Scribd
Upload
28 views12 pages

Adobe Scan 23 Sept 2023

Dfffg

Uploaded by

Anshuman Sahoo
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF or read online on Scribd
28 views12 pages

Adobe Scan 23 Sept 2023

Dfffg

Uploaded by

Anshuman Sahoo
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF or read online on Scribd
You are on page 1/ 12
Linear Programming Problem —Simplex Method “One step for better is good enough 4:1. INTRODUCTION As discussed earlier, the Simplex Method, also called the ‘Simplex Technique’ or the ‘Simplex Algorithm’ is an iterative procedure for solving a linear programming problem in a finite number of ‘eps, The method provides an algorithm which consists in moving from one vertex of the region of asible solutions to another in such a manner that the value of the objective function at the succeeding vertex is less (or more, as the case may be) than at the preceding vertex. This procedure of jumping from one vertex to another is then repeated. Since the number of vertices is finite, the method leads to an optimal vertex in a finite number of steps or indicates the existence of an unbounded solution In the present chapter, we shall introduce and explain the computational procedure of the simplex method. The theory behind the method will be first developed and then the computational techniques explained and illustrated. The following important definitions are necessary for understanding and developing the theory that follows : Definition (Basic solution). Given a system of m simultaneous linear equations in unknowns (m caxp where &y is constituted of cost components corresponding to %y Definition (Optimum basic feasible solution). A basic Maximize z = ex subject to: Ax = ‘5 called an optimum basic feasible solution if z, = ¢y x, > function for any feasible solution, Jeasible solution xq 10 the L.PP b and x20 2 where 2* is the value of objective PROBLEMS id all the basic feasible solutions of the equations 2x + 6 + Oy + y= 3 Gry + diy + Any + Oxy = 2 404, Find all the basic feasible solutions of the following L.P.P. without using the simplex alogorthm Maximize z = 2xy + 3x + dxy + Txy subject to dy Ay tay tay 8 4-2 + 6 - Ty y2Q iL AM 403, , 405. If xy = 0. xy = 0,15 = 044 on eo ey a dy + 4x + y= 12 Hairy + Big + Bey + my = 10: : Wve problem. find two more extreme point sdlutions of the above probl OPERATIONS RALSEARGH <0 4:2. FUNDAMENTAL PROPERTIES OF tfc ational procedure for obtuning. an We are now well set to discuss the theory behind th oa raperton of tie claligie iaiiie ‘ optimum basic feasible solution 10 a given L.P.P. The J simplex algorithm. LPP. that now follow, will allow us to arrive at the much awaited Se naan Theorem 4-1 (Reduction of a Feasible Solution to # Basic Feasible Solution): LPP. hay « feasible solution, then it also has a bast feasible solution. Proof. Let the 1.PP he to determine x s0 as to maximize subject to the constraints Ax = b,x > 0; where A is an myn real matrix 1 xn real matnces respectively. Let p (A) = Since there does cxist a feasible solution, we must have IA, b) = ptA) and m m Let £0). Op, 1. Oy} be a Set Of constants (not all zero) such that 1B) + yay +. + Uy Suppose that for any index r,4, ¢ 0. Then a, = $ Ye 7 ie 8, ‘ Zang + | Sau em see Gee on |. Thus, we have 4 solution with not more than p — | aon are positive, we shall choose a, in such way tha” Sn-Zer0 components. To ensure that these 20 for all jay, 91 LINEAR PROGRAMMING PROBLEM—SIMPLEX METHOD This requires that either cy, = 0 or a Pa) Hay > 0 and 2: Thus, if we select a, such Som [2a of «, then for each of the p~ 1 vi 7 bles x) — 4, a, 8 NON Megative, and so we have a feasible solution with not more than pI non-zero components, Consider now this new feasible solution with not more than p—1 non-zero components. If the corresponding set of p — 1 columns of Ais linearly independent, case (i) applies and we have arrived at a basic feasible solution. If this set is again linearly dependent, we may repeat the process to urine at a feasible solution with not more than p-2 non-zeto components. The argument can be repeated, Ultimately, we get a feasible solution with associated set of column vectors of A linearly independent, The discussion of case (i) then applies and we do get a basic feasible solution. This completes the proof, Corollary. There exists only finite number of basic feasible solutions to an L.P.P. Proof. Exercise for the reader. Theorem 4-2 (Extreme Point Correspondence). A basic feasible solution to an L-P.P. must correspond to an extreme point of the set of all feasible solutions and conversely. Proof. Let the L.P.P. be : Maximize z= e'x, xe R" subject to the constraints: Ax=b x20 where A, b and ¢ are real mxn, mx 1 and nx 1 matrices respectively. Let p (A) = m. Let S be the set of all feasible solutions to the L.P.P. Also suppose that x is a basic feasible solution [7] 0 where xp is an m x1 vector, such that for a non-singular sub-matrix B of A, Bxg If possible, let x be a point of S, such that there exist x), x; € S such that x, ¥ x, and x= Ax, +(L-Am O 0, we can write 2 2 E yay te gab. Gel te ee 2 - £ (+ €0)a = b. (some a, + 0) ‘Tres. the two different points XY = (+E Gy NEO nw H+EG), 0. 04 = 29) = (Hy -Edy, -EOy -44 5 )-EOp, 0, 0... 0) samsfy the constraints Ax” = b. Also. since x > 0, by selecting € such that 0c < min | A Jmh2.. the ferst Ss of xi) = = ‘ P component: and x will always be positive. As the remaining (n — p) x) and x@) » are feasible solutions different from LINEAR PROGRAMMING PROBLEM—SIMPLEX METHOD 8 degenerate extreme point), then one can select (m — p) additional vectorts from A with their corresponding variables equal 10 zero, such that the Fesulting set of vectors are linearly independent. This completes the proof Remarks 1. The number of extn the preceding theorem that there is. onl conversely. So there is one-to-one correspondence hetween the exe solutions. But basic feasible solutions are finite in number, thecekan set of feasible solutions are also finite in number 2. An extreme point can have at most m positive x/ Heme points and the basic feasible fe, the extreme points of the convex js where m is the number of constraints Theorem 4-3 (Fundamental Theorem of Li LP.P. is a convex polyhedron, then there ¢: basic feasible solution must be optimal, Proof. Let the L.P.P. be to determine x so as to Ineur Programming). If the feasible region of an xists an optimal solution to the L.P.P. and at least one clx subject to: Ax = b and x20, where ¢, x € RY, Then the feasible region S of the L.P.P. is given by S={x|Ax=b, x20}. Since S is a convex polyhedron, it is non-empty, closed and bounded. ‘he objective function dz = efx, x € S is continuous on S, which is non-empty. closed and bounded, therefore, z attains its maximum on S. This proves the existence of an optimal solution. Now, since S is a convex polyhedron, it has finite number of extreme points. Let these be x, Xe vn % € S. Clearly § = . Therefore, any x € S can be expressed as a conver combination of the extreme points, say k X= Egy: 20, Say= 1 for all j= 1,2, k Let max. (eT) j= 1, 2, k). Then for any x € 5, ‘ y = ET Lajx) = Lay (Tx) < Bayz, = zo zS% forany xe S. Thus, the maximum value of z is attained only at one of the extreme points of S. That is, at least one extreme point of S yields an optimal solution, Now, since each extreme point of S corresponds to a basic feasible solution of the LP.P. therefore, at least one basic feasible solution is optimal. This completes the proof. z Remark. tis easily seen that if the optimal value of z is attained at more than one extreme points of S. then every convex combination of such extreme points also provides an optimal solution to the L-P.P. ‘Thus. the optimal solution to an L.P.P. is either unique or infinite in number, SAMPLE PROBLEM 406. Ler x 4 and x; = 1 be a feasible solution to the system of equations 2x; - xy + 2ay = 2 ay + bx 18, Reduce the given feasible solution to a basic feasible solution. Solution. The given system of equations can be written as Ax = b, where 2-12 ele a-(7 7. »-[ 3] OPERATIONS RESEARCH ak (A) = 2. there are only two linearly independent columns of A. Thus, if the column since. spectively, then vectors of A be denoted by a, my and ay Lothen ay = yay t Where a's are sealars not all zero. I we let en This ives 2 = 2m, — oy and O = ay 4 Ae, Solving these equations, we fel ty = 879 and olay Now ‘ I ia . fe min | >of win {do} the new solution with not more than two This indicates that vector ay can be removed, Hene' non-zero variables is given by Se2-2xhen, hy PP. is 4p = 9/2 and xy = 13/4 with x = 0 Thus, a basic feasible solution to the given L (non-basic). Note. If we let ay = 1, another basic feasible solution sy = 26/9 and ay = 4/9 PROBLEMS feasible solution to the system of equations ata ted 407. The column vector [1.1.1 is 2st Reduce the given feasible solution to a basic feasible solution. 408. Consider the system of equations. at dy tay 27 2ey - apt My ye 1 and 44 = 0 is a feasible solution, Reduce this feasible solution to two Here y= os hon different basic feasible solutions [Meerut M.Se. (Math.) 1998; Dethi B.Sc. (Stat.) 2000] 409. If x xy = 3. 44 = 1 be a feasible solution of the following Linear Programming problem, then find the basic feasible solution Maximize <= x1 + 2x; + 4x3. subject to the constraints : Dy ta ayy a Hh Be hag t Se = I, ave ky 20. 6] is not a basic feasible solution to the system ney tye? 410. Show that the vector [1, 0. nome? dey + 3 Theorem 4-4 (Replacement of a Basis Vector), Let an L.P.P. have a basic feasible solution. If ‘one of the basis vectors and introduce a non-basis vector in the basis set, then the new 4 ayy ay =O. we drop Seuation obtained is also a basic feasible solution. izing ¢ = ex subject to the constraints Ax \d mx I real matrices respectively. Proof, Let us consider the L.P.P. of maxi RY" and A, b are mx x 2 0; where x7, € © Let p(A) = ™: LINEAR PROGRAMMING PROBLEM—SIMPLEX METHOD 7 Suppose that there does exist a hasic feasible solution xq. sot a Bry = b. xy > 0 where B forms a basis set for the column vectors of A. Cleutly, for any column vector ae A,we have y= vy by + yyby + * Yay by = By, where by € Band yy, are suitable alas, Now. we Know that ifthe basis vector, for whieh the coefficient y, is non-zero i replaced by ye A then the new set of vectors also forms a Now, for y,, # 0, we can write and therefore (1) gives » aon b= Fam + [2 My ier is Yy tet My ” i o = £ | am - xp,2H |b, + 22 Gr in Yj i Thus, the new basic solution is y having as its components, the variables ‘ %y 5 Nar Nei = Xp) - Xp, 2 (i Qowm: i#n and %y, = “8 oq 9 . We shall _now show that Xy is feasible also, that is, the new basic variables %,, are also non-negative. Two cases do arise : Case (i). 2g, = 0. In this case the new set of basic variables is obviously non-negative, since we have assumed the existence of a basic feasible solution xy, Case (ti). xp, # 0. In this case we must have y,, > 0. This requires that for the remaining ¥,(i # A) either y,, = 0 for i # r, or XB Nae Ma My 2g Rw PO and ier and Bs. for yy <0 and ier My Me Ya My So, if we select the index r(y,; # 0) in such a way that or Min 2 b> ier we les } then, the new set of basic variables are non-negative; and hence the basic solution %, is feasible. Remark. If after replacement of basis vector, the new basis matrix is en) + then, the new basic feasible solution is where b, = by fori # rand 6 4,284» a i * XBr where ty = aq ~ Age ier and Spy: are the basic variables. Mi My Definition (Net evaluation). Let Xp be a basic fe: wible solution 10 the LP.P. : Maximize z= ex, where Ax =b and x0. Let eg be the cost vector corresponding 10 Xp. For each column vector a, in A, which is not a column vector of B, let iE Mabe y is called the evaluation co! corresponding 10 &, Theorem 4-5 (Improved Basic Feasible Solution). Le Xs vox subject 10 AK = D.X20- solution obtained by admitting @ non-basis column vector a; in — ¢; is negative. Then %, is an improved basic feasible Then the number 9 FG and the number % called the net evaluation responding 10 % be a basic feasible sol - Gis lution to the LPP. Maximize = Let %q be another basic feasible the basis, for which the net evaluation solution 10 the problem, that is AP by he > co Xo Proof. The L.PP. is to determine x, so as to Maximize 2 = ex; 6x7 cousinins: Ax=b and x20, b’e RY, wher A is an mxn real matrix. We are given that xp is a basic feasible solution. Let zo = Ca Xa Let 4, be the column vector introduced in the basis, such that z - 6 < 0 be the new basic feasible solution, then Let b, be vector removed from the basis and let Ry vy » = p,-£ and Spr Br, 2 5 cx? € R” subject to the ae Thus, the new value of the objective function is f= 0 Cadi > bg, = 6) Uv xp /yq > O1- Hence, the new basic feasible solution Xp gives an improved value of the objective fi jive function. Remark. If the given basic solution h: a : appens to be a non-de; en x, ae = > generate one, then xp, 0 and a Coote the value of z, occurs. If the given feasible solution happens to be a ee, fe one, then an increase in the value of z, depends on xg, /),, In any case, z; ~ cj < 0 implies that : ves at Z can never be ess than zy, ee Mag = 0 for at least one j for which y, > 0, i ion is obtained which gives an un Ine of th changed value of the objective f junction. Proof. {t follows from the above theorem that =m - (y= 6) 2H. I~ Dy > 0 =] . + 2, 2.) m; then another basic Theorem 4-6 (Unbounded Solution). Le ; jon). fora least one j, for which yy $ 0 (i a L oad exist a basic feasible solutio 7 i + 25 wey m), and 5 a exist any optimum solution to this L.P.P. mh and z — cy is negative, the oe pe » then there does noi 97 LINEAR PROGRAMMING PROBLEM—SIMPLEX METHOD Proof. Let the given L.P.P. be to determine x so as to Maximize == ex: x7 € RY subject to the constraints Ax=b and x20, ble R™ where A is an mx real matrix. Let a basic feasible solution to this problem be xp, so that Bxy=b and xy 20 with the value of the obje; fe function % = Cup = 2 cai XBe Now, we can write b= Bxp + 9; - Eg; a € A. Gis a scalar E ab) + ba - 6 E : EG - En) + 8 If E > 0, then (xg; — & yj) 2 0 since yy < 0. This shows that there exists a feasible solution whose (nr + 1) components may be strictly positive, But, in general, it may not be a basic solution. The value of the objective function for these (m + 1) variables is given by B= E calsm - By) +84 =F, ewan — 8, carry - But - <0 and E>0 Late as bste, Hence, there is no limit to the optimum value of z and, hence there exists an unbounded section to the given L.P.P. Remarks 1. If for some aj for which zj~ cj > 0 and yy $0, i= 1, 2, .. m, then there exists a feasible solution to the L.P.P., such that the (m + 1) components of the solution may be strictly positive and 4-0 as b> te, 2. If yy € 0 for at least one j for which z - ¢ = 0, then 2 = zp Theorem 4-7 (Conditions of Optimality). A sufficient condition for a basic feasible solution 10 an LP.P. to be an optimum (maximum) is that 2; - c; 2 0 for all j for which the column vector € A iy not in the basis. B. Proof. Let the L.P.P. be to determine x so as to Maxi z= ex; 6 x7 RY subject to the const Ax=b and x20 where A and b are mxn and mx I real matrices respectively. Let p (A) an m xm submatrix B of A as a basis matrix, Is: ‘m so that we can choose Let us assume that there exists a basic feasible solution x, to this L.P.P. Let eg be the cost vector corresponding to the basic variables. ‘Then Bxp = b, xp 20 and 29 = cpxp. Now. for all those j for which a, € B, we are given that z ~ c, 2 0. Let a, = b, for all such j for which a, € B. Then y= BT the unit vector (since y, = Ba) OPERATIONS RESEARG), = eng % 3-92 H- 9 and y-9 zi 6, 2 0 for all j for which a, © A. Now. let x be a feasible solution, Then ° Ginee a € B. egy = 6) (since x > 9) (since (since ey y, for all j tor which a, € B. Now, since ee eee le or ae = yyy therefore, the above inequation can be written as E esx 2 E63) OF guy 2 ex oF wz 2 here :* is the value of the objective function for the feasible solution x. Hence, < is an optimum for that basic feasible solutions for which yO > 0 for all j such that ae B Corollary. A necessary and sufficient condition for a basic feasible solution to an L.P.P. to be an opnimum (maximum) is that 5 ~ ¢, ® 0 for all j for which a, € B, Proof. It follows immediately from theorems 4-5 and 4. Theorem 4-7 proves the sufficiency of the condition, Theorem 4-8. Any convex combir optimum solution to the problem. Proof. Let x1, x2, .... x, be k different where A and b are mx~n and mx 1 real Let z, be the optimum value of z, Obviously, : ‘ where x) 20, x 20... 2 0, MSS Ax Now, any combination of x,, X, i ‘ X= 2 ox, where o sitive scalars such that z a, 7 ‘ Te O,'s are Uae Axe ACS : a in, WM) 4 7 jay MAX, Bebe g ; =b (ince 2a, = 1 is, Further, sinee x, 2 0 and g, s g's y "20 fora i= 45 non-negative. 2a k theron ' fon : Thus % is a Feasible solution ty the Lep. © every component of x will be

You might also like