Reagjloltead RIM|Nagpur, University BCA Part-II (3° Semester) il’ Solved Series THAKUR PUBLICATION NAGPUROPERATIONS RESEARCH-I TP SOLVED SERIES For BCA [Bachelor of Computer Applications] Part-II, Third Semester Students of ‘Rashtrasant Tukadoji Maharaj Nagpur University (RTMNU)’ Edition 2019 Reprint Edition 2020 Copy right © Aut Rights Reserved This book is sole subject to the condition that it shall not, by way of trade or otherwise, be lent, resold, hited out, or cthenvise circulated without the publisher's prior written consent, in any form of binding or cover, other than that in which itis published and without including a similar condition. This condition being imposed on the subsequent purchaser and ‘without limiting the rights under copyright reserved above, no part of this publication may be reproduced, stored in or transmitted in any form or by any means (electronic, mecha jotocopying, recording or otherwis ‘ritten permission of both the copyright owner and the below mentioned publisher of this book. ithout the prior Books are Available for Online Purchase at: www.tpplorg.in THAKUR PUBLICATION PVT. LTD., NAGPUR + Lucknow + Meerut * Hyderabad * Pune * Jaipur * Chennai * Rohtak * + Jalandhar * Ahmedabad * Bengaluru * Bhubaneswar * Bhopal * Kerala * Kolkats Plot No. 109, Nawa Nakash, Behind Jaiswal Restaurant, Near Choti Maszid, Nagpur-440017. Mob. 09595029116, 8840084584, 9 318594 / 22.Syllabus UNIT-1 jon to Operation Researc! felopment of Nature of OR, Characteristics of OR, Introduction to Operation Research (OR) Origin and development of OR, Nature emcieslicerot OR Classification of Problems in OR, Models in OR, Phases of OR, Uses and Limitations of OR, Met Applications in OR. Linear Programming ~ Concepts of Linear Programming Model. Mathematical Formulation of the Problem, Graphical solution methods. UNIT - IL Linear Programming Methods ~ Simplex Methods, Big M methods, Dual Simplex Mettiod, TwoPhase methods, Duality in Linear Programming ~ Formulation of Dual Problem, Application of Duality UNIT - TH ‘Transporation Problem ‘Mathematical model for Transportation Problem, Types of Transportation Problem UNIT- IV Assignment Problem ~ Zero-One Programming Model for Assignment Problem, ‘Types of Assignment Problem, Hungarian Method, Branch and Bound Technique for Assignment Problem. Retailers / Distributors NAGPUR 1. Book World, Swami Samarth Complex, Opp. Home Mahila Sobhagruh 282, North Bazsar Road, Gokul Peth, Nagpur-#40010. Ph, 0712-2562999, Mob, 9422864426, 2. Central Book Shop, Near Jhanshi Rani Chowk, Nagpur. Mob. 09823073909 3. Laxmi Pustakalaya and Stationers, Kelibagh Road, Mahal, Naypur-4400032, Ph, 0712-2720379, 2727354 4. Pragati Book Depot, Near Rajvilas Talkies, Mahal, Nagpur. Ph. 2721354, Mob. 9823030242 5. Renuka Book Depot, Near Rajvilas Talkies, Mahal, Nagpur. Mob. 0976540613 6. Venus Book Centre, Opp. Rajaram Library, Ram Nagar Road, Mangalam Arcade, Dharmpeth Extn., Nagpur. Ph. 2520781, 2536314 7. Vijay Book Depot, 477, Gola Chakka Marg, Near Karachi Jeneral Store, Nagpur. Mob. 9122534217 WARDHA. & Gandhi Book Depot, Arui Naka Square, Near Sai Mandir Wardha. Mob, 8275285568, 9420682100, 07840979193, GONDIA 9. Shree Mahavir Book Depot, Pal Chowk Rail Toli, Gondia-441601. Ph. 07182-225032TP Soived Series (Operations Researct-l) RTMNL UNIT 2 SIMPLEX METHODS & DUALITY eee SIMPLEX METHODS Ques 1) Whit is simplex method of LPP? What are the different steps of simplex algorithm’? Or Discuss the principles of simplex method. Or Explain various steps of the simplex method involved in the computation of optimum solution to a linear Programming problem. (2016 [05}) Ans: Simplex Methods One can conveniently solve problems with two variables. If we have more than two variables, the solution becomes very cumbersome and complicated. Thus, there is 2 limitation of LPP in solving all types of complex real life problems, where the variables are always more than two, For such linear programming problems solutions can be found with the help of simplex method. For solving a linear programming problem which involves ‘two or more variables, simplex method can be used. It is an iterative method and is based on the assumption of obtaining an optimal solution to the problem by the extreme point of the feasible region, Simplex method is an algebraic procedure in which a series of Tepetitive operations are used and we progressively approach the optimal solution. Thus, this procedure has a number of Steps to find the solution to any problem, consisting of any umber of variables and constraints, bowever problems with ‘more than 4 variables cannot be solved manually and require the use of computer for solving them. ‘This method developed by the American mathematician G.B Dantizg, can be used to solve any problem which has 4 solution, The process of reaching the optimal solution ‘hrough different stages is also called iterative, because the same computational steps are repeated a number of times before the optimum solution is reached. Steps of Simplex Algorithm’ Steps of the Simplex Method ‘The steps involved in simplex method are as follows: Step 1) First the Jinear 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. ‘Supp 7) Until an optimam solution is reached. steps 3 to 6 are to be repeated. Principles of Simplex Method The principles of simplex method are as follows: 1) Simplex method is based on the assumption of obtaining an optimum solution to LPP by the extreme point of the feasible region 2) This an iterative process for searching a better comer point until the value of objective function is improved by shifting to another comer. Ques 2) Write short note on the following: 1) Slack, 2) Surplus and 3) Artificial Variables 4) Basic and non-basic variables Or ‘What is Artificial Variable? (201802) Or Define the terms: (2018(244)) i) Slack variable i) Surplus variable iit) Basic variable and non-basic variable. ‘Ans: 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 inequalities into equality. For example, ax+aySb) can be redefined as, ax + aay +S, =b; Here, S, is the slack variable Slack variable = Total resource — Used resource Slack Variables are the unallocated portion of the given limited resources and are non-negative in nature. These variables help in a more detailed ‘economic understanding of the solution. 2) Surplus Variables: Surplus variables convert > inequalities into equalities. For example. aX +a, ¥>b; can be redefined as, ax tay -S:=b: conven < Here, S;is a surplus variable Surplus variable = Production - Requirement Surplus variables are the additional amount over the required minimum level and are also non- negative in nature.Simplex Methods and Duality (Unit 2) 4) The concept of basic asic Variables is associated with the solution of the linear programming problem with multiple decision variables. The LP model with multiple decision variables can be explained by using the simplex method, ‘The general form of LP model where multiple decision variables are used can be defined as Max z = cx, Subject to the conditions, Ax=b x20 Here, c is a row vector defined as ¢ = (€}y €2-..Cy)sX is the column vectors and defined as x column vector b can be represented as b ba)", and A is the coefficient matrix, The general form of the solution of the above- mentioned LP model is xs] _[B! xx} Lo Here, xs = (Xi, Xb2. +--+ Xo)" and the matrix Bis known as the basis matrix and the matrix has m linear independent row from the coefficient matrix A. This solution is known as the basic solution for the mentioned problem. So, the basic variables can be defined as the m variables which can take any value other than zero. Moreover, if the variables satisfy the non-negativity condition of the LP model, the basic solution created by them is called the basic feasible solution ‘The remaining variables are known as the non-basic Variables. The values.of the non-basic variables are set to be zero. If the LP model is answered by using the simplex method, all the slack or surplus variables will be the basic variables for the first iteration. Artificial Variables: In many cases, constraints are represented in @ combination ofs, 2 and = in quations. Sometimes, this problem is not solved even her 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 ‘and. are completely fictitious. Artificial ident ‘enipies are designated as — M for maximisation problems and + M for minimisation problems where M per unit is a very large penally assigned in objective function “The summary of extra variables needed to add in the jing Problem to convert it into given Linear Programming Standard form is given in table 2.1 bar ste 2 inp |e Coctident 0 | Presence cree | vate | ese |e Ww Added” | Varies tn | Varabes Oofectve | ttn Functlon | Soludon Wins |" faieay po 0 Ne equal (2) | | Greater than | Subtract 0 0 | No oregano | aun ® wrtteand | me [om | Yen wu aici | vate | Taarwiey [Attonty [=m Po | Yes ie Li [variate L Ques 3) What is difference between slack, surplus and artificial variable? ‘Ans: Comparison between Slack, Surplus and Artificial Variables ‘Table 2.2: Comparison of Variables _ Particulars | Slack | Surplus | Artificial Variable | _Variable_| Variable Mean Unused | Exoess No physical Resources | Amount of | oF economic of dle | Resources| meaning. tis, Resources. | utilised. fictitious Whei to be 2 and © used inequality | inequati Co-efficieht | +1 = in the constraint | _ Co-efficient | 0 a 7M for in the minimisation objective and -M_ for | tunetion 7 maximization ‘Use as Initial | Used as | Cannot be | Initially used program — | starting | used since | but iter on variable point tunit matax | eliminated Aitiat | condition 48 Table or | not satisfied tableww i) | Presence in| Helps 1 | — Indie | the Optimal | interpret Ifeasihle Table idle and Solution key Ques 4) What are the advantages and disadvantages of simplex method? ‘Ans: Advantages of Simplex Method 1) Simplex method is used for those problems which ‘have more than two decision variables 2) Simplex method is a well-known and easy 10 use method. Through this metho, one ean easily program the algorithm on computer. For computation, any Function can be quickly included in the program software —<$—$— .Dam by just altering the function evaluation. The method is lume consuming when done manually but is easy and quick when done with calculators and computers. This ability of the method to program it on machines. has ‘made it popular in advanced mathematics 3) Though the method is simple the identification of errors can be difficult. The simplex method unlike the ‘graphical method deals with an LPP having more than two decision variables. Disadvantages of Simplex Method 1) Simplex method has its limitations when solving LPP. It is used only in those business situations where a decimal quantity is appropriate, Also, it can be used only when more than two variables are used in a problem, Though this method is very efficient in these situations, but in real-life situations this method does not work because there are hundreds of variables, 2) Simplex method is best adapted in a certain LPP only. ‘This method is used only when the problem is expressed in a standard form with three conditions. Another disadvantage is that the values of the constraints must be ‘non-negative for all variables and must be expressed in = form, with the RHS value to be positive. Ques 5) What are the applications of simplex method? Ans: Applications of Simplex Method Following ate the applications of simplex method in different areas 1) Production: This method is very useful in production situations. With a given set of inputs, the method produces multiple outputs. This helps the manager to decide the best combination oatput to be produced to ‘maximise the company’s profits. For example, dairy product manufacturing company has to decide that for a given quantity of milk how much butter, ghee and cheese has to be produced. In such production situation, simplex method comes handy and useful in ‘minimising costs and maximising profits 2) Advertising: Advertsing is an integral part of any business and there are various media for doing so such as television, newspaper, magazines, radio, hoardings, Pamphlets, etc. All of these have different costs and varying levels of effectiveness. The frequency of advertising is also considered. These variables have to be addressed under the resource constraint of the money that can be spent on advertising. 3) Distribution: Consider the case of a firm having its Production units located at three different places A, B and C. This is a complex multivariate’ problem. Distribution units of this firm are spread over three other places X, Y and Z. The transportation schedule from the production units to the distribution units can bbe decided with the use of simplex method, However, this transportation schedule is worked-out under some inequality constraints. For example, cost of transportation has to be minimised, and the product supply and demand requirement should be satisfied between the production unit and the distribution unit BCA, Third Semester TP Solved Serter (Operations Research) RTMNL 4) Investment: Simplex method techniques help in making the optimal decision. This is useful in investment situations as well. When a firm seeks to invest but is faced with various available alternatives. then the method helps in making a cheice For example, a firm wants to invest £20 crore and bay three alternatives A. B and C for investment, Expanding the existing production capacity is represented by altemative A, Starting manufacturing a new related product is represented by altemative B and diversifying into a new venture is represented by altemative C. Each of the three alternatives involve a cash outlay of 710 crote. Which of the alternatives has to be chosen is 10 be decided by the firm using the simplex method 5) Inventory: Managing an inventory involves accuracy ‘The manager of the inventory faces two major challenges. One is what to order and other is how ‘much to order, Using simplex method, these problems can be effectively handled. Ques 6) Write the simplex algorithm for maximisation ae? Or Discus the steps of maximization case of simplex method. Ans: Simplex Algorithm (Maximisation Case) ‘The steps for the simplex algorithm for obtaining an ‘optimal solution to an LPP are as follows: Step 1: Linear program is converted into a standard form. Step 2: All right-hand side values of the constraints are checked for non-negativity. In case of a negative value, the corresponding value in equation of the constraints is ‘multiplied by ~1 to get the non-negative values. Step 3: ‘The inequalities of the constraints are converted into equalities by using slack or surplus variables. Costs of these variables are put equal to zero Step 4: An initial solution to the problem is obtained and is putin first column of the simplex table. Step 5: Net evolutions A, = ZC, (j ‘computed by using the relation Z-C=CeX-C, sn) is Examine the Sign i) Hall the net evolutions are non-negative then it can be said that the initial basic feasible solution is an optimum solution. ii) In case of atleast one negative net evolution, we have to move to the next step, ‘Step 6: In case of more than one negative net evolution, choose the most negative of them, ‘The corresponding column is called entering column, i) Inthe entering column. if all the values are <0, then it can be said that the solution to the problem is ‘unbounded, ii) If atleast one variable has a value > 0 then it can be Said that the corresponding Variable enters the basisSimplex Methexls att Duality (Unit 2) Step 7: Ratio of Xy/Entering column is computed and the lowest of these ratios is selected, The row that cormespons 10 this lowest or minimum of ratios. is known as leaving row, The con both, the ente as the key ek mon clement that is ‘column and the leaving row ‘AC oF pivotal element present i known Step 8: This key clement is divided by the heading clement to convert it (0 unity, AML the other elements are divided using clementary row transformation to convert them to zeros. Step % The computational procedure instep 5 is ‘peated ll an optimum solution is reached or an unbounded solution to the problem is indicated, Ques 7). Solve the following LPP by simplex method: Maximise Z =100x + 60y + 402, ject to contraints x+y +S 100, 10x + 4y + 57.5 600, 2x + 2y + 62300, and x20,y20,and220, Ans: To convert it into an equation, we add a slack variable S), Ss, and Sy on the left hand side to get, x+ytz+S)= 100 10x + 4y + 52+ S2= 600 For Sy row 1 2 1 6 i- ° o- o o ye " 100 = 600L = 40300 so ‘Simplex Table II: Non Optimal Solution Basis [Tx | y | [Si 8 [Ss] bay at [66.675 (Minimum isit 0 | o | us |u2] 1 |-1710]0| 40 yatvey > [x fioo) 1 [a5 lief [10 [oo [iso SsLo [o [ois s [os [1 [1a0|150 ¢,_[i0d] 60 [iofoto fo [100] 40 [s0fo [10 [0 0 [-20f1ofo| 10 fo Incoming variable Since value of 22 ~ 2 is not optimum solution. New entry (Corresponding valuéof key row) (Corresponding valueof key column) 20 (<0), {max case], Hence it 2x + 2y + 62 + Ss = 300 Oldvane~ ee ‘The problem can now be expressed a follows: Max Z=100x + 60y + 407 +08; + 0S: + 05s. For Row x For Row Ss 1-0x2x321 0-0x82-0 Subject to constraints 3 K+ y+ 2+ S,+0S2+0S;= 100 1615 2xSxda0 10x + dy + 5z +05; + Sp-+0Ss = 600 a) 2x + 2y + 6240S) +0S2+ S3=300 s duc$nd ‘Simplex Table I: Non Optimal Solution bi Boe one Sx5 wasis| x | y_| % [Ss{Ss]Ss) bi [bvav 33 S[o) 1 | 1 [2 [1{ofofseo| 100 5, |o| & | 5 [| ¢|0 [600] 60 (Mtinimum| |Vatue) > s fol 2 [2 | & [olo| [300] 150 «, | 100 | 60 | 40 [olo}o 3 To lolol Simplex Table IIE: Optimal Solution _ was [x Ty] 2 | ss | ss [St ~eo | 40] 000 | 5 OO ar fio} 1 fo | v6 [= 25 [6 | 0 | 1008 | 5 A A coming vara [100 Too | a0 [0 [0 Po Incoming variable [100 | 60 | 2007 | rows | 308 [ot a-6_10 10 [160% | 100s | 208 [0 4 <0, fmax ease], Hence it is not Since all value of 2% ~ optimum solution. New entry (Corresponding valueof key row) x (Corresponding valucof key column) = Old value= ee em Key number ‘Thus, the optimal solution is x Max Z = L00x + 60y + 402 (00/3, y = 200/3, 2 = 0 +602 0x0 3 = 3333.33 + 4000 + 0 = 7333.33p-w Ques 8) Solve the following LPP using simplex method. Max 15x; + 6x: + 9x44 2X Subject to 2x; +x; + Sxy + 6x45 20 3x) +42 + xy + 28x45 24 Ty + Tq S70 and X45 X25 X3y X40 Ans: After adding slack variables ), S; and Sy the standard form of LPP becomes Max Z= 15x, + 6x2 + 9X) + 2xu4 0S) +082 + 0S; Subject to constraints 2x) + x2 + 5x) + 6xy+ Sy + 0 3K; + Xa + 3X3 + 25x4 408) + 82+ OSs Tx; + Ox: +0K3 + Tae Xie 2 M4 Kt St So ‘The initial basic feasible solution is Sy = 20, S: = 24, Ss BCA, Thitd Semester TP Solved Series (Operations Research) RTMNU Since all z, ~ 6, 2 0, the solution 1s optimal and is given by Max Z.= 132, x, Ques 9)_Use Simplex method to solve the following 3X1 + 5X2 +4Xy Subject to the Constraints 2X) + 3X, <8 2X, 45X55 10 3X) + 2K) + 4Ms < 1S Where Xj, Xa Xi20 Ans: The problem can now be expressed as follows Max Z=3X; + 5X2 + 4X34 0S) +08: + 05s. Subject to constraints 2X, +3Xz +S, + 0S: + 0S, =8 10 2X2 4+5X; 408+ S34 053= = 2 Oy = 32 = a5 Byte” enon baste. ‘The Inia 3X) + 2K; +4Xs +08, +0S;+ S; simplex pes 7 Non Optimal Sot Xi, Xo, Xa, St, Se $320. Simm le : Non Optimal Solution 3 CAS [Now applying the simplex method, we have following table SJol2 [its [6 lilo fo [20 ‘Simplex Table I: Non Optimal Solution a G[3]s | 4 [oolo slo} 3 | 1 | 3 [25 Jo] 1 | o [24 Bho xe fee Xs SS Slo) 7 [0 fo [7 foro [7 fo stots {2 {ato [ifofot sa es Listetete op ote Soy oo 2 [5 folifol— 5 Sete ete fet 2 fe sot as [3 {2a folopr| ase eet eee % | 0] 01 0 fololo Esl 6l-s17) G-z1 3 1s 14 fojofo 1 i T Incoming variable Incoming Variable ‘As some of z, - c 0 the current basic feasible Entering = Xz, Departing = S,, Key Element = 3 solution is not optimum. 2; ~ c) = ~ 15 is the most Ry(old/3 = Ry(old) x 1/3 Negative value and hence x, enters the basis and the Ry(old) ~ 2Ryinew) variable S; leaves the basis. Ry(old) - 2Ryinew) Simplex Table Il: Non Optimal Solu ‘Simplex Table II: Non Optimal Solut ables Tx] uy (SS [1b] bm, Gq t3ststa to folo 1 Wia= 13 BGs] x [x1 x: [xs | 5, [5,]8, [atin Ratio 0/0) ws | 3 | 326 }1|-20 Jo] 4} ovtinimum pol St ss fos {i {0 Via fofol = ‘Value) > s:[o [as [as] | [38 1 fol tans us {1 {35a fol 14 [ols] wa =4 Sito faa teat ots fos Toll asia =15 [=7 | isa fo [= 75] Sohne te tee te 6 [st 2 fol o Jo — - s_[as|i25_fol-s Jo | Incoming Variable =1 16 | 123 ols Jo 7 Entering.= Xs, Departing = S2, Key Element = 5 * Incoming variable R,(new) = Ri(oldy/5 = Rxold) x 1/5 Since % - €, < 0, the solution is not optimal and Ri Ri(old) therefore, x2 enters’ the basis and the basic variable 8; Ry Ry(old) ~ 4Ra(new) leaves the basis, ‘Simplex Table I; Non Optimal Solution a | 3 [slat 0 Tole Solin setae BIC Xn_| XTX 5, ‘S| Min Rao xa] xs xs [Si [82 [Ss] bs s[w3 [23 Tifol aso fol 4 rary 2 [ans |—4n3 Tol 1 [ans stop — ‘si 0 | sors [aiasy/ 0 0 -2ns ast worm 2 ols Z,_| 4s |S]4l ims] ao) ‘a[ces Tah ae lq= Lins lobo -i7is—wsto a T 6|9 ofolo Incoming Variable Si} 93 |3{3]0 Entering = Xi, Departing = Ss, Key Element = 41/15 ofis 3 [fo Ry(new) = Ry(old)/41/15 = Ry(old) x 15/41 Ry(new) = R\(old) = 2/3Ryinew) Ri(new) = R(old) + 4/15Ry(new)‘Simplex Methods and Duality (Unit 2) Simplex Table IV: Optimal Solution c |afs{a] oo | o | o [B\ cs] Xw |xi)xlx) ss | ss | ss m| § | soi | ol i [o[ isai | war |-101 x) 4 | 641 | 0 1] oar | sar [avai 0 Xi} 3 | 8941 | 1) 0] 0 | 241 | 12781 15/41 | s ° % [3 GZ] 0 4 | asa | 24a | rat 0 | 45741 | ~2a/a1 | 11/41 Since all C; - 2; $0, hence optimum solution is arrived at with value of variables X, = 89/41, Xz = 50/41, X Maximise Z = 765/41 2/41 Ques 10) Solve the following LP problem using simplex method. (2016 [05)) Maximise 2 = 4x + 3x, Subject to constraint 2x1 + x2 S 1000, x; + x $800, x; $400, x3 $700, x1,%2 20 Ans: The problem is converted to canonical form by adding slack, surplus and artificial variables as appropriate: 1) As the constraint 1 is of type "s’ we should add slack variable S 2) As the constraint 2 is of type “S’ we should add slack variable S2 3) As the constraint 3 is of type *<’ we should add slack variable Ss 4) As the constraint 4 is of type “S' we should add slack variable S« ‘After introducing slack variables Max Z = 4x) + 3x2 + 0S: + 0S2 + 053 + OSs 1000, x1 + 2+ S2= 800 100 and Subject to 2x14 x2+S) Xi +S; = 400, 2+ $= X1, a, Si, Sz, Ss, $420 ‘Table 2.3: Non-Optimal Solution pa Positive maximum C,-~ Z, is 4 and its column index is 1. So, the entering variable is x Minimum ratio is 400 and its row index is 3. So, the leaving basis variable is S, :. The pivot element is 1. Entering = x,, Departing = S3, Key Element Rs(new) = Re(old) R,(new) = Ry(old) ~ 2Ry(new) R,fnew) = R,fold) - R(new) Ri(new) =Re(old) Table 2.4: Ne G 1a Co] Xe eblele @ aks Z= 1600) t IG [o JE i {| erect Positive maximum C, ~ Z, is 3 and its column index is 2. So, the entering variable is x-. Minimum ratio is 200 and its row index is 1. So, the leaving basis variable is S,, . The pivot element is 1. Entering = x2, Departing = S1, Key Element Ry(new) = Ri(old) Rx(new) = Ra(old) - Ri(new) Ry(new) = Ry(old) Ra(new) = Ry(old) — Ry(new) c [4]3[olofolo B [cy] Xm | x1 x2|Ss)8s]5s]5« zo] | % [0 [ololololo e-z[at]>ofofolo ‘Table 2.5: Non-Optimal Solution G [s3]e[ololo [| Mie Ratio | B lal x amsisisisy x _|3| 200 [ola|ijo|-2iopaz BCA, Thint Semester TP Solved Series (Operations Research-I) RTMNU Positive maximum C,~ Z, is 2 and its column index is 5 So, the entering variable is Ss ‘Table 2.6: Optims [fo a Minimum ratio is 200 and its row index is 2. So, the ° leaving bisis Variable is Sy, a 1 100.0 is}e[ofaia} c. The pivot element is | Bee le Enteri Departing Zz zaps Rxew) = Rol) C c=zZolo Ry(new) = Ri (old) + 2Ry(new) Since all C,-Z, <0 Rynew Rynew) = Ryold) ~ 2Rxnew) 200, x; = 600 Max 2600 ‘Ques 11) Solve the following LPP using simplex method: Maximise Z = 3x, + 2x; subject to x; +x; $18 In tm S28 xi +2x: $20 with x20, Ans: After introducing slack variables, we have following equations: Max Z = 3x; + 2x: + 05; +08; + 0S; Subject to Rit +S)=15 28 20 2x +E +S: xit 204 Ss and x), xs $), 53. S320 After applying the simplex method, we got the following table: Table 1 (Te 73 72 Jo Jo To B X ]x |} si/s:| sy 3) 3 [i filifofo Ss [ol 2 1[o]1]o eS) Z=0 ofolofofo Z-G|3[=2[ofofo T Incoming Variable Negative minimum Z, - Cis -3 and its column index is 1. So, the entering variable is x}. Minimum ratio is 14 and its row index is 2, So, the leaving basis variable is Ss. The pivot element is 2, Entering Variable = xy, Depaning Variable = S;, Key Element = 2 Ro (New) = Ry (Old) + 2 Ry (New) = R, (Old) - R; (New) Rs (New) = Ry (Old) ~ R; (New) Hence, optimal solution is arrived with value of variables (2018/05),Simplex Methods and Duality (Unit 2) px Table 2 | cg 3 2 o 0 o Min Ratio Bo |e} Xe fafa |S] Ss |S) Xe * So }o) ru fofryfay a fof a (3) 2 z 2 x 3 14 1 Lp Ol d [0] te 2 2 T 2 S ;O} 6 Joy 3 Jol ilit 6 = -> za4 2 2 3 2 Za 4/3] 3 Jo] 3 fo 2 2 Z-G\O} 1jol3 Jo 2 2 7 Incoming Variable Negative minimum 2;—C; is + and its column index is 2. So, the entering variable is x: Minimum ratio is 2 and its row index is 1. So, the leaving basis variable is Sy. ‘The pivot element is 3 . 1 Entering Variable = x3, Departing Variable = S,, Key Element = —_ Ry (New) = R; (Old) x 2 R New) = Rs (OND ~ 5 Ri (New) Ry (New) = R; (Old) - 2 R, (New) Table 3 G 3 2 o 0 0 B Go Xe x % Si S: Ss X 2 2 0 1 2 “1 0 x 3 13 1 0 =I 1 o | Ss 0 3 o 0 3 1 1 Za} a 3 2 1 1 0 Z-G 0 0 1 1 o | Since ali Z,~ C20 Hence, optimal solution is arrived wiv value of variables as: and Max Z = 43pat BCA, Third Semester TP Solved Series (Operations Research 1) RTMNU Ques 12) Discuss the simplex method for minimisation case? Or Draw the flowchart of simplex algorithm for minimisation case. ‘Ans: Simplex Algorithm (Minimisation Case) Minimisation case is converted into Maximization case by Min Z are followed as they are similar, Constraints may have 2 and = signs. Max (— Z) and all other steps of maximisation case ‘After ensuring that all b, 20, they are considered in the problem. Anificial variable having no physical meaning have to be ‘considered because the basis matrix cannot be obtained as an identity matrix inthe starting simplex table This antficial variable technique is used so that the starting basic feasible solution is achieved and the simplex procedure as usual is adopted tll the optimal solution is obtained (figure 2.1), ut the LPP in standard form >| Find the init basie feasible solution * Construct the starting simplex table t Update the new simplex table so obtained Compute ‘Select the row ‘corresponding to min. ratio to find the vector leaving Find the vector X, entering the basis sothat A, = min the basis T ‘Solution can be improved. ‘Compute the ratio (R=b/a) Corresponding toany 4,< 0, are all the elements of entering vector, Xi<0? Figure 2.1: Flowchart for Simplex Method Ques 13) Use simplex method to solve the LPP. Minimise Z =x, 3x; + 2x5 Subject to 3x, -x2+2%)$7 ~ 2x, + 4x $12 = 4x, + 3x2 + 8x) S10 XX) G20 ‘Ans: Since the given objective function is of minimisation we convert it into maximization using Min Max Zax) +3x2- 2 Subject to Constraints 3x; — x2 + 2k) $7; = 2x, + 4x2 5 125 ~ 4x) + 3x2 + 8x5 10 6, = CoX,~6 and Find key element and ‘Solution \ cobain the new solution by No gf under testis ‘matrix wansformation Ieany 4,<07 ‘optimal Stop f Yes ~ Solution is| ‘unbounded Stop Max (-Z),Simplex Methods and Duality (Unit 2) bas We rewrite the inequality of the constraints into an equation by adding slack variables S;, S2, Sy and the standard form of LPP becomes, Max Subject to Constraints Bx m+ 2 +8) ~ 2x) + 4% + Sp = 12; = 4x; + 3x4 B+ Ss = 10 Xty Xay Xa, 8), $, Sy 20 Z=—Xy + 3x2~ 2x5 +08; + 08; + 05) + The initial basic feasible solution is given by $, = 7, S2= 12, Ss = 10. (x1 =%2= X= 0) Simplex Table I: Non Optimal Solution Basis xu & » [fs] ® |S | & bf = s | 0 3 1 2 1 0 o | 7 al s | o [2 4 o fel: fole 12/4 = 3(Minimum Value) —> | s [o | -4 3 « [of o [1 [0 103 =3.33 « = 3 oo fo = x 0 0 o [olelfo ae) 1 3 2 [ol[elfo = F Tncoming variable Since 22 —€2 <0 the solution is not optimum. ‘Simplex Table 1: Non Optimal Solution _ Basis xu | x Si Si Ss |b bay | s fo] #2 0] 2 i 1 0] 10 | 4@sfinimum Value) » [3 | 2 1 [0 fo m7 0 [3 [30 s [o| -s |[o|s |o | -# [1 | 1 |-s2 & =I 3 | 2 [0 ° ° 7 % =az_ [3 | o | 0 a © _ T Incoming Variable Since 2; ~¢; <0, the solution is not the variable S; to leave the basis. ‘optimum. Improve the solution by allowing ‘Simplex Table IIL: Optimal Solution the variable x; to enter into the basis and { Basis % % x Si S Ss » Loa 2 1 o 45 5 wo 0 4 % 3 0 1 215 us a0 0 3 ts ° ° 0 10 1 =e 1 i 2 a 3 25 us a5 0 Ee of as is a 0 Since all z, ~¢, 2 0, the solution is optimum. = The optimal solution is given by Max Z.= 11 xy 24,02 = 5,450bas Ques 14) What is the artificial variables tech solving linear programming problem (LPP)? Or Write short notes on the Big-M and Two-phase simplex method. for Or Explain in detail two-phase method to solve the Linear Programming having artificial variables. (2017[05)) Or Write the steps of Big-M method to solve LPP. (2018(03)) Ans: Artificial Variables Technique in Simplex Method For solving the LPP there are two methods as follows: 1) Big-M Method/Primal-Penalty Method: The Big- M Method is a method of removing artificial variables from the basis. Anificial variables holding no physical meaning are assigned with variables which are undesirable for the objective function. To minimise the objective function Z, a huge positive price, also called penalty is assigned to each artificial variable. To maximise the objective function Z, a huge negative price, again a penalty is assigned to these artificial variables. For a maximisation problem the penalty willbe designated by - M and for a minimisation problem +M, where M > 0. Steps of Big-M Method ‘An LPP can: be solved by the Big-M Method through the following steps: Stepl: LPP is expressed in a standard form using surplus and artificial variables. The co-efficient of the surplus variable is assigned with a zero and the objective function of the artificial variable is assigned with +M, a huge positive number for Minimisation cases and ~ M for Maximisation case. Step 2: Zero value to the original variables is assigned to obtain the initial basic feasible solution. Step 3: Calculate and then examine the values of z in last row of the simplex table. )) For optimal solution to the current basic feasible solution, all 2 620. ii) We can say that the problem has an unbounded optimal solution if for a column, k, 24 ~ cy has the most negative value and that all other entries in this column are positive. iii) In a condition where one or more % - 6 < 0, then the-basis with the largest negative % - c, value will use a variable. That is, Min (7,—6:2,-¢,< 0) n-G The column to be entered is known as key or pivot column. Step 4: Simplex method of the maximisation case determines the key row and key clement, Follow the same method for Big-M method. BCA, Thint Semester TP Solved Series (Operations Research 1) KTMNL ation, convert by applying Min (7) = nplex Method: Linear programming problems with artificial vanubles can also be solved by another method called the two-phase method. This method is an alternative to the Big-M method as the name suggests, the method divides the procedure into two phases: i) The first phase of this method is the process of minimising the sum of the artificial vanables to get a basic feasible solution 10 the problem considering the given constraints, This iy known as auxiliary ii) The second phase starts with the basic feasible solution obtained at the end of Phase 1 In this phase, the original objective function is optimised Steps of Two Phase Simplex Method The iterative procedure of the wo phase simplex ‘method algorithm is given below Step 1: Put the linear programming problem into its standard form, Then, check for starting basic feasible solution of LPP, i) If the starting basic feasible solution already exists then go to Phase 2 directly ii) If the basic feasible solution does not exist then 20 t0 next step, Le., Phase I Phase I Step 2: This step involves variables which are required for the basic feasible solution, Artificial variables are added to the left side of each equation to complete the equation that lacks the required staring basic variables. An auxiliary objective function is constructed to minimise the sum of all autificial variables to get a basic feasible solution to the problem ‘The new objective is, Minimise Ay Maximise A Where A, (i= 1, 2... m) are non-negative artificial variabl Step 3: The new auxiliary LPP is then solved using simplex algorithm method. At the least interaction following three possible cases may be seen: i) When Max Z* < 0 and at least one positive Value artificial variable is present in the basis then the original L.P.P. will not have any feasible solution, ii) When Max Z* = 0 and at least one zero value anificial variable is present in the basis, then the original L.P.P. will have feasible solution. In such a ease, proceed to Phase 2 to get the basic feasible solution. We may proceed directly (0 Phase 2 in order to get basic feasible solution or else proceed t Phase 2 after cial basic variableSimplex Methods and Duality (Unit2) bar iil) When Max Z* = 0 and there is no artificial Phase I Variable inthe basis, then a basic feasible Step 4: This phase starts with the optimum basic solution 10 the original L.P:P_ has been found. feasible solution of Phase 1. The variables of the Go to Phase 2. objective function are assigned actual coefficient and a zero value to the artificial, variables that appear at zero value. (Ques 15) What is Difference between Two Phase and Iig-M Method? Ans: Difference between Two Phase and Big-M Method between two-phase and Big-M n ji Two Phase Method Big-M Method — Tes easy 10 obuain ital basic feasible solaion because @ | Tew diicalt to bain initial basic feasible solution Because a to basic variables, ero co-eficint is not attached to busi variables. - Tis easy and less time consuming. _ Ts tedious hecause variables (penalties) are involved. - This method does not mvolve any assumpiions at the original | This method uses assumptions for constraints. system of constraints, Ques 16) Solve by Big M Method. Minimise Z = 60x, + 80x, Subject to. x22 200 xis 40e X1 + x2 = S00 x22 0 ‘Ans: Converting the problem into the standard form by adding slack, surplus and artificial variables in the set of constraints and assigning appropriate cost to these variables in the objective function, the problem can be re-writen as follows: Min, Z= 60x, + 80x3 + 05; + 0S; + MA; + MA: Subject to. Ox; #12—$) +02 + Ar + 0A. X, + O.X; 4 0.5) 48; 40.A) + 0.8; = 400 x1 #24 0.81 + 0.52 + OAL +A: Whereas x1, 52, Siv SA AZO From the above system of equations, the al basic Feasible solution can be displayed in the following simplex tables: “Simplex Table I: Non Optimal Solution © [> [__ Contrbation per Unit of so [ol o[M[M T |_| Basie Variable | Solution Vi sr Si |S: | A | Ar] Mie. Ratio M Ay 200 0 tot t}ol 205 0 - 00 1 o}1pope, = M 300) ofop opr, 309] wom | =w[o[M[M = 9 Mw 2m[ Mol olo T Incoming variable Simplex Tuble H: Non Optimal Solution ‘Contribution per Unit oo [w| 0 [o[M Basie Variable | Solution Value |x, | x: | S:_| S| As | Min, Ratio X 200 o [tp fofof = 5: 00 a “AS __ 30 a a 30> 2 “Ta000 + 300m | M80 = 80 | 0 | ME c-4 lw po [so-m[ofo T coming Variable Simplex Table IL: Optimal Solution © [> [| ___ Contribution per Unit o[w[elo T{_ ariable | Solution Value |u| | Si_| Ss 0, [200 oft ° © 100 ote H we 300 se s40m | 40 | 90 | -20 | 0 { 7 co oo |» ToD38 BCA, Third Semester TP Solved Series (Operations Research-l) RTMNU As al the values in the index row C, ~Z, are either positive or zero further improvement in the objective function is not possible, Hence optimal solution is x; = 300, x; = 200, and Ques 17) Solve the foliowing simplex problem using Big-M method: Minimise C= 2x, + 4x; Subject to constraints 2x; +132 14 xi+3m 218 Xtm212 Xm 20 Ans: Since the given objective function is of minimization we convert it into maximization using Min C = Max C=~2x) - 4x; After introducing the necessary slack, surplus, and artificial variables, the augmented problem is given here: Maximize C=~ 2x) = 4x +05; +05; ~MA\ -M. Subject to 2x; #248) = 14 Xi+3x2-S2+A\ Xitm + Ars 12 Xi, Xr» Si, Sp, Ar, A220 BG [_% x, S18, |S: ‘Min Ratio ] AL =M 4 2 to [0 if a: | =M 1 H o[-1fo $= As | = 2 i oo fr re ¢ = oot ¢ =iMi M[M | M. GG [ama = [= | =M i Tncoming Variable Entering = X;, Departing = A;, Key Elemem = 3 Ry(new) = R,(oldy/3 = Ry(old) x 1/3 Ri(new) = R,(old) ~ Ry(new) Ry(new) = Ry(old) ~ Rivnew) Table 2.8: Non Optimal Solution . BIG] % Tels [ os Ts Tar] a; Ta, [Min Ratio A[-M] 8 oy fof ao pt is et as [ats 5 Ail-M{ 6 23 oo ig ro is 9 q 2 ato 0 0 _[-M M. -M C= a | | MaMa | =| 2M aa [MT G=G | 7aM-23 [0 [-M | 33M-43 T-Mt0 [-samM+ano | 7 Incoming Variable X1, Departing = A, Key Element = 5/3 (old)/513 = R(old) x 3/5 (old) ~ 1/3R(new) (ld) ~ 2/3R,(new) on Solution BI Gy Xs [XTX & Sy Sy AL Ay A, [Min Ratio Pal =2 [aus To [38 vs__[o] 35 =us— fot = Palas To 15 =ys— To [=i 2s to| = Al-M] 145 Toto] as) s__[-r] 215 a G [ato oto] MM C_[-2 [=a] -215M + 975-15 + 68 | M [275M 25 | 13M 65 [Mt =C[0 To [asm — 25 [ism — 65 [-M[=7/5M = 2/5] 15M + 6/5] 0 I ‘: Incoming Variable Entering = S,, Departing = Ay, Key Element Ry(new) = Ry(oldy/2/5 (old) 5/2Simplex Methods and Duality (Unit 2) pa» R\(new) = R (old) + 3/5R,(new) Runew) = Rx(old) ~ 1/5R,(new) Table 2.10: Optimal Solution 1G) ™ [x puls1 s 1S ‘Ay [ Ay _| Mim Rati | {| -2) 9 1}o]o] 12 |-32 <2 32 x=] 3 Pe po tin tia ia} 12 _ a =| 32 | [ [e-fatatot oto “MM f | G [2fatoviyt =I 1 I ¢-cToToTo] -i =) [-M [-M+1] -M+1 X Maximise C Minimise C Ques 18) An advertising agency wishes to reach two types An initial basic feasible solution is obtained by setting x) = of audiences — Customers with annual income greater x = S; = 0. Thus, S, = 20, A; = 3, Sy = 5 and Max than 15,000 (Larget audience A) and customers with 3M. annual income less than 715,000 (target audience B). The total advertising budget is €2,00,000. One programme of This initial solution is shown in simplex table 2.11 TY. advertising cost 250,000, one programme on radio advertising ®20,000. For contract reasons, atleast three ‘Table 2.11: Initial Solution _ programmes ought to be on T.V. and the number of radio Basis [x _[xq]541S41Ss] Ax [bs [bya rogrammes must be limited to five. Surveys indicate that Sol 5 [2[1[0[ 0] 0 |20] 205=4 a single T.V. programme reaches 4,50,000 customers in Ma folotifol1|3}an=3 >] target audience A and 50,000 in target audience B. Ome sfol 0 fifofojifo{st = radio programme reaches 20,000 in target audience A | 5 [1 fololol-Mi and 80,000 in target audience B. Determine the media mix z_| -M_[0[0[M|0|-M ‘to minimise the total reach. Use simplex method to solve =a [-M—3|-1[0/M|o] 0 ‘the problem. T Incoming variable ‘Ans: Let x; and x2 be the number of programs in TV. and respectively. The LP problem can be formulated a8 As 2,~ ¢, value in x; column of table 2.11 is the largest follows: negative value, enter variable x, to replace basic variable Total TV Customers = Customers in target audience A and A, into the basis. For this apply following row operations: Customers in tatget audience B R,(new) ~> Rx(old) /I(key element), "= 450000 + 50000 = 500000 atomers in target audience A R(te¥) > Ri(old) ~SRe(new) Total Radio Customers = and Customers in target audience B Hence the new solution is shown in table 2.12: 20000 + 80000 = 100000 ‘Table 2.12: Improved Solution Maximise (total reach) z = 500000 x; + 100000 x2 = Sx, + x2 Basis |x | x |S, | S, |S) | bi | byay Subject to the constraints, sfofof2 tits ots} seis 30,000x; + 20,000x2 < 2, 00,000 a é : o a of3{ - (Or 5x; + 2x2 $20 (Advertisement budget) toto} rote yi st x23 (Advertisement on T.V.) stoperes 2 aj ns5 (Advertisement on Radio) ete terete and xy, 220 + After introducing slack/surplus and/or artificial variables Incoming variable in the inequalities of the constraints, the LP problem in ee ‘The solution shown in table 2.12 is not optimal as 24 ~ cy Maximise z = 5xy + x2 + 0S; + 0S; +0S3—MAy value in S; column is the largest negative. Subject to the constraints, 5x +2x2+ Si x)-S:+ Ai m+ Si=5 and x, X35 51, Sos Ss, A120 Ri (new) — Ryold) + Ry(new) ‘Thus, enter variable S2 to replace basic variable Sj into the basis. For this apply the following row operations: Ry (new) ~ R, (old) /5{key element), 20po Hence the new solution is shown in table 2.13: ‘Table 24 ‘Optimal Solution Since all 2, ~ cj 2 0 in table 2.13, hence, the optimal solution is: x, = 4 programs in T.Y. and x; = 0 in radio with Max (total audience) 2 = 20, 00, 000, ‘Ques 19) Use two-phase simplex method to solve Maximise 7 = 5x, ~ 4x, + 3x, (2017105)) Subject to 2x; + x2- 6x; = 20 6x) 45x; + 1045576 Bxi- 3x +6550 XX) ZO ‘Ans: Introducing slack variables S,, S$, 2 0 and an artificial variable A, 2 0 in the constraints of the given LPP, the problem is ref BOA. Shad Semester TP Sokved Series (ppratina tosest’%\, OM Since wy ~ can oxtisnan wlation te the massliaey LAE has not been obtained Stenplex Table IN: Mew Otienal Sehatioon [al xe [As [1% sfololt [asent zn a7 | nlolifol aa | am [ois ‘ne [o[0 Since all % ~ 6, 2.0, an optimnurn waitin to the aailiary LPP has been obtained, Also Max Z* = 0) with nes atificia vvatiable in the basis, We go te phase U Phase I: Consider the, final sirnplex table of phase | Comsider the actual cost associated with the original variables. Delete the artificial variable Ay column fron th table as itis eliminated in phase I. Initial basic feasible solution is given by A, = 20, $, = 76 and S, = 50. Phase I: Assigning a cost-1 to the artificial variable A, and Cost 0 to other variables, the objective function of the auxiliary LPP is, ie inte ded ipl: Ota Bt _[-a)o] sa] of -17 far . S lolololasali] an lan x Slilol Jol ina ts, SStertitan Maximise Z*=0 xy +0 x24 Oxy +05, +08; —1A, Subject to 2x4 4X2 -6%y +A) =20 6 xy +5 x2 + 10 1548, = 76 Bx; -3.x2+ 6554S) =50 X12 Xa Si, Sap AL ZO. Simplex Table f: Non Optimal Solution Baris [x1 [, [9 [Ay]5i[82[b,[ 6/4, ‘ail -1[ 211 [-6] 4 [0] 0 [20] 202 10 [s [ole 1 [076] 7616 = 12.66 s,[ ols | 1 [50] 5018 = 6.25 | ¢ fo ofo 2 ofo Ff 6fofolo icing ab Since 2) ~¢) and z ~ cy is negative therefore this is not Optimum solution for mazimization case. Simplex Table I: Nom Optimal Solution vs [Afsil $ [ o [ bd 3 1 D a -15/2] -14] 1572] 3071| “314/772 1S4729) “Ws [25/4 0 s ele]— [E Incoming variable Since all 7, ~ 6, 2 0-an optirmurn basic feasible woluticm tat been reached. Hence, an optimutn feasible solution to the Biven LPP is 2) = 55/7, x2 = 30/7, x; = 0 and Max Z = 15577 Ques 20) Une two phase method to: Minimise Z = x) + x2 Subject to 2x, +24 +727 x, %20 ‘Ans: The problem is a problen of minimisation. So, for our convenience, we will convert this problem of minimisation to maximisation by taking the objective function as: 2! = —2 Maximum 2 = ~2=—1)~ x2 Purthes, all the b, (4, 7) are positive. Introducing the surplus variables x; and x4 the constraints become: 2x) 42-4 Ky 4 Tg -%4=7 ‘Here, we cannot get the starting basic feasible solution. we introduce the artificial variables (positive variables) x, and t, ‘Thus the above equations can be written as: 2x tp Ky tae eh Ky + xg — at = 7 Cheatly x5 < x3, XoSunglex Mettoatn wont Dati (ini 2) Phase sautables Faking 4) = OS Ox OL we get Wa Age 7. Hhiy phave consists of the removal of artificial 0 Let, Ay and Ay denote the artificial column vectors comtesponsting 0 artificial variables xg and xy eespectively. We construet the First table as follows: " able 2.14 Now, we first remove one by one the each of artificial column yeetor Ay aud Ay from the basis matrix. We ean remove A> and introduce Y; in its place in the s matrix, For this, divide the second row by 7 and then subtract it from the first row. Thus we get the next transformed table: ‘Table 2.15 a] | om off 7 1 U Whenever the artificial x, becomes zero (non-basic), we forget about it and never consider the corresponding vector Ap again for re-entry into the basis matrix, Now, we proceed to remove Ay and introduce Y; in its place in the basis matrix. For this, we multiply first row by and subtract + times of this mew from the second row, Thus we get the following table: ‘Table 2.16 Bl mls fs |» [om TA 2I Tea [2 pli |o || 3 fi w | o | 1 | 4] ot | Bp || | —) tai fio fo fo.) 0 » Lp Ls Since, art ariable xs become zero 80 we forget about A, and not consider it again, Thus we get the following solution in phase I: 21 us 10 Ciara bat which is the basic feusible solution with wl proceed to get the optimal solut ch we can) 1 by simplex method, Phase follows: Now, construct the starting simplex table as [Minima Rai] 10 BAG Minimum Z = —maximum 2’ B (Ques 21) What are the special cases in simplex 1 Or ‘Write short notes on the following: 1) Infeasibility 2) Degeneracy 3) Unboundedness 4) Maltiple Optimal Solutions Ans: Special Cases in Simplex Method While using simplex’ method for solving LPP, one may encounter many special situations which are summarised below: 1) Infeasibility: Now it is very cleur that @ solution is said to be feasible if it satisfies all the constraints and the non-negativity conditions of a problem, However, in real life situations, itis possible that the constraints are not consistent or there is no feasible solution to the problem, In such cases, a non-feasible solution is obtained also known as infeasible solution and the situation ts called unfeasibility. Also, a non-feasible solution occurs if fan artificial variable appears in the basis of the solution for being optimal 2) Degeneracy: Problem of degeneracy is caused when an LPP ig solved using simplex method and a. situation occurs in which there is a tie between two orDaz more basic variables for “leaving the bases’. This means that the Values of one of more basic Variables in the solution column is equal to. zero or the num ratio to identify’ the basic variable to “leave she Kuss fs not unique. Degeneracy also occurs during iterations in the simplex method, Therefore, when there is a tie in the ninimum ratios for Teaving the basis, then the selection is made arbitrary, However, one can reduce the number of iterations for amiving at an optimal solution by adopting the following rules: i) Detect degeneracy and divide the coefficients of the slack variables by the corresponding positive numbers of the key column in the row starting from left to night Comparing from left to right, the row that contains the smallest ratio becomes the key row. id In a situation where there is a tie between a slack and an artificial variable to leave the basis, the artificial variable should be given preference for leaving the basis. Also, in such cases. degeneracy need not be resolved Unboundedness: Unbounded solution occurs. when there is no constraint on the solution and the decision variables can be increased infinitely under the same restricted conditions. Such unbounded solution occurs in niaximisation 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 in minimum ratio column are either infinite or negative. Multiple Optimal Solutions: Sometimes an LPP may arrive at an optimum solution but not a unique one. This means that more than onc optimum solution exists. In such cases, the following are the characteristics of the final table: i) An optimum solution, ii) Forall basic variables, (Z,—C,) values are zero. 1) For non-basic variables, (Z,—C) values may or ‘may not be 2ero. For atleast one non-basic variable, (Z,~ C;) value will be zero, More than one optimum solution is possible. For all possible solutions value of objective function Z remains the same. ii) Basis in final tableau contains basic variable X, vic. X). Xa... and non-basic variables S, viz. Sj S> ‘Ques 22) Solve the following LP problem using simplex ‘method. (2016 (05) Maximise z= 6x, + 4x, Subject to constraint 2x, 43x50 3x +2x, S24 xytem23 xy 20 3) 4 vy) BCA, Thind Semester TP Solved Series (Operations Research-t) RTMNU ‘Ans: Introducing the non-negative slack variables. Sy, surplus variable, S., an artificial variable A\, the LPP becomes Maximize z= 6x, + 4x2 + 0S, + 0S:~MAy Subject 10 2K; + 3X; 48) +082 +055 3x) + 28: +08; + So +05: x1 +X2 + 08) + 0S,-S, 4A Xie Xa, Sy Sy Se AY 20 30 An obvious IBFS(Initial basic Feasible Solution) is S, = 30,8)= 24, A= 3. Initial iteration: ‘Table 2.18: Non-Optimal Solution 6 | 4 [ololo | x} ]Si [$2] Se] A: [ Min, Ratio 2 [3 [i fofoto 3 [2 Joi fofo | 2-8 1] | t fofof-fa Gee Exeot ato To TMT 3 Ge 3 8: 2-6, ‘Second iteration: Table 220: Opti clels 7 Cu | Ye | Xe | xr Sy o{s [ufo 0 ots [se 1 eta pa[t ° zu | [efo ° Since all z, ~ ¢; 2 0, the current basic feasible solution is optimal. The optimal solution is x, = 8, xj= 0 with ‘maximum of 2 = 48. Alternative Solution Since z;~ ¢, = 0, for the corresponding non-basic variable Xa, alternate solution exist, Therefore the solution as obtained above is not unique. We can bring x; in the place of S, of S), Introduce x2 into the basis in place of S, the new optimum simplex table is obtained as follows: - S. Co) Ys | Xe [|e ss [S + [a [auspoli pas toasto 0 [si fssfofol ws] ws ]a x, | 2s [1 [0 | as | as fo = 1 lefefo [2 fo‘Simplex Methods and Duality (Unit 2) Since z, ~ ¢, 2 0, the current basic feasible solution is optimal. Another optimal solution is x, = 12/5, x2 = 42/5, with maximum of 2 = 48, ‘Thus if two basic (optimum) feasible solutions are known, an infinite number of non-basic feasible solutions can be oblained by taking a weighted average of the (wo solutions. Thus if x; = [8, 0. 14, 0, 5}, xp = [12/5, 42/5, 0, 0, 39/5] are two BFS, then a NBFS bas We have resolved degeneracy as below: S [S18 Sp fon ar ES, Tos 1s* [o/s] beast non-negative (+ve value) ‘Simplex Table 1: Non Optimal Solution BA + (12/5)(1—2)} {| Contribution per] 5] 3 0 |o 42/5(1-2) > unit xtahe, +(1-) 14a ~ » +] aie [soutiol x | a2 |si{ s {ss]atin. 0 Variables| on Ratio. | 5k +39/5(1-a) Values | 8442 lo s, | 0 fo] 35 |1]-wslolo | ¥ 5 42/5-42/5) s| x | 2]. | % ws ols | =| lo| sy | & fo -215|1 [ans | o | 39/544 L 0 3 2 10 of] 1 fof - fo ‘Where 0 < A < 1. We can verify the solution x will always give the same value of z, for all 0 (Maximum Value) Incoming 9D4 BCA, Third Semester TP Solved Series (Operations Reseasch-1) RTMNU Ques 24) Solve the following problem using simplex method: Maximise Z = 5x)+ 6x; + x: Subject to the constraints ‘9x, +3x,- 2x Subject 2y, + ay. 240 By, 4 ys 235 ynye20Simplex Methods and Duality (Unit 2) par Hence, we find that 1) The primal problem is of the maximization type while the dual is of the minimisation type 2) The constraint values 60 and 96 of the primal have become the co-efficient of the dual variables y, and y2 in the objective function of the dual in that order, while the coefficients of the variables in the objective function of the primal have become the constraint values in the dual. 3). The first column of the coefficients in the constraints in the primal has become the first row in the constraints in the dual, and the second column has similarly become the second row, 4) The direction of inequalities in the dual is the reverse of that in the primal. Thus, while the inequalities in the primal are of the type S, they are of 2 type in the dual. ean represent them in matix form as given in igure 2.2: wien gy ee By 4 e388 €9C-8) joo st Figure 23: Primal and Dual Formulation Example 1; A company has three production facilities S,, $2 and Ss with production capacity of 7, 9 and 18 units (in 100s) per week of a product, respectively. These units are to be shipped to four warchouses Dy, D2, Ds and D, with requirement Of 5, 6, 7, and 14 units (in 100s) per week, respectively. The transportation costs (in rupees) per unit between factories to ‘warehouses are given in the table below: Di De Capacity 9 10 7 S. S S 40 20 1 Demand 5 is s ‘Example 2: The concept of duality in transportation problem is applied inthe following manner. Reproducing transportation data for ready reference in table 12. In table 12, there are m =3 rows and n = 4 columns. Let u, tz and us be dual variables corresponding to each ofthe supply constrain in that order. 30 70 30 0. 3 a 3 Simitary, v,, vs: Vs and vs be dual variables corresponding to each of demand constraint in that order. The dual problem then becomes oe Ds > Ds De & 9 w 30 10 Sh 30 ‘0 oO S ® 70 20. Demand : i mH 70. o. Maximize Z = (Tu, + 9u2 +18us) + (5¥1 + 8¥2 + 7¥3 + Lava) Subject to the constraints uy ¥9 $30, uy + v5 550, ult ves 10, w+ v2 530, ur vs $40, Ure S60, uy+v2S8 ust ¥5 $70, Us+ eS 20, and uy, v, unrestricted in sign for all i andDg BCA, Third Semester TP Solved Series (Operations Research-1) RTMINU For interpreting the data let us consider the constraint u, + v; S19 or vj < 19 —u,. This represents the delivered. Value ofthe commodity at destination D, which shouldbe less than or equal to the unit cost of transportation from S, to D, ‘minus the per unit value of commodity at D1. A Similar interpretation can also be given for other constraints. Now, the optimal values of dual variables can be obtained either by solving this LP problem or by reading values of these Variables from the vansportation table that constrain the optimal solution, : ‘One may verity that the total transportation cost at optimal solution obtained by the MODI method would be the same as obtained by putting values of v,'s and v's from optional transportation table in the dual objective function: i 4 Maximize Z= Yau, +Db,v, 7 i ‘Ques 28) What is the relationship between primal and dual? Ans: Relationship between Primal and Dual | ‘The relation between the standard and its corresponding dual is given by following table 2.20: Table2.20 - Primal a 1)_ Objective iso maximise 1)_ Objective is to minimise 2) Variable x, 2) Constraint j 3) Constraint i 3) _ Variable yi 4) Variables x, unrestricted in sign 4) Variable y, i unrestricted in sign 5) Constraint iis = type [5 Constraint jis = type 8) © Type constraints 6)_2 Type consraints | 8) _< Inequality constraints 8) 20 Variables } Mathematically it is shown below: Primal Dual Max Z = c14) + 6282 + ans + Coke Subject to the constraints BRL 2 se + ay SDs x20: Min 2° = byw + basta +n +, Subject to the constraints: yy 4 y+ + Rein 2 6) re 12cm w20 i Min Z = c1xy + C9k2 + + Coke Subject to the constraints | ami tas Ay 2B, x20: ay Max 2* = bw) # BW. # 0s + Day ‘Subject tothe constraints: y+ Baja + + yy SG, pent 1W/ unrestricted, f= 12 oom | Minimise Z=¢)x) + ¢:x2 bb Og Subject to the constraints ayn + + Bahn = by Maximise Z* = byw + baw tant Data Subject to the constraints; ay) +2 + Hg Wn SE Beh Qt 1, unrestricted, 12120 Manimive 7-= €141 4 62% Subject « the constrain: Minimise Z* = byw, + bs; Foo tan ‘Subject (0 the constraints: Ay BW + Bg j=, unrestricted, i OO en |simplex Methods and Duality (Unit 2) Ques 29) Discuss the principles of duality? ‘Ans: Principles of Duality ‘The symmetrical form is used to construct the dual from the primal or primal from the dual. The rules for the same are 1) The maximised objective function of the primal tec the minimise octet fH 2) The inequality signs are reversed in duality. This ‘means that for all maximisation primal problems with S ype constraints, there will exist @ minimisation dual problem with 2 type constraints and vice versa. However, this reversal of inequality is true for all constraints except the non- negative condhtions. 7 5) Th earn notre) comms variable in the dual and vice versa, This ‘Deans for a primal problem with m constraints and o variables, there is a dual problem with m variables and constraints 4) The right hand side constraints by, bs, .... by of the primal become the coefficient of the dual variables Yis Yivs-s Ym in the dual objective function Z* Also the coefficients ¢1, C2, .... Cy of the primal variables Xie X55 oe % iM the Objective function become the right hand side constraints in the dual. 5) The matria of coefficients of variables in primal is a transposition of the matrix of the coefficients of Variables in dual and vice versa. Ques 30) State and prove duality theorem. (2018(05)) Ans: Duality Theorem, If either the primal or the dual problem has a finite optimal solution, then the other one also possess the same, and the optimal values of the objective functions ofthe two problems are equal, Max Z, = Min Z, Proof: Let the following LP problem represent a primal problem: Maximise Z, = cx subject to, Ax Sbor Ax +Is=b)x20 where I is the identify matrix of order m and s is the slack variable. Let xp = B'b be an optimal basic feasible solution to this primal problem and ca the cost vector of the basic variables, Then according to the optimality condition, we have, ¢ - 2, $0, for any vector a, of A but not in B Therefore, g- ~ 4 (Bla) <0. for all j This is equivalent t0 cS ¢gB'A, when written in matrix notation, Dag Let y* = epB!. Then cS cpB™A becomes ¢ < y* A or ch unrestricted (redundant), The dual variables “y, and ys unrestricted” are dominated by y,>Oandy,20, Eliminating redundancy the restricted variables are y, 20and y, >0.ps0 Ques 33) Obtain the dual problem of the primal problem: Minimise Z = x; ~ 3x; ~ 2x (2017(05)) Subject to constraints: 3X1 ~ 2 + 2x3 $7 2x, ~ 4x, 2 12 4x. + 3x2 + 8x3 = 10 Xi, X12 0 and xy is unrestricted. Ans: All 2 constraints can ulightag ton eth et be converted to < type by MAX 2 = X1~3x3~ Ixy Subject to 3X, =X +245 $7 aw ~ Im +4u S12 2.Q) = 4x1 + 3x) + 8x) = 10 @) and Xi, X2, x32 0; Since 3" constraint in the L primal is equality, the corresponding dual variable ys will be unrestricted in sign, Dual is as follows: MIN Z, = 7y; ~ 12y2 + 10y; Subject to 3yi=2y2-4ys21 Yi + dy + By; 2-3 2yi+ 82-2 and y,y2 0; ys unrestricted in sign ‘Ques 34) Write the dual of the following LP problem: Minimise Z = 3x; - 2x; + 43 Subject to the constraints Ixy + 5x; + 4x3 27 6x +x 43x) 24 Tx, - 2x3 - x; $10 x) = 2kz+ 5x3 23 4x, + Tx, — 2x22 and Xj) X35 % 20 ‘Ans: Standard Primal Since the objective function of the given LP problem is of minimisation, the direction of each inequality of type has to be changed, The standard primal LP problem 30 obtained is: Minimise Z = 3x, - 2x2 + 4x3 Subject to the constraints, 3x) + 5x2 + 4x3 27 Ox +x +3 24 Tx) $2: +%52 xy - 2m + 5K 23 4x) + Tx2- 2022 Dual Primal If y,, Ya. Yo, Ye and ys are dual variables corresponding to the five primal constraints in given order, then the dual of this primal LP problem is: BCA, Third Semester TP Solved Series (Operations Research) RTMNY Maximiso-Z* = Ty, + dy ~ 1Oys + 3ys + 2ys Subject to the constraints 391 + 6y2~Tys + Yat dys $3 Sys ya 2ys—2yat TysS-2 4y) + Bya + yy + Sys ~2ys $4 and Yue ¥aYs yu ys20 Ques 35) Obtain the dual of the following LPP: (201 Maximise Z = 2x, + 5x2 + 6x os subject to Sxyr 6x;—%5 <3 =2x +m + 4x5 54 xy — 5x2 + 3x51 — 3x, - 3x2 + 7x3 56 ‘with x1, X2, X32 0. ‘Ans: Primal Max Z, = 2X1 + 5x2 + 6x) Subject to 5x1 + 0X45 $3 2x te t4us4 X1- 5243x351 ~ 3x1 —3xn + 7m $6 and x, X2, X3 20. Dual Min Z, = 3y, + dys + ys + 6ys Subject to . Syr—2ya + ys— 3ye22 by, + ¥2~ Sys 3y229 yi tye + 3ys 726 and yL, yo. Ys. Ya2 0. ‘Ques 36) Obtain the dual of the following LP problem: Minimise Z =X, + 2X2 Subject to the constraints 2X, 44X25 1600 Xi- X= 30 X.210 Where, Xi, X:20 Ans: Since the objective function of the primal LP is of minimisation, change all < type constraints to 2 type constraints multiplying the constraint on both sides by =I. Also write = type constraint equivalent 10 wo constraints of the type 2 and S. Then the given primal problem can be written as: Minimise Z,= x) + 2x2 Subject to the constraint ~2x)— 4x22 — 160 X12 30 i= x25 — or — x +> — 30 X, 210 and x), x2 20Simplex Methous an! Duality (Unit 2) best Lety1. 92. ¥9 and y4 be the dual variables corresponding Ans: Standard Primal tothe four Constraints in the given onder Changing the third constraints, the given LPP is The dual of the given primal problem can then be Minimise Z= 2x) + 3x + 4% formulated as follows: Subject to the constraints 2xy + 3X2 + SK Maximise Z, = —160yy + 3092~ 30y, + 10y4 dart Senet Subject to the constraints mai = dna 65 2-5 Bit y-yit yes |, Xj, X22 0, %5 unrestricted. 4yi-y2+ nS 2 and YuYnyuya2O Dual Primal ‘The dual of this problem is: Lety = ys ys (ya. ¥s2 0). The above dual problem then Maximise Z* = 2y, + 3y2— Sys reduces to the form: Subject to the constraints Maximise Z, =—160y, + 30y + 10ys 2yi+ Bya~ys $2 3yi+y2—4ys=3 Subject to the constraint Syit Ty2—6y) 54 ity tyes L Yin ¥220, ys unrestricted. 4y)~yS2and y,,ya2 0; y being unrestricted in sign ‘The standard form of the dual is: Maximise Z*=2y, + 3y2 —5(y} -¥3) Ques 37) Write the dual of the following LPP in standard form: ‘Subject to the constraints Minimise Z = 2x + 3x + 4x 2y, #3y2 = (yy) S2 Subject to the constraints By, + ¥2 My -¥9)=3 2xy + 3x; + 5x22 Syy + Ty2 — 6ly3 ~ YS 4 wekitonss Yodo vh20 Xi,X22 0, xs unrestricted. ‘Where, ys, being unrestricted, is equal to ¥4~ y% ‘Ques 38) What is dual simplex method? Write the steps for solving LPP using dual simplex method. Or ‘Write the algorithm for dual Jmplex method. Also draw the flowchart of dual simplex method. Ans: Dual Simplex Method ‘The simplex method is used to find the basic feasible solution to a LPP and the procedure /algorithm is said to be complete when the optimal solution is achieved. This means that when all c;~ 2) $0 for maximisation problems and c)~ 2, 2 0 for thinimisation problems is achieved, it is said that optimal solution is achieved and the procedure should be stopped. In simplex method one or more solution values (i, xy) are negative and optimality ¢~ 2, both for maximisation and minimisation is satisfied, then the optimal solution may not be feasible (because cj ~ 2; = cj ~ caBa, is completely independent of the vector b). Such cases are dealt by finding a starting basic, Le. ¢ ~ z) $0 for a maximisation problem With exception to the feasible solution that is dual feasible, ‘The best way to deal with such cases is to use a variant of the si simplex method helps to retain optimality while bringing the primal back to feasibility (i.e. iplex method called the dual-simplex method. The dual 0 for alli). Dual-Simplex Algorithm ‘The dual simplex algorithm can be applied using the following steps Step 1: Determining an Initial Solution: The LPP is converted to a standard form by adding slack, surplus or artificial Vartables to obtain an initial basic feasible solution, This solution is presented in the initial dual-simplex table. Step 2: Test Optimality of the Solution: The dual-simplex method is not required to be used in cases where all solution Values are positive (e., Xn, 2 0 for all i). This is because simplex method is sufficient to find the improved solution. Otherwise go to step 3. ep 3: Test Feasibility of the Solution: Current solution is said to be infeasible if all elements in row r and colurnn j are poutive (re. yq 20 for all j) but for which solution valve is negative (ie, xye< 0). Hence go to step 4Ban BCA, ‘Third Semester TP Solved Series (Operations Reseatch-l) RTMNU Step 4: Obtain Improved Solution i) For obtaining improved solution, select a basic variable associated with the row (called key row) that has the larges, negative solution value, ie, x5, = Min (Xp, : Xp, <0). ii) The minimum ratios only for those columris are determined that have a negat variable associated with the column is selected for entering into the basis, for which: fe element in row 1 A non-basic 98 nl y, <0} tora Yar Ya " A key element is the element (ic., ya) which is atthe intersection of key row and key column, The improved solution can bbe obtained by making ya. as | and all other element of the key colurnn zero. However, it must be noted that key elements always positive, Step 4: Revise the Soluti sible solution, Repeat steps 2 to 4 for either an optimal solution or a non- In figure 2.3 a flow chart of solution procedure of dual-simplex method is shown: “Transform the LP model into standard oem 4 ‘Setup nal simplex table ¢ > __ come and =e rales Do a 6-5 0and al (Optimm soltion mad. ‘bined . 1) Select key row with ages negative solution vals (ie, xy <0 for alli) 2) Select the key column wih Min {(6,~2)! 6 Yo <0) aio, where ya the ‘epative clement in the key 10 v 1) Determine an clement (called key element) atthe inersection of key row sod Key colon 2) _Deyeloo the nex imorove solution with the wal metho ‘Figure 23: Flow Chart of Dasl-Simplex Method ‘Ques 39) Use dual simplex method to solve the following LPP Max Z=-3n-% Subject to x +%:21 2x #31222 xm 20 Ans: Convert the given constraints into < type. Max Z=-3x-m Subjectto ~x,—xp<-1 = 2x) — 3x, 5-2 x20 Introducing slack variables 8), S2< 0, we get Max Z =~ 3x) xp + 08, + 082 Subject to =x, — x24) <—1 2x, ~ 3x2 + 825-2 X12, Sy, $220sunplex Methods and Duality (Unit 2) De: An intial basic (infeasible) solution of the modified LPP is S, =~ 1, S=~2: aeeee ‘Simple Table I: Non-Optimal Solution x % Si = =1 i 0 i Incoming variable Since all Z,~ C, > 0 and all b;< 0 the current solution is not an optimum basic feasible solution. Since bs =~ 2, the most negative, the corresponding basic variable S2 leaves the basis. Also since max (Z;— Cay’ ax<0) ‘where x, is the leaving variable max (3/ - 2, I/ = 3} =~ 1/3 = Z2 ~ Ca/am the non-basic variable x2 enters the basis. Drop $2 and introduce x2 ‘Simple Table It: Non-Optimal Solution Basis x x St S by s [8 3 0 1 ap = > a “1 23 I L ° =13 23 Zi = 23 = 0 1B = 28 G =3 -1 0 o ZG 78 0 0 13 t Incoming variable Since all Z;—C,>0 and by =~ 1/3 < O the current solution is not an optimum basic feasible solution. 1/3 the basic variable S; leaves the basis. Also since max {Z,~ Cyfas, au <0) = max ((1/3)(-V/3)...(/3(-1/3)) 1 corresponds to the non-basic variable Sp +. Drop S; and introduce S:. Basis x & S 1 =3 x 1 =i Z, a 1 Cs 3 oO =o z 1 Since all Z,C) <0 and also b, 0 an optimum basic feasible solution has been reached. The optimal solution to the given LPP is x, =0; x2 = 1, Maximum 4 ‘Ques 40) Use dual simplex method to solve the LPP. Max Z=-3x-2m Subject to x1+%221 xtms7 xy + 2x22 10 ms3 x1 %20 |Ans; Interchanging the 2 inequality of the constraints into S, the given LPP becomes Max Ze — 3x12 Subject to =x %e-} x+ms7 =x) 2x - 10 0x, +053Ds BCA, Thin Semester TP Solved Series (Operations Research) RTMNL Su, the standard form of the LPP becomes By introducing the non-negative slack variables S). . Max Z=~3x,~28; + 0S, + 08, + 0S; + 0S, Subjectto —x)-4 +, xith+S: =x) 28+S)=— 10 Ox, #424 S.=3 ‘The initial solution is given by 1S: 2 Incoming variable Since all Z, ~ C, 0 and some b, < 0 the current solution is not a basic feasible solution. bs = - 10 being the most negative, the basic variable S; leaves the basis. Also Max (2, ~ Cau, ae < 0) = Max (3/1, 2/- 2) =~ | the non-basic variable x, enters the basis. Tableau Ws |ts]o]-|o] oft Incoming variable Drop S, and introduce x;. 2.<0, S; leaves the basis. = Sirce all Z,~ C, 2 0 and ail b, 20, the current solution is an optimum basic feasible solution. ©. The optimum solution is Max Z = ~ 18, x; =