-Negative Conditions: x, and x; art the number of machine parts produced by machiat (MI) and machine (M,) and they cannot have negative value. ‘The non-negative condition is expressed as: Xue 20 Now, we can write the problem in complete form of LPP as follows: Max (Z) = 40x, + 100x, (Objective functions) Subject to, “12x, + 6x, £3,000‘Linear Programuning (Chapter 2) ‘unit of product Y also results in two units of a by- product Z at no extra cost. Product X sells at profit Of {10/unit, while Y sells at profit of T20/unit. By Product Z brings a unit profit of %6 if sold; in case it cannot be sold, the destruction cost is t4/anit. Forecasts indicate that not more than 5 units of Z can be sold. Formulate the problem so that the quantities of X and ¥ to be produced, keeping Zin mind, $0 that the profit earned is maximum. Solution: Formulation of L.P. Model Step 1: The key decision to be made is to determine the number of units of products X,Y and Z to be produced, Step 2: Let the number of units of products X, ¥ and Z produced be xi, x2, Xs, where 4X5 = number of units of Z produced = number of units of Z sold + number of units of Z destroyed = x3 + x4 (82y). Step 3: Feasible alternatives are sets of valucs of Kis Xz, Xp and Kg Where Xi, X2, Xs, X42 0. Step 4: Objective is to maximise the profit, Objective function (profit function) for products X and Y is linear because their profits (€10/unit and 20/unit) are constants irrespective of the Humber of units produced: A graph between the total profit and quantity produced will be) a straight’ line However, a similar graph for product Z linear since it has slope + 6 for first part, while a slope of —4 for the second. However, it is wise linear, since it is linear in the regions (0, 5) ‘and (5, ¥). Thus splitting x, into two parts, viz. the umber of units of Z sold (xs) and number of units Of Z destroyed (x,) makes the objective function fo: product Z also linear. ‘Thus the objective function is maximise Z = 10x, + 202 + 6x3 = 4x4. Step 5: Constraints are: On time available on operation I: 3x; + 4x2 $ 20, ‘On time available on operation II: 4x; + 5%2 $26, On the number of units of product Z sold: x; <5, ‘On the number of units of product Z produced: xq = 3 + X4 OF 22 + Xy+X4= 0 and + ay Kay X$y X20 2.2. SOLUTION OF LINEAR) PROGRAMMING PROBLEMS(LPP). Linear Programming problem (LPP) can be basically solved by wo methods. Tis is shown in the figure below: [Methods of Solving LP Problems | — ee Graphical Method ‘Simplex Method LPP wich involves |] (LEP vtich involves tw or ‘only two vaiables) ‘mor than Evo variables) — ‘Simple Case 1) Big Metnod [2) Two Phase Method (Only slack variables ae added wile convecing the problem soto (Surplus varable ae standard form in ieatty macs DRA al ail A Yarables are aed to 2S entity atic) a 1) Graphical Method: Graphical method is @ very basic method of solving linear programming problems. Equations are tepresented in'a graphical form and are solved using the intersection points and finding the values of the bounded area, 2) Simplex Method: The most general and powerful technique to solve linear programming problem is Simplex method. Itis an iterative method which solves the LPP in a limited number of steps or indicates that the problem has unbounded solutions. Simplex method is designed to solve a number of linear equations simultaneously with _more/less variables. The* computational procedure is repeated until an optimal solution is dete 2.31, Introduction Graphical method indicates the constraints and determines the ‘feasible region’ on’ the graph, The feasible region is the area that contains all possibly feasible solutions to the problem, ie, those solutions which satisfy all the constraints of the problem, The graphical method is applicable to solve the linear programming problem which involves two decision variables. One can arrive at an optimal solution to LPP by evaluating the value of the objective function at each vertex of the feasible region. It will only occur at one of the extreme points, For example, et consider the following LP problem: Maximize Z=4x+3y ‘Subject to the restrictions, xtyss x+2ysB8andx),m2026 (Wait In the figure 2.1 points B, C, D, E is vertices for the feasible region shaded area. . Pigure21 2.3.2. Steps of Graphical Solution ‘Steps involved in graphical method of LPP are as follows: Step 1) Formulate the linear programming problem. } ‘Step 2) Inequality in the constraints is converted as " ‘equality, Step3) Constraint lines are consinicted as equations Step 4) Then “feasible” solution fegion is Step 5) Then corer points of the feasible region are located. Step 6) On the comer point, the value of the “objective function is calculated. Step 7) The point where the objective function has «optimal value is selected. Step 8) On. the same graph Paper, objective function is also constructed by. assuming some arbitrary value of Z. ‘Step 9) The value of the objective function can be calculated ifthe optimal values occur at the omer points of the feasible region and the one which gives the optimal solution is selected. 2.3.3, Maximisation Programming Problems Example 5; Use the graphical method to solve the following LP problem. Maximise Z=3x,+ 2x; Subject to the restrictions, 2ny$x 540 + ky +mS08 2x, 4 3%;5 60 and x), x:> 0 Solution: For the purpose of plotting the above equation on the graph, one needs to convert ‘inequalities into equalities and find out the point of line. Linear MBA Third Semester (Operations Research) MDU_ nena Beanies feasible retin 24) 15840 The feasible region is given by OABCD. The value of the objective function at each of these ‘extreme points is as follows 3,425, S50 59%0=0 ‘The maximum yaluc of the objective function Z = 64 occurs at the extreme point B (16, 8). Hence ‘optimal solution to the given LP problem is: 5,=16,x, Example 6: Solve the following LPP Maximise Z = 2x; +x; Subject to: x: + 2k 10, n+ns6, Xm 52, 2x21, andxy, x0. Solution: For the purpose of plotting the above equation on the graph, one needs to coavert inequalities into equalities and find out the point of Aine,making cei egies hours work by. skilled man and ihe ofinny cel Fequires 1 our by'a skilled maa. Thus we have, 2x +y S40 The making of a deluxe model requires 2 hours Work by semi-skilled man ordinary model requires hours work by a semi-skilled man. Thus we have, 2x+3y 580 Total profit = Z=10x 48y, which is to be maximised. ‘Thus, the mathematical formulation of the given linear programming problem is, Maximise Z=10x + 8y Subject to 2x+y S40 2x 449/580 x20 y20 For the purpose of plotting the sbove equation on the graph we convert inequalities into equalities and find out the point of line. o. x 3300 ¢450= 900 “[_#, (150, 300) [31504 45-300 = 1650 c (0.420) [3x04 4x420=1680 the problem is of maximisation and the ‘maximum value of Z is attained at a single vertex, hence this problem has & unique optimal solution. ‘The optimal solution is x; = 0, x2 = 420 and ‘maximum value of Z = 1680, c le 9: A small-scale manufacturer employs fivé skilled men and ten. semi-skilled men. He article in two quantities, a deluxe mode! ordinary model. The making of a deluxe fequites 2 hours of work by a semi-skilled a SBA noter ack Ordinary model requires 3 hours of work by semi-skilled and 1 hour by a wy ‘The feasible region determined by the system of Consicaints is shown in figure 2,2: Ont y=40 ly] 2x 44y =80 Oe Point 11 } x+y =40 20 y=0. ae 2e¢3y = 80 x=40, y=0 (300, 0)‘the optimal solution to the given LP problein a 2 a 4S = Y= 7p and minimum value of Z 900, Ress? in Example 11; A farmer is engaged in breeding pigs. ‘The pigs are fed on various products grown on the farm, Because of the need to ensure nutrient ‘constituents, it is necessary to buy additional one or {wo products, which we shall call A and B, The amutrient constituents (vitamins and prot of the product are given below: Minimam Requirement of Natricat Constituents Product A costs €20 per unit and product B costs 140 per unit: Determine how much of products A aand B must be purchased so as to provide the pigs ‘nutrients not Jess than the minimum required, at the Towest possible cost: Solve graphically. Solution: Mathematical formulation of the above: nim Z=20n +40, (Cost 36x, +61, 2108 ) function) 3x, +12: 236 20x,+10x, 2100] (Nutrient nung20 Constraints) (Non- negativity)The value of the objective function at each ofthese extreme points is as follows: “sina, | Vie of Obie) [Corner Poiat|Coordinates| Functions 2.220%, + dex, a oie 70 | B 26 20 fa 42 10 D 120) 240 Here, one find that minimum cost of £160 is at point C(4, 2), So, x) =4,m=2,2=160 Thus, the optimum product mix is to purchase 4 Units of product A and 2 units of product B in onder to maintain minimum costs of £160. Example 12: Solve the following LPP Minimise: Z = 3x, + 2x, ‘Subjectto; 5x; +x; 2 10, a x +x26, i _ mi44n2 12, 41,%220. MBA Third Semester (Operations Research) MDU. Solution: For the purpose of plotting the above equation on the graph, we convert inequalities into ‘equalities and find-out the point of ine: Ski +x= 10 witG=6 H+ 4x,=12 [Point 5m += 10[5)=0, = 10 0. =6 0.8: x Xt4n= 12) x, The region is unbounded upwards and to the Tight of ABC and is shown in shaded atea in the figure above, The value of the objective function at each of these extreme points is a8 follows Extreme Point Objective Function Coordinates Value (x) Xx=3x, + 2x, A(O, 10) 3(0) + 2(10) BUl.5) 3(1) +265) = 13 c (4.2 | 34) +20)=16 DU20 12+ 20) =36 ‘The minimum value of objective function.z = 13 Sceurs at the extreme point (1, 5). Hence, the optimal solution to the given LP problems i ¥=5 and min-z= 13,Sad a4 METHOD 24.1. Introduction i solving a linear programming fe . Problem which involves two or ‘More variabl i les, simplex. method It is an iterative method and is based” on a 3 i the Pree ot OF obtaining an optimal solution to the Problem by the extreme point of the feasible tegion, 2.4.2. Principles of Simplex Method 1). Simplex method is based on the assumpion of 6 40 optimum solution to LPP by the ‘xtreme point of the feasibie region, on 2) It's an iterative process for searching + beter Fomex point until the yalue of objective function is improved by shifting to nother comer. 2.4.3, Steps of Simplex Method The steps involved in simplex method are as follows: Step 1) First the linear programming problem is formulated. Step 2) Initial simplex table with slack variables in maximisation solution is set-up. ° Step 3) Inclusion of decision variables is determined. Step 4) The variables that are to be replaced are determined. ‘Step 5) Then the new raw values for new variables are calculated Step 6) Next, the remaining rows are revised. Step 7) Until an optimum solution is reached, steps 3 t06 are to be repeated 24.4. Slack, Surplus and Artificial Variables In Simplex Method, the inequalities with < or > Signs are changed to equalities using slack, surplus and artificial variables, Let us understand these variables: 1) Slack Variables: Slack variables convert < inequalities into equality. For example, ayxtay inequalities into equalities. For example, aX+aY>b ; canbe redefined as, ayxtasy-Sr=br i Here, 5; isa surplus variable. 5 ~ Surplus variable = Production: = Requirement Surplus variables are the additional amount over the required minimum level and are-ttsd ‘on-negative in nature. ~ 3) Artificial Variables and Its Uses: In miltiy Gases, constraints are represented in’2-a combination of," and = in equatigas; Sometimes, this problem is not solved even afler introducing surplus variable in the equation. In such a condition, artificial Variables are introduced to arrive at a feasible solution. These artificial variables do not have any” physical identity and are completely fictitious ne Anificial variables are.designated as — M for maximisation problems and + M~ for tninimisation problems where M per unit is Nery large penalty assigned. ia objecqye function. The summary of extra variables needed to addin the given Linear Programming Problem to convert it into, standard form, ie given in table 2.1 Ss Table 24 Types of |" Extra | Coefficient of | Presence | Constraint | Variable to | Extra of beAdded | Variablesin | Variables | Objective | in tnitiat Function __| Solution Max | Min z =A Less than [Addonly | 0. | 0] Yer ‘orequal to | slack Hi) | variabie Greater | Subiract 0 | © | Ne thanor | surls “ jeqalto()| varableand | | sat | yeg ald Anificial variable % Equilto [Add only | —M | aN Ye © aniical : variable ugSolution: Rewrite the inequality of the constraint into an equation by adding slack variables S,, S and Sythe standard form of LPP becomes; Max Z = 15x; + 6x + 9x + 2+ 05, + 082-4 as, Subjeetto constraints 2x1 + x24 5x5 + 614+ S) + 0S; + 08,= 20 : 3x 4 x3 + 3x5 + 25,408) +S, + 05) = 24 $ Tx; + Oz +013 + Tx, + OS; 40S, + Ss =70 : %15 3%, X4 Si, Su S320 ‘The initial basic feasible solution is $, = 20, S; = =%=% == 0 non basic) ‘The Initial simplex table is given by: ‘Simplex Table I: Non Optimal Solution. PARAS 1s} 6[t{olo(20) 207: 2 1/3]25]0]1/0}24] s¢vtinimum Value) —> a} 0} 7{olo] 1 |70) Tor 10 [9] fofolo . ofefofoloro “ =|-[-2)olaro (Conesponding valucof keyrow) a 2 _ ‘Commnngaet eyes) Tncoming t a + variable « As some of %— ¢ <0 the current basic feasible solution is not optimum, 2; - c, =~ 15 is the most negative value and hence x, enters the basis and the variable S2 leaves the basis. ‘Table 1: Nan Optimal Sotution Is;|_S._ IS Zann [if =z0 oa 3 33 =) 258 |0] us |e vs | fra] |‘nea Pogaraing (Chapte2) 2) When Max Z* = 0 and a last one zero value atifcial variable is present in the bass, then the original LPP. will hae feasile solution, In such a case, proceed to Phase 2 to get the basic feasible solution, We may proceed directly to Phase 2 inorder to get basic feasible soltion or else proceed 10 Phase 2 after eliminating the artifical basic variable When Max 2* = 0 and there is no artificial variable in the basis, then a basic feasible solution to the orginal LPP. has been found. Goto Phase 2 2 Solution: Phase I: First we introduce surplus and artificial variables, and then rewri 82:0 coefficient tothe decision variables and a coefficient | tothe arti Minimisation Z=Or, + 0x, +08; +0S:+ A, + Subject t9 Ri, Si, SA, Ar>0. Sinpes Table | tn Vals pee as uo Since allz)-«; £0, so this is not optimum solution. So we sel ‘Simplex Tble 1: Non Of Phase Il Step feasib object and a appear at zer0 value Tmmple 27: Soe Goat tworpase metod. The poten Manse Subject 20% + S0xs—S, +05; +A, +00; = 4800 80x; +50%+05; -$:+0A, + A,— 7200 ‘Tesolin to this problem shal be obtained using simplex raethod.Itis given 42 This phate sta lube of Pe en ve freon a sa “0 ale 1 te Sa Z= Mis ty 10 20% + 50> 4909 “ei 804, +5042 7299 4%20, : 2 oh aa ite the objective function cial variables. The problem Since Zp ~ Css most poi aon intable 105 | [mes Non-Optimal Solution ; t 3 on zs iB Bae — a) es [o pipe] a0 LU @)>|2 lata ow since s8Z,-¢, <0. (ns ee ss Be woof] TT] The opel sion ge al Example 2 tac meh ine et mex postive 5 ; tier timal Satutios ei Toyo. oy 0 ay1 0 24 A=Z-C) 0 a (| 3805 id the basic variable x, leaves the positive, the current feasible jsolison is not optimum (Minimisation Case), Therefore, S; F : lex Table V: OpinalSaation He 40, Vi 24 oO oO 1 G Basic pesenles Sakai a mr = Si s a x 144 a5, 1 | 0 | = =150 0 St 2400, oo | 0 1 =I Z 194/5 24 0 =12/25 A= %4-C =8/5 0 [0 | 1285 all Z,~C, <0. (Minimisation case) this is optimal solution, {solution given by Simplex table V is x, "Example 28: Use two-phase simplex method to solve Maximise Z=5xi-4m+30 Subjectto — 2xi-#%:-6%)=20 ag 6x, +5 x2+ 10%) 576 Bx —3x + 6x5 550 x1,%%320 O and x3 = 144, T6and S: = 50. se Zt = 0/40 x1 + Ox; + 05, + 0S:— 1A; Dy +X -6 x; + Ay= ta t6x) +5 =50 i Sa Ar 20 Solution: Introducing slack variables S;, 52 = 0 and an artificial variable A, > 0 in the constraints of the given LEP, the problem is reformulated in the standard form. Initial basic feasible solution is given by Ay =Linear Progranuning (Copter 2) Table : Non Optimal Solution 2.493, Difference between Two Phase wasis [ie [5 [AL]S:[S,[b|b/ay and Big-M Method Af=t|2 [1 |-6) 4 [0] 020] a02=10_| Table below shows the difference between two (S| 0] 6[s lie] o| 1/0 }76) 76 = 12.66" ‘phase and Big-M method Ss] o[s |S] 6 (0) 0/1 [s0|swe=625 ‘Two Phase Method Big-M Method @_[o}ofol-ilolo : Te is easy fo obtain inital | It @ difficult 10) obtain 4 [2/=1/6 | /ol0 ‘basic feasible solution | inital basic feasible el2j-fe[o[olo bectute a zero co-efficient | solution becaute @ zero 7 is attached 1 basic | co-efficient isnot Tocoming variable ‘Since 2) ~e, and 2, — cis nepative therefore this is ‘not optimum solution for maximization case. ‘Simplex Table I: Nou Optimal Solution Basis fe, [my TANTS ST by | byes | Ai] =1 0 | ya |=isa].1]0 =v) in| san-> S| 010 [294] 12 1 | -34 17772 | 154729 x01 [va] 34 Lolo) im [24 sjofo | o [tool | x, | 0[=74] 1572. Oo] | =e [ol-ra} isa oop if Incoming variable Since z:— c: < 0 an optimum solution’ to the auxiliary LPP has not been obtained ‘Simplex Table I11: Non Optimal Solution attached 10 basic variables. This Gary and (eed time [Tt is tedioul because variates: consuming, variables (penalties) are involved. This method does not | This method uses inyolve any assumptions | assumptions for at the original system of | coostraints Constraints 2.4.10, ‘Special Cases in Simplex Method ‘While using simplex method for solving LPP, one may encounter many special situations which are summarised below: Special Cases in Simplex Method Sciitone Unbeunded Solutions Degeneracy Since all 4 q > 0, an optimum solation to the ‘uxiliary LPP has been obtained. Also Max Z* = 0 with no anificial variable in the basis. We g0 to phase TL. ‘Phase IT; Consider the final simplex table of phase L. Consider the actual cost associated with the ginal variables. Delete the artificial variable A; column from the table as itis eliminated in phase 1. ‘Simplex Table: ‘Basis : lelelsia[=le gel alee Bais [xfs] | A 15,1 5, 1] [Acca Opel =o oli tsaa| a7 fo lula S| noo) aserr| aon [1 | arr [sar] a 0 | tol —a7 | 34 [0 fins ss ebaslile Sols ofojo | —1 Jojo z ofo|-o | 9 jo] o a ofojo | o tito 2.4.10.1. Alternative Optimum Solutions Sometimes an LPP may arrive at an optimum solution but not @ unique one. This means that more than one optimum solution exists, In such cases, the following are the characteristics of the final table: 1), An optimum solution, 2) Forall basic variables, (Z, ~C;) values are zero: 3) For non-basic variables, (Z,~C.) values may or may not be zero. 4) For atleast one norcbasic variable, (2 ~ €) value will be zero, 5). More than one optimum solution is possible, 6) For all possible soiutions value of objective: function Z remains the same. 7) Basis in final table contains basic variable X, viz. Xi %2,... and non-basic variables S, vie, Si, Syee erates har rset ayynct bat Br AE sec hos tx ass lcm fs a ‘Heme, ths prchiem bas ao oxiqpe sciatica tet hss ape ope sh. 24102 Unbounded Solation Unteomatet shai aces wen Ger: crete i tion a trio rabies ane incon iitly oats Ga sae eweltons. Sock etborated whan tans is meiimion prions. vhese te inition Waite cn be mate ifiely lage sates oe se comstrits. Tis haggess in A sper metic za to in rhe 2 esi 4 replacement 165, ‘ae negative. Ths ipl : indefmtly widest deiving, one of vatiables to 2x0. The piven sohation is, bonded, =46 Wait plen Table Epive I solution is: xy © 4, x; mL sy Gand Z = 14. Now, inthe Sin Ble 1, a itive ne ‘of vero. This means abjectivefanton valve. Inet east eae made a basic variahle nd the resulting basic feasible solution will also have the objective function valic equal to 14, This means, an alternative optimal solution to this problem exists which can be obtained by 8 basic variable as shown below: a 1 (Opecatonn Heeart MEL MIVA Thin HExuupte 30; Maxine Zoe Say) fing 4 Subject to the constraints Yay May 22 8 AN Day ky 2 ayeday ay 8 Ais Xap Ki 20 ick Varlables ay, #3 ancl ay ‘Solutlon: tntecaluc e reste in landsid form oe the oblens cnn follows: Male Zoo 5x) 4 6x3} yO 1-Ony 4 Oy Sect to he eonstalis . Dey By Ant gy eS . Aq) + 2am hh 92 seta Hayes Hie Mr Ki Sip) 20 at Davie feasible solution iv ofva setting x) 8) 8 00th ‘Sotion xj] xy |xyfay|y) 4) Values f_. rauation (or Iden 3 {2 0 __ |G Tiasie | Solution | Varlnbles Vals G=Z, Thus alternative optimal solution is: are Sax,= Bandar Z=14 Hence, this problem has no unique solution but has multiple optimal solutions. 2.4,10.2, Unbounded Solution Unbounded solution occurs when there is no consteaint on the solution and the decision variables can be increased infinitely under the same cesieicted conditions. Such unbounded solution occurs in maximisation problems, where the decision variable can be made infinitely large under the same constraints. This happens in the iteration stage of simplex method when all the entries i inimarn ratio column are cither infinite or negative From Simplex Table Hl, one observes that the solution is not optimal beewse a positive value appears in the C= Z, row. Accwaing to this Xy is the incoming variable and the casresponding column is pivol column, Computing the replacement ratio, we notice that all the three ratios are negative, This implies that we can increase xy indefinitely without driving one of the asic Variables to zero, The given solution is, therefore, unbounded.