Duality in Linear Programming “Man is simply t0 recognise that there are different approaches for solving problems” 5:1. INTRODUCTION Associated with every linear programming problem (maximization or minimization) there always exist another Linear Programming problem which is based upon the same data and having the same solution. The original problem is called the primal problem while the associated one is called its dual problem. It is important to note that cither of the two linear programming problems can be treated as primal and the other as its dual. The two problems, thus, constitute a primal-dual pair. The concept of duality is based on the fact that any linear programming problem must be first put in its standard form before solving the problem by simplex method. Since, all the primal-dual computations are obtained directly from the simplex table, it is logical that we define the dual that may be constituent with the standard form of the primal . . GENERAL PRIMAL-DUAL PAIR Based on the standard form of primal, there are two important primal-dtial pairs >) Definition 1. (Standard primal problem) f Maximize 2 = ¢)x) + ¢p% + + GX, subject to the constraints = 5: ky + Oak, + ose + Minky = is 420s a Dual Problem Minimize 2* = byw, + B,W, + +. + ByWq subject to the constraints Man Bab ji, + day Hone + Oyj 2 5 6 De see May Bebe CT aym + aye nym 2 Aw 7 Wy (= 1s you m) unrestricted. are the primal variables, w;’s the dual variables and the other constants have their Note that usual meanings. Definition 2, (Standard primal problem) Minimize z= ¢)Xy +,€3% + + Gy%q Subject to the constraints = ky + 0% + one + Oink = bys 4205 Dual Problem Maximize 2% = byw; +-DpW + se + Dplq Subject to the constraints cum Mat Sa Si w (aed, 2, «...m) unrestricted.~ 130 OPERATIONS RESEAR = Remark, From the above definitions one may easily observe the following: ery primal constraint, there is a dual variable, (b) For every primal variable, there isa dual con (©) The coefficients of the di bles in the constraints are the Primal variables except that they are transposed, i.c., columns in pF rows in the dual coefficient matrix. (@ The number of dual constraints number of dual variables is exactly equal to the mi (©) The objective coefficients of the primal var constraints, whereas the right-hand side constants o coefficients of the primal problem. ‘The information regarding the primal-dual obje variables may be summarized in the following table : (a) For me as the coefficients of the coefficient matrix become the is exactly equal to the number of primal variables, whereas the umber of primal constraints, ‘bles become the right-hand side constants of dual fF the dual constraints become the objective ive, type of constraints and the signs of the duat Standard primal Dual objective Objective Constraints Variables ‘Maximization| ‘Minimization > Unrestricted To Minimization Maximization s Unrestricted 5:3. FORMULATING A DUAL PROBLEM Various steps involved in the formulation of a primal-dual pair are : Step I. Put the given linear programming problem into its standard form. Con: ler it as the primal problem. ‘Step 2. Identify the variables to be used in the dual problem. The number of these variables equals the number of constraint equations in the primal. Step 3. Write down the objective function of the dual, using the right-hand primal constraints. If the primal problem is of maximization type, the je constants of the dual will be a mininfization problem and vice-versa. ‘Step 4. Making use of dual variable identified in Step 2, write the constraints for the dual problem. (a) If the primal is a maximization prob’ m, the dual constraints must be all of "2" type. Tf the primal is a minimization problem, the dual constraints must be all of *< typé. dual () The column coefficients of the primal constraints become the row coefficients of the constraints. 7 (c) The coefficients of the primal objective function becomes the right-hand side constant dual constraints. (d) The dual variables are defined to be unrestricted in sign. Step 5: Using steps 3 and 4, write down the dual of the given L.P.P. sof the standard form of the primal is constraint associated with at [Note : ‘The duat constraint comresponding:to an artificial variable in th always redundant, hence it is never necessary to consider the dia anificial variable ‘dual pair is canonical form, the pri Remarks. J. {f the given linear programming problem is ‘said t0 be symmetric. 2. If the given linear progréitming problem is in its standard form, the p unsyrmetric. a mal-dual pair is said ¢0 be| ae ALITY INV LINEAR PROGRAMMING. 431 pul . PRIMAL-DUAL PAIR IN MATRIX FORM standard Primal Problem Definition 1. (Standard primal problem). Find x7 € RY so as to maximize z= ex, € € RY subject t0 the constraints + Ax and x= 0, b © Re" where Ais an mx real matrix. Problem. Find w" © RW so as to minimize 2* = blw, b © RM subject to the constraints : Alw eel ce RY where AT is the transpose of an mxn real matrix A and w is unrestricted in sign Definition 2. (Standard pri subject to the constraints : Dui ial problem). Find x" € R" so as to minimize z cx, ce RY Ax=b and x20, ble Rm where A is an mxn real matrix. Dual Problem. Find w" € RY, so as fo maximize zt = bTw, b € R" subject to the constraints Alw scl, ce RY where AT is the transpose of an mx real matrix A and w is unrestricted in sign. me SAMPLE PROBLEMS S01. Forinulate the dual of the following linear programming problem : Maximize z = 5x, + 3x, subject to the constraints + Ky + Skp $15, Sty + 2x $10, x, 20 and x20. Solution. Standard Primal. Introducing slack variables s, > 0 and s 2 0, the standard linear programming problem is: Maximize Sx, + 3x + O's, + Oss subject to the constraints Bx + Sky + sy + Os, = 15, Sxy + ey + Oy + 5p = 10, xy 2 5, 5p 20 Dual. Let 1 and w, be the dual variables corresponding to the primal constraints. Then, the dual problem will be Minimize 2* = 15, + 10, subject to the constraints : Sw + Sw 2 5, Swy + 2p 23 wy + Ow, = 0 Ow, + wy 20 w, and w) unrestricted (redundant), } > my 20 and w, 20 The dual variables “Ww, and w» unrestricted” are dominated by w, > 0 arid w, > 0. Eliminating tedundancy, the restricted variables are w, > 0 and w, 2,0. 502. Write the dual of the L.P.P. : Minimize 2 = 4x; + 6xy + 18x) subject to the constraints + xy + ky 23 xy t+ ey BS and xy sy xy 20 Delhi BSc. (Stat) 2006] = Oand 5, 2 0, the standard form of Solution. Standard Primal. Introducing surplus variables s the LPP. ints Ax, + Ory + [Bry + Oss + Ouyy subject to the const ny + Bey = 5) = 3 Oxy tty Baym sy = Sp AD AG Spe dy 2,0,OPERATIONS RESEapy, cy al COMSIFAIM, the dua 132, esponding (© Dual, Wf wy and wy be he dual AP ples corresponding presen wal = juny + Sivy_ subject (© the constraints © wt 3 soy + 2g SHB vey Omg Sa aE ME 6. my 20 ON pry $04 ane OHNE wy and 2 unrestricted (redundant). Eliminating redu indancy, the dual problem is . Eliminating todas mize ct = 301 + 502 subject to the constraints ey amy tomes 6 2 ig; wy 20 and w 2 0 : ing lit g propiem + vrite the dual of the sowing linear programming pro" 1 eo alone Peisay — Diy + Ay sublect rahe constraints ¢ Get 8 AGE OAT day 24 Fey dy 85 510 5x23 ey + Te — 289 22 my 20, x 20, and ¥ 20. [Visvesvaraya M.B.A. (Dee:) 201) Solution. Standard Primal. Introducing ‘surplus variables 51 2 0: $2 > 0, 54 2 0, 55 2 0 and g slack variable s3 > 0, the. standard form of L.P.P. is: Minimize z= 3%) > 2% + 45 * Os, + Oss, + sy + O55 subject 10 the constraints : oa +S 1 2 4 5 3x + 5m +47 7 6 + ot 3u-—2e4 Ixy 2 - +53 = 10 3 xy — 2 + 5 dx, + Ti ~ 25 - 55 ape ps pe Se 830836 846 5 Dual. If w,(j = 1s 2 3: 4 5) are the dual variables correspondin; in the given order, the dual of the standard primal LPP. is : = Tw, + ivy + Ors + 3g + 2s 20. 1g to the five primal constrains subject to the constraints + Maximize z* = Buy + 6w + Tws + we + 4s $3 Sw + Wy — 2Wy — 2Wy + Ts $2 dw, + 3p — wy + Sing ~ 2ws $4 =, £0, = $0, wy $0, ~ m4 $0, ws $0 wjUi = 1, 2, 3, 4, 5) are unrestricted in sign. > 0, wy 2 0, Wy <0, wy 2 O and Ws 20 Eliminating redundancy, the dual variables are : Nate in the primal iff we multiply the third constraint throughout by —1, the dual variable w3 will be20 504. brain the dual of the following L.P.P. : es rm * 3x, + xy subject fo the constraints : 1 = 6, xy + 2p + Sky = 45 Xp xy X20, [Madras BE. 199 11 standard form. Considering # ® Solution, Clearl; i 7 ly, the give i ‘ anlaton Cleary, the given linear programming roblem Minimize 2* = 6, + 4w si straints we rk = ji 1 + div, subject to the constraiitts : dy + 99 2 me 3 + 3wy + 2we 2 3, wl 23, here w, and w are dual variables, w+ Sip 212 wy and wy are ‘unrestricted, ”puauitY IN LINEAR PROGRAMMING ; 133 505. Obtain the dual problem of the folk 3 ins following primal problem: \ Minimize 2 = x) ~ ax ~ 2x5 subject to the constraint sei constraints : VOM A BG ST Quy dey 2 12, wy + diy ~ 30y BS wy + 8wy $ =2 201 ~ Buy S 2 wo = $0 wy, Wp and wy unrestricted, Eliminating redundancy, dual variables are w, < 0, w, 2 0 and w, unrestricted. This is re-written as follows : | on amet | = ms ang m2 0 Maximize 2+ = Tw, + 12, + 10w, subject to the constraints : Bay + 2wy — dws SL, wy + diy — 315 2 3-2, = 85 Ww) $0 and wz 2 0, wy unrestricted. Note. In the primal, if we multiply the frst inequation by -I; the dual variable w will be 2 0 instead of < 0. PROBLEMS 506. Formulate dual of the following L.P.P. : (@) Maximize z= 4x, + 2p subject to the constraints: yt 234-3224 4520 and 20. Annamalai MLB.A. (Nor) 2009] (6) Maximize _z = 2000x + 3000) subject to the constraints Gr + 9y $100, 2+ y $20; x20 and y20. [Osmania M.B.A. (Sept) 2001) 307, Formulate the dual of the following LPP. Maximize z= 10x + 8rp_ subject to the constrains #225, Wy ZIA Ht d_ed xj 2.0 and xy is unrestricted. [Osmania MBA. 1999) $08. Obtain the dual problem of the following LPP. : Maximize f(x) = 2t) + Sty + 6xy subject to the constraints Sey + 6 em Sh 2) tH tay S4 HSH HAN Sh mtay 0 it i =Wp- Mp / 2 0 in the constraints of i = Pe oud wiela f the dual and write w = w, - 2 The standard form of dual then is to Minimi: = br ize g(w) = wor, ~ 2), b” € R subject to the constraints : ANG ~ 2) ~Iys =e, ce RM Wp Wo and s > 0, Problem as our standard prima Considerin, wil te this linear programming the associated dual problem Maximize ih(y) = ¢, EY CER" subj Oy Tat Jginniee sgheonsains 5 “YS 0 y2 0) and y unrestricted,— suit LINEAR PROGRAMMING 135 ou inating redundancy, the dual problem may be re-wi Maximize h(y) = ¢ Ay sb] Ay > bj yao. his problem, which is the dual of the dual problem. is just the primal problem we had started ii ritlen as © E RY subject to the constraints : Ay = bo bre Re with ‘This completes the proof. Theorem 5-2 (Weak Duality Theorem). Let x, be a feasible solution to the primal problem Maximize f(x) = ex subject to: AX cl, w20 Proof. Since x, and w, are the feasible solutions to the primal and its dual respectively, we must have : Axo Sb, X20 and ATw, 2 e%, wy 20. Thus © S WTA OF €X5 $ WoT AN, S Mlb or €%) € by {since w!b = bY wh Theorem |. Let x, be a feasible solution to the primal problem Maximize f(x) = ex subject to: Ax Sb, x20 and w, be a feasible solution to its dual : Minimize g(w) = b"w subject to: ATW 2c, wz 0 where xT and © € R®, wT and bY © RM" and A is an mXn real matrix. If cx, = DT Wy then both x, and w, are optinum solutions to the primal and dual respectively. Proof. Let x5 be any other feasible solution to the primal problem then theorem 5-2 gives ex SDP, Thus CX" SX. (since ex, = BF w,) and hence x, is an optimum solution to the primal problem, because primal is a maximization problem, Similarly, if w,* is any other feasible solution to the dual problem, then bTw, < b’w,* and thus \, is an optimum solution to the dual problem, because dual is a minimization problem. Theorem 5-4 (Basic Duality Theorem). Let a primal problem be Maximize f(x) = x subject to: AX el, w20 wh bh eR" I x, (w,) is an optimum solution to the primal (dual), then there exists a feasible solution 0 (%,) to the dual (primal) such that ex, = be Proof. The standard primal can be written as Maximize z= ex subject to: Ax + € Riis the slack vector and I is th where x,7 sociated identity matrix. aOPERATIONS Reseyq ie al, where Xp” € R" is the optimay ,, 7 : ml of A. Then the optimum pint 136 fon to the prt ohuti imal B being the opt ean optimum s° 7 [xpo O14 en b Let x, feasible solution £ objective function is = Bb, xp = B = Oxy = HM fated With Xp 1 simpl soci is the cost vector as S ven by the opt lex table are given by sin all ay € A eg ila — cj forall ay € =~ G* legBtg -0 forall get — ¢, 2 0 for all j. This gives 2 where ¢5 Now, the net evaluation ~ omtitnal, we must ave J 7 for all j Since xg is optimal. ee Hi f I in matri; cpBAZe and Gp B12 0 (Gn matrix fom, or eee or . the above become ATw, 2% and wy 20, wo" € RM" problem. Moreover, the corresponding dy, Now, if we let Bo! ey” = Wo This means that Wy is a feasible solution to the dual objective function value is, cg Bo! b = cy Xp = Soe Dig = Wold 5 a feasible solution w, to the dus Thus, given an optimal solution x, to the primal, there exist such the cx, = b’w,. Similarly, starting with w,, the existence of x, can be proved. 7 Corollary. If Xq is an optimal solution to the primal, an optimal solution to the dual is given by B-!cy, where B is the primal optimal basis. Note, Observe that B! eg” represents optimum zj — cj under primal slack columns. Theorem 5-5 (Fundamental Theorem of Duality). if the primal or the dual has a finite optinun solution, then the other problem also possesses a finite optimum solution and the optimum values of the objective functions of the two problems are equal. Proof. Consider the primal-dual pair Primal. Maximize f(x) = ex subject to: Ax 7 Dual. Minimize g(w) = b’w subject to: ATw > c? and w> 0. ‘ecessary condition). Let a feasible solution x, (w,) be an optimum solution to the primal (dua!) Problem. It then follows from Theorem 5-4 that there exi ! s lere exists i Vo (X cu (primel) problem, such that ex, = bTw,. oe. 0. > Xt now follows from Theorem 5-3 that w, must be optim: This proves-the necessity of the condition, (Sufficiency). it follows from Theorem 5-3 Theorem 5-6 (Existence hs 5-6 (Bs Theorem). If either objective function yalue, then the aller moby P . oo et Lt {he given primal problem have vecine function, say +e, there existe. af a fea x — & If possible, let n feasible solution w. of “he arten " €% SB That fs, br ial. the primal or the dual problem has an unboundé! Mas no feasible solution, yf the unbounded solution, ‘Then for any value o! ies Ve es Solution say x yielding this solution, i I, there exists wl solution, Then from Theorem 5-2, for & Teasible solutiy ii suc! Wo ion si "1 > Now as b ig const: X, to the primal int ‘ant and w, has to satisfy the consti!rr quALlTY IN LINEAR PROGRAMMING " ar, 2 € therefore the dual objective function Fw, 5 Must be pr, > & Hence the problem has no feasible solution, py a similar argument it can be jas ao feasible solution wndard results on duality nite. ‘This contradicts the result shown that when ‘an unbounded solution, the primal zed as follows : - Primal roe ual problem __FFeasbe ston Na fase su — Fee Soton Optinun Dual Uabounded else Soon Pina Und Unum Tesh Yn the following section we shall see how th the relationship betw ‘ ie above existence theorem helps us in understanding fen the optimum values of the primal and dual variables. COMPLEMENTARY SLACKNESS THEOREM Theorem 5-7 (Complementary Stackness). Let x, and w, be the feasible solutions to the primal (max. e's | AX SD. x 20} and its dual { min. biw | Alw 2, w =O} respectively. Then, a necessary and sufficient condition for x, and wa to be optimal to their respective problems is that Wel ~ AX.) = 0 and x,T(ATwy ~ &?) Proof. Necessity. Let & = w,(b ~ Ax,) and B = x,'(ATw, ~ €'). Since xy. w, are feasible solutions to the primal and dual respectively, we have @20, P20 and a+ B= wlb-x,'e% Now, if 4 W, are optimal, then ex, gives a = O and B = 0, Thus the conditions are necessary. Sufficiency. Let the given conditions hold for the feasible solutions x, and w,, That.is, ct iw, so that ct + B = 0, But since c 2 0 and (b > 0, this, 0 and B = 0. ‘Then = a+ B= weTb~ xfer = €X) = BIW Xo and wa are optimal. Thus the conditions are sufficient. Corollary 1. if x° and w° be feasible solutions to the primal and dual problems respectively, then they will be optimal if and only if 12 seam and ap CE, aun ~ @) = 0. Proof. From the above theorem, x° and w° will be optimal, if and only if Pb = Ax.) = 0 and x,P(€" ~ Afwg) Consider the first set of conditions. Since each term in the summation non-negative, it follows that wy'(b - Ax, is aysf)= PSL am,= nc 138 OPERATIONS re Stn Similarly, the second set of conditions is equivalent to 5 E, qui? ~ ) = 0. Ps Corollary 2. For optimal feasible solutions of the primal and dual systems, whenever th variable is strictly positive in either system, the ith relation of its dual is an equality, "eit Proof. It follows from corollary , that = b (ith primal elation) and se 0 EF ayue = (ith dual elation Corollary 3. For optimal feasible solutions of me primal and dual systems, whenever ith rela of either system 1s satisfied as a strict inequality, then the ith variable of its dual vanishes, Proof. If follows from corollary J that E asp w= 0 and E agwe>q = sf Remarks. The conditions of corollary 1 can also be written as and 5? sWgy = 0 i where sy) i8 the ith slack variable in the primal problem and nq is the jth surplus variable in the ual, Thus, the theorem relates the variables of one problem to the slack or surplus variables of te other. The above relations are called ‘complementary slackness’ because they imply that whenever a constraint in one of the problems holds with strict inequality (so that there is slack in the constrain, the complementary dual variable vanishes. :7. DUALITY AND SIMPLEX METHOD Since any L.P.P. can be solved by using simplex method, the method is applicable to both the prinsl as well as to its dual. The fundamental theorem of duality suggests that an optimum solution to th associated dual can be obtained from that of its primal and vice versa, If primal is a maximization problem, then following are the set of rules that govern the derivatiot of the optimum solution Rule 1. Corresponding net evaluations of the starting primal variables | = Difference between the left and right sides of the dual constraints associated with the starting primal variables. Rule 2. Negative of the corresponding net evaluations of the starting dual variables 2 Difference between the left and right sides of the primal constraints associat with dual starting variables. Rule 3. If the primal (dual) problem is unbounded, then the dual (primal) problem does "0 any feasible solution. Note, Jn rule 2. dual problem is to be solved by changing the objective fiom minimization 10 Me ‘maximization. have- pualitY IN LINEAR PROGRAMMING 139 SAMPLE PROBLEMS 515. Use duality 10 solve the following L.P.P, : Maximize z= 2x, + x subject to the constraints : Sy 2z S10 xy +5256. y= SQ np = Dey SNE ty ty 2O. [ethi B.Sc. (Stat.) 2006] Solution. The dual problem of the given primal is Minimize 2 wey 422 Dy + wy = wy — Img 2 Me ye Hye Wy 14 2 O. Inoducing surplus variables s; 2 0, s, 2 0 and artificial variables A, 2 0, Ay 2 0, an ini tusic feasible solution is Ay = 2 and A, = 1. (The primal constraints associated with 5}, 53. A; and azare: —8) $00— 2 $0.5, < Mand xy wa 2 fl 1 1 1 @® 1 = 3M -3M¥10_—2i46 2 Onset o First Iteration, Introduce y, and drop yy. iz : es oe eee Faas Yea ie Voie Yo ype Se M 32 0 rR GD 2 =1 12 1 Wr _-10 ie 1 win -t Oe cigeeo 2 a 0 kua 73M 12MM SMR 053M S32 Second Iteration. Introduce yy and drop y,. Yu Ye vn Ya yy ora Ys) Yo yy Ye vs 1 0 1B 1 43 28 iB BUS v1 Mf aati 0 13-18 1B B aaeeee 12 0 3 0 3B 4B 8B M143 M83 Final Iteration. Optimum Solution. Yo Wp Mt ¥ Ys Y4 Ys Yo 7 Ys Ys wa -12 «0 1 wm -IR SSC Ye 312 3 l oO -e in na in ae =10 2 0 0 14 2 Ma MD Thus, an optimum feasible solution to the dual problem is wy = 0, wy = 3/2 and wy = 1/2; min 2 (— 10) = 10. ‘Also, the primal constraints associated with the dual variables A, and Ay are xy ¢ Mand x cy = 230, where ey = (2 5 0 hus, we Bet 4 = 430, a = 460 and ay = 480, Tp Tie svow give $558 62th Sho 3 ah este $-0=2 ‘To obtain the values of by, Dy and bs, we perform iteration on the starting primal table Initial Heration. Introduce yy and drop ys. 3 2 5 0 ° o e Yo xy yi v2 Ys Ya Ys Ye 9 Y= 430 2 1 1 0 0 ° Ys m= 4603 ° ® 0 1 o ° Yo ay = 480 4 ° 0 ° L ‘i 0 3 2 3 0 0 ° First eration. Introduce y3 and drop yy. 6 Yo Xo vt v2 n Ys Ys ¥6 0 Ys 200 -12 @® 0 1 12 0 5 va 230 3 0 1 0 nn ° ° ¥% 480 L 4 0 ° 0 1 1150 92 2 0 0 5h. 0 Second eration. Optimum Solution, Ss Yo Xp yi vo 3 Ys Ys Yo 2 Y 100-14 1 0 12 =u ° 7 % 230 3A o t ° in o BOE ety 80 4 0 0 2 2 1 eats 1350 4 0 0 1 2 0 by = -1/4, by = 3/2 and by = ~4. Note that the values of e,, ¢ are also readily available.) (i) The dual problem is Minimize 2° = ayw, + aw, + aywy subject to the constramts : Wy + Bip + wy B 3, wy hwy + Awy = 2% wy + 2p + Oy BS mw, 20, w, 20 and wy 20, d with the starting primal vai 1 oe dual constraints associate and w, > 0, Wes xy xy and Xe are wy 2 0,142 ‘Thus, we have the following informa Thus using Rule J we get cq am -0 = Q=m-0 > O=my-0 => ‘The optimal dual objective is Min, 2° = 1,350 = aywy" + ayia" PROBLEMS Write down the dual of the following L.P.P. and solve + 518, Maximize <= 8x +-4xy_ subject to the constraints ny + 2p $30, 2p #48 S 2H pe ZO. (Osmania 519. Minimize 15xy + 10x subject to the constraints : Bey + Sry 2S, Sup + Ixy 23, Hy wy BO. 520. Maximize ¢ = Sxj + 2x2 subject to the constraints : 6x, #26 4m $3212 4 + 224 ZO {Sambalpur MBs 521. Minimize = = 2x + 9x2 + x subject to the constraints xy + dey + 255 25, Sy Hy +2 24 and xy ty 4 20. 522. Consider the following U.P.P. Maximize z= x, + Sty + 3x3 subject to the constraints : ay + 2ty tay = 3 Qep ay = AG a tye my 20. (a) Write the associated dual problem, : (6) Given the information that the optimal basic variables are . solution. 523. Given the LP.P. : Maximize z= 2xy + 4ty + 4xy — 3xq subject to the constraints : AAR FG TA Ae Hany = BE xp ay ay ay BO. Use the dual problem to verify that the basic solution (x). x2) is not optimal. and x3, determine the associated optimalé 524. Find the optimal value of the objective function for the following problem by only inspecting #8 Minimize z= 10x, + 4xy + Sy + xq subject to the constraints : i 3) = Tey # Bry # O5xy 2150; 20 (= 23.4) (Karnataka BB. (Pe)! 525. Use duality to solve the following L.P.P, : Maximize 2 = xj ~ 4 + ks + 2ny_ subject to the constraints: ate hay SST xp vag dys -2p 20 (F214 2340 mn Hed [Delhi B.Se. (Mel 526. Use’duality in solving the L.P.P, : Minimize 2 = sag, itimine @ = 2x ~ xq + xy 4 Sy ~ 3xg_ subject to the consta a BHR By ty tay tage 2s 20 (f a. 4. 5). 527. Consider the problem - Maximize z= 2x ~ 5x) subject to the cons ot, Ht 2D DW toy + ry SG ay ~ ay + By () Write its dual. - 4 ‘ we (ii) Solve the primal and then find the solution of the dualR PROGRAMMING. iy IN LINEA ual 143 28, Solve the following LPP. 2+ 3 subject 10 the constr “tea Oy Meany 20. oblem ts SB ary Hay eyed Find te solution of the corresponding dual p 529. Consider the problem : Maximize = from the final mplex tableau. (Delhi W.Se, (Stat) 1998] subject to the constraints ; br + Ae bay < 40, - meh, Mite 2 0, Write the dual of this problem. Solve the dual and from there find the solution of F the primal problem, Wethi Se. (Sat) 2000) and D respectively and Manimize 2 = dn + 6% + 355 4 24 subject to the constraints : LSt + On + 445 + any < 550 Ant tet xs 700 2H + Bm + y+ 2x < 200 Fy tay 420 $50, Ina productmix problem sy. x3, 3 and xy indicate the units of Products A, B,C we have 10 (Machine J, hours) (Machine 11, hours) (Machine /11, hours) (@) Formulate the du. of the above problem, (Solve the dual. From the solution of the dual, read the soltion of the primal problem. (0 I machine 1 remains out of order for 24 hours, what will be the loss tthe manufacturer? {2 W te profit contribution from product A increases from RS. 4 to Re, 5 Per unit, what will be the effect on optimal product mix? (Delhi M.Com. 2007} S31. Consider the following LP! Maximize z= Sx; + 2x, + 3xy subject to the constraints : Ait Sy + 2S by, ay Sty ~ Gy o where by, by are constants. Fo sp al table is obtained: 5 by ay tay values of by, b, the following opt o Yo Xo y v2 Ys vs vs 5 yb 30 1 6 2 1 ° 0 ¥s 10 € 3 +1 1 z= row 150 a 7 d ji constants a, 6, ¢, d and e can be obtained from the data of the given LPP and the optimal feasible ‘ndition of the table, Determi (©) the values of R.H.S. constants by, by. (0 the optimal dual solution (from the given table). (© toe values of a, and e associated with the variable 4 ods Given below are the object funtion, the constraints and the simplex tableau fora linear programming "mix problem : Maximize z= 2x + Sey + Bxy + Os, + sp + sy subject to the constraints: Gry + By + Any 596 (hours/week, Department 1 Dey + my bd 5 40 (burdveck, Depart 1) Sx, + 34 4 Dey 5 60 (ourdweek, Department 22) Mp ty ay B09