Professional Documents
Culture Documents
3 Comparison of Alternative
3 Comparison of Alternative
Industrial Engineering
Volume 15 Issue 1 Version 1.0 Year 2015
Type: Double Blind Peer Reviewed International Research Journal
Publisher: Global Journals Inc. (USA)
Online ISSN: 2249-4596 & Print ISSN: 0975-5861
ComparisonofAlternativeSolutionsinLinearprogrammingModelingusingtheDualSimplexMethodandDualityMethodfromPrimalProblemEstablishingImplementationthroughtheSimplexMethodology
© 2015. Francisco Zaragoza Huerta & Roció Guzmán López de Lara. This is a research/review paper, distributed under the terms
of the Creative Commons Attribution-Noncommercial 3.0 Unported License http://creativecommons.org/licenses/by-nc/3.0/),
permitting all non commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.
Comparison of Alternative Solutions in Linear
Programming Modeling using the Dual Simplex
Method and Duality Method from Primal
Problem, Establishing Implementation through
the Simplex Methodology
2015
Francisco Zaragoza Huerta α & Roció Guzmán López de Lara σ
Year
Abstract- The Document Research shows one way to visualize Dantzig, who followed John Von Neuman studies Duality
Alternative Solutions for the Same Problem using at the same time, Kantovich Russian scientist working 21
Mathematical Programming Tools Different Solution original, with parallel models seeking optimization problems in
the potentially what efficient search optimization, the idea of economics.
T
value sought.
he problem of solving a linear system of
inequalities goes back at least to Fourier, who was The linear programming problems with more
born after the method of Fourier-Motzkin than two variables are treated through the simplex
elimination. Linear programming is presented as a method is as an iterative process that allows improving
mathematical model developed during World War II to the solution in each step. Mathematical reason for this
plan the costs and returns, to reduce the costs inside improvement I is that the method is to move through the
army and increase losses of the enemy. The linear corners of a polyhedron to a neighbor vertex such
programming models were worked in the decade of the increase or decrease (depending on the context of the
objective function is maximized or minimized), since the
40s during world war secretly by George Bernard
number of vertices having a finite polyhedron solution
Author α: Industrial Engineering Department. ITESM Campus Irapuato. always it will have a feasible solution and extreme point.
e-mail: francisco.zaragoza@itesm.mx
Author σ: Architecture Faculty. FAUMSNH. Simplex Method: Application of the simplex
e-mail: rociogl@hotmail.com method is used when the problem is sufficiently large, it
© 20 15 Global Journals Inc. (US)
Comparison of Alternative Solutions in Linear programming Modeling using the Dual Simplex
Method and Duality Method from Primal Problem, Establishing Implementation through
the Simplex Methodology
is designed for linear programming problems whose variables. The simplex technique involves generating a
parent has the ability to spread the number of non - zero series of solutions in tabular form, called tableaus. By
is small. There are implementations of the simplex inspecting the bottom row of each tableau, one can
method for solving linear programming problems with immediately tell if it represents the optimal solution.
sparse matrices restriction. There were several different Each tableau corresponds to a corner point of the
variants of the simplex method that takes into account feasible solution space. The first tableau corresponds to
the particularities of the various special classes of linear the origin. Subsequent tableaus are developed by
programming problems. shifting to an adjacent corner point in the direction that
yields the highest (smallest) rate of profit (cost). This
II. Methodology process continues as long as a positive (negative) rate
Simplex: a linear-programming algorithm that of profit (cost) exists.
2015
i. Initialization
22 Transform all the constraints to equality by introducing slack, surplus, and artificial variables
as follows:
Global Journal of Researches in Engineering ( G ) Volume XV Issue I Version I
Table I
b) Test for optimality: Case 1: Maximization problem; e) Step 3: Solve for the new Basic Feasible solution by
the current Basic Feasible solution is optimal if every using elementary row operations (multiply or divide
coefficient in the objective function row is a row by a nonzero constant; add or subtract a
nonnegative. multiple of one row to another row) to construct a
Case 2: Minimization problem; the current Basic new simplex tableau, and then return to the
Feasible solution is optimal if every coefficient in the optimality test. The specific elementary row
objective function row is non positive Iteration. operations are:
c) Step 1: determine the entering basic variable by Divide the pivot row by the “pivot number” (the
selecting the variable (automatically a non-basic number in the intersection of the pivot row and pivot
variable) with the most negative value (in case of column).
maximization) or with the most positive (in case of For each other row that has a negative
minimization) in the last row (Z-row). Put a box coefficient in the pivot column, add to this row the
around the column below this variable, and call it product of the absolute value of this coefficient and the
the “pivot column. new pivot row.
d) Step 2: Determine the leaving basic variable by For each other row that has a positive
applying the minimum ratio test as following: coefficient in the pivot column, subtract from this row the
• Pick out each coefficient in the pivot column product of the absolute value of this coefficient and the
that is strictly positive (>0). new pivot row.
• Divide each of these coefficients into the right III. Dual Simplex Method
hand side entry for the same row.
Simplex method starts with a non-optimal but
• Identify the row that has the smallest of these
feasible solution whereas dual simplex method starts
ratios.
with an optimal but infeasible solution.
• The basic variable for that row is the leaving
variable, so replace that variable by the entering Simplex method maintains the feasibility during
variable in the basic variable column of the next successive iterations whereas dual simplex method
simplex tableau. Put a box around this row and maintains the optimality.
call it the “pivot row”. Steps involved in the Dual Simplex Method are:
All the constraints (except those with equality transformed to equality sign by introducing required
(=) sign) are modified to ‘less-than-equal-to’ sign. slack variables.
Constraints with greater-than-equal-to’ sign are b) Modified problem, as in step one, is expressed in
multiplied by -1 through out so that inequality sign gets the form of a simplex tableau. If all the cost
reversed. Finally, all these constraints are transformed to coefficients are positive (i.e., optimality condition is
equality sign by introducing required slack variables. satisfied) and one or more basic variables have
Modified problem, as in step one, is expressed negative values (i.e., non-feasible solution), then
in the form of a simplex tableau. If all the cost dual simplex method is applicable.
coefficients are positive (i.e., optimality condition is c) Selection of exiting variable: The basic variable with
satisfied) and one or more basic variables have negative the highest negative value is the exiting variable. If
values (i.e., non-feasible solution), then dual simplex there are two candidates for exiting variable, any
method is applicable.
2015
one is selected. The row of the selected exiting
Steps involved in the dual simplex method are: variable is marked as pivotal row.
Year
a) All the constraints (except those with equality (=) d) Selection of entering variable: Cost coefficients,
sign) are modified to ‘less-than-equal-to’ sign. corresponding to all the negative elements of the
Constraints with greater-than-equal-to’ sign are pivotal row, are identified. Their ratios are calculated 23
multiplied by -1 through out so that inequality sign after changing the sign of the elements of pivotal
Cost Coefficients
ratio =
− 1 × Elements of pivotal row
Graph 1
The column corresponding to minimum ratio is • If one problem has feasible solutions and a
identified as the pivotal column and associated decision bounded objective function (and so has an optimal
variable is the entering variable. solution), then so does the other problem, so both
e) Pivotal operation: Pivotal operation is exactly same the weak and the strong duality properties are
as in the case of simplex method, considering the applicable
pivotal element as the element at the intersection of
• If one variable has feasible solutions but an
pivotal row and pivotal column.
unbounded objective function (no optimal
f) heck for optimality: If all the basic variables have solutions), then the other problem has no feasible
nonnegative values then the optimum solution is solutions
reached. Otherwise, Steps 3 to 5 are repeated until
the optimum is reached. • If one variable has no feasible solutions, then the
other problem either has no feasible solutions or an
IV. Duality Method unbounded objective function.
• Every linear program has another linear program Relations between Primal and Dual
associated with it: Its ‘dual’. The dual complements • Weak duality property If x is a feasible solution to the
the original linear program, the ‘primal’. The theory primal, and y is a feasible solution to the dual, then;
of duality provides many insights into what is
cx ≤ y*b. Strong duality property If x* is an optimal
happening ‘behind the scenes.
solution to the primal, and y* is an optimal solution
V. Duality Theorem to the dual, then; cx* = y*b
The following are the only possible relationships • Given an LP in general form, called the primal, the
between the primal and dual problems: dual is defined as follows;
Table II
• Complementary Slackness; A pair, x, π, respectively feasible in a primal-dual pair is optimal if and only if:
ui = πi (a’ix - bi) = 0 for all I (1); vj = (cj - π’Aj) xj = 0 for all j (2)
Idea of the Primal and Dual Algorithm; Suppose we have a dual feasible solutionπ. If we can f ind a primal
feasible solution x such that; xj = 0 whenever cj – π’Aj > 0 then: Ai x = bi for all i and hence;
(1) holds because x is a feasible solution to the primal (2) holds because xj = 0 whenever cj – π’Aj = 0. Thus the
complementary slackness relations hold, and hence x and π are optimal solutions to the primal problem (P) and the
dual problem (D).
• Complementary Solutions; At each iteration, the simplex method identifies x, a BFS for the primal, and a
complementary solution y, a BS for the dual (which can be found from the row coefficients under slack
variables). For any primal feasible (but suboptimal) x, its complementary solution y is dual infeasible, with
2015
cx=y*b; For any primal optimal x*, its complementary solution y* is dual optimal, with cx*=y*b Insight;
Year
24
Global Journal of Researches in Engineering ( G ) Volume XV Issue I Version I
Graph 2
Graph 3
Max z = 3x1+5x2
subject to
x1<=4
2x2<=12
3x1+2x2<=18
x1,x2>=0
2015
First iteration
Year
Cj X 3 5 0 0 0
X b X1 X2 H1 H2 H3 25
0 H1 4 1 0 1 0 0
Second Iteration
Cj X 3 5 0 0 0
X b X1 X2 H1 H2 H3
0 H1 4 1 0 1 0 0
5 X2 6 0 1 0 1/2 0
0 H3 6 3 0 0 -1 1
Zj 30 0 5 0 2 1/2 0
Zj-Cj -3 0 0 2 1/2 0
Third Iteration
Cj X 3 5 0 0 0
X b X1 X2 H1 H2 H3
0 H1 2 0 0 1 1/3 - 1/3
5 X2 6 0 1 0 1/2 0
0 H3 2 1 0 0 - 1/3 1/3
Zj 36 3 5 0 1 1/2 1
Zj-Cj 0 0 0 1 1/2 1
Solution Comprobation
Z*=36 X1+H1=4 2(2)=4
X1*=2 2X2+H2=12 2(6)=12
X2*=6 3X1+2X2+H3= 3(2)+2(6)=18
Z*=3(2)+5(6)=36
Min W = 4Y1+12Y2+18Y3
subject to
Y1+3Y3>=3
2Y2+2Y3>=5
Y1,Y2,Y3>=0
First iteration
2015
Cj X 4 12 18 0 0 M M
Year
X b Y1 Y2 Y3 H1 H2 A1 A2
M A1 3 1 0 3 -1 0 1 0
M A2 5 0 2 2 0 -1 0 1
26 Zj 5M M 2M 5M -M -M M M
Zj-Cj M-4 2M-12 5M-18 -M -M 0 0
Global Journal of Researches in Engineering ( G ) Volume XV Issue I Version I
Second Iteration
Cj X 4 12 18 0 0 M M
X b Y1 Y2 Y3 H1 H2 A1 A2
18 Y3 1 1/3 0 1 - 1/3 0 1/3 0
M A2 3 - 2/3 2 0 2/3 -1 - 2/3 1
Zj 3M+18 -2/3M+6 2M 18 2/3M-6 -M -2/3M+6 M
Zj-Cj -2/3M+2 2M-12 0 2/3M-6 -M -5/3M+6 0
Third Iteration
Cj X 4 12 18 0 0 M M
X b Y1 Y2 Y3 H1 H2 A1 A2
18 Y3 1 1/3 0 1 - 1/3 0 1/3 0
12 A2 1 1/2 - 1/3 1 0 1/3 - 1/2 - 1/3 1/2
Zj 36 2 12 18 -2 -6 2 6
Zj-Cj -2 0 0 -2 -6 2-M 6-M
Solution Comprobation
W*=36 Y1+3Y3-H1+A1=3 0+3(1)-0+0=3
Y1*=0 2Y2+2Y3-H2+A2=5 2(3/2)+2(1)-0+0=5
Y2*=3/2
Y3*=1 Z*=3(2)+5(6)=36
© 201
5 Global Journals Inc. (US)
Comparison of Alternative Solutions in Linear programming Modeling using the Dual Simplex
Method and Duality Method from Primal Problem, Establishing Implementation through
the Simplex Methodology
Max Z = -4Y1-12Y2-18Y3
subject to
-Y1-3Y3<=-3
-2Y2-2Y3<=-5
Y1,Y2,Y3>=0
First iteration
2015
Cj X -4 -12 -18 0 0
Year
X b Y1 Y2 Y3 H1 H2
0 H1 -3 -1 0 -3 1 0
0 H2 -5 0 -2 -2 0 1 27
Zj 0 0 0 0 0 0
Second iteration
Cj X -4 -12 -18 0 0
X b Y1 Y2 Y3 H1 H2
0 H1 -3 -1 0 -3 1 0
-12 Y2 2 1/2 0 1 1 0 - 1/2
Zj -30 0 -12 -12 0 6
Zj-Cj 4 0 6 0 6
Third iteration
Cj X -4 -12 -18 0 0
X b Y1 Y2 Y3 H1 H2
-18 Y3 1 1/3 0 1 - 1/3 0
-12 Y2 2 1/2 - 1/3 1 0 1/3 - 1/2
Zj -36 -2 -12 -18 2 6
Zj-Cj 2 0 0 2 6
Solution Comprobation
Z*=36 -Y1-3Y3+H1=-3 -2(0)-3(1)+0=-3
Y1*=0 -2Y2-2Y3+H2=5 -2(3/2)-2(1)-0+0=-5
Y2*=3/2
Y3*=1 Z*=-4(0)-12(3/2)-18(1)=-36
established.
Year