Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                
Scribd
Upload
23 views

Linear Programming

Linear programming (LP) is an optimization technique for finding the optimum value of a linear objective function subject to linear constraints. LP involves decision variables, constant parameters, and assumptions of proportionality, additivity, divisibility and certainty. The simplex method is commonly used to solve LP problems by moving between corner point solutions until an optimal solution is found.

Uploaded by

simonquanzihan
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
23 views

Linear Programming

Linear programming (LP) is an optimization technique for finding the optimum value of a linear objective function subject to linear constraints. LP involves decision variables, constant parameters, and assumptions of proportionality, additivity, divisibility and certainty. The simplex method is commonly used to solve LP problems by moving between corner point solutions until an optimal solution is found.

Uploaded by

simonquanzihan
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 6

LINEAR PROGRAMMING (LP)

Definition
LP is an optimization technique for finding the optimum
(maximum or minimum) of a linear objective function
subject to linear and nonnegative constraints.

Forms of LP model
Our standard form: Max Z = S cj xj
S aij xj £ bi, for i=1, ..., m
xj ³ 0, for j= 1, …, n.

Z = objective function;
xj = decision variables;
aij, bi, cj = constant parameters.

Some transformation operations: minimization, greater-than-


or-equal-to constraints, free variables, and equality
constraints.

Assumptions
Proportionality: cjxj and aijxj.
Additivity: linear sum of contributions.
Divisibility: noninteger decision variables.
Certainty: deterministic variables and parameters.

Terminology
Feasible solution = a solution for which all the constraints are
satisfied.
Feasible region = the collection of all feasible solutions
Optimal solution = a feasible solution that has the most
favorable (maximum or minimum) value of the objective
function.
Basic solution = an augmented corner-point solution.
Basic feasible solution = an augmented corner-point feasible
solution.

Graphical solution method


Two-variable problems.

Properties of corner-point solutions


Property 1. (Fundamental Theorem of LP)
(a) If there is a feasible solution, there is a basic (corner-
point) feasible solution;
(b) If there is an optimal feasible solution, there is an optimal
basic feasible solution.

Þ The search for an optimal solution is simplified since only


basic feasible solutions need be considered.

Property 2. There are only a finite number of corner-point feasible


(CPF) solutions. The upper bound on the number of CPF
solutions is
"+$ (" + $)!
! %=
$ "! $!

Þ A finite search technique is possible but inefficient Þ the


simplex method is more efficient.

Property 3. If a corner-point feasible solution has no adjacent


corner-point feasible solutions that are better, then there are no
better corner-point feasible solutions anywhere; i.e., it is optimal.

Þ The region of feasible solutions is convex.

The simplex method:


The algebra of the simplex method.
The simplex method in tabular form.
Interpretation of the simplex tableau
• optimal solution (values of the optimum slack variables
are ignored);
• status of resources based on values of slack variables;
• per-unit worth of resources (shadow prices).

Artificial variable methods


• for constraints of the ³ type; and
• for constraints of the = type.
Big-M method
Two-phase method.

Special cases
• Degeneracy.
• Infeasible solution: Incompatible constraints.
Unbounded solution.
• Multiple optimal solutions.

Other Methods for LP ( Khatchian, Karmarkar,...).


The Algebra of the Simplex Method

Step 1 - Transform the inequalities into equations using slack (or


surplus) variables.

Step 2 - Select a corner-point solution as a starting basic feasible


solution [usually, we start at (0,0)], and compute the
corresponding value of the objective function.

Step 3 - Move to an adjacent corner-point feasible solution that


would increase the objective function at the fastest rate, and
compute the corresponding value of the objective function.
(Repeat this step as often as needed.)

Step 4 - Conclude that the current corner-point feasible solution


is optimal when none of its adjacent corner-point feasible
solutions could give a better objective function value.

Criterion for Selecting the Entering Basic Variable:

"Select the nonbasic variable that, when increased, would


increase the objective function at the fastest rate.''

Criterion for Selecting the Leaving Basic Variable:

"Select the basic variable that reaches zero first as the chosen
entering basic variable is increased.''
The Simplex Method

Step 1. Introduce slack variables (to transform the inequalities


into equations). Select the slack variables as basic
variables and the original variables as nonbasic ones.

Step 2. - If there are non-basic variables with (cj-zj) >0, select


the variable xj with the largest value of (cj-zj) as the new
basic variable xe. Go to Step 3.
- If all nonbasic variables have (cj-zj) £ 0, an optimal
solution is obtained.

Step 3. Compute the ratio between bi (or xi) and the positive
coefficients aie of xe. The smallest ratio determine the
leaving variable xl.

Step 4. Perform a change of basis (by Gauss-Jordan


elimination method) in order to find the values of all
new basic variables and the objective function. Go to
Step 2.

You might also like