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

Linear Programming Problems

This document discusses key concepts in linear programming problems (LPPs) including: 1) The objective function and optimal objective function value aim to maximize or minimize the value of Z. 2) LPPs have properties of proportionality, additivity, certainty, and divisibility. 3) Slack appears in underused "<=" constraints while surplus appears in overused ">=" constraints. Binding constraints have zero slack/surplus. 4) LPPs can have 0, 1, or infinite optimal solutions depending on the feasible region. The optimal solution(s) will occur at an extreme point, not in the interior region.

Uploaded by

Mayank Apte
Copyright
© © All Rights Reserved
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
72 views

Linear Programming Problems

This document discusses key concepts in linear programming problems (LPPs) including: 1) The objective function and optimal objective function value aim to maximize or minimize the value of Z. 2) LPPs have properties of proportionality, additivity, certainty, and divisibility. 3) Slack appears in underused "<=" constraints while surplus appears in overused ">=" constraints. Binding constraints have zero slack/surplus. 4) LPPs can have 0, 1, or infinite optimal solutions depending on the feasible region. The optimal solution(s) will occur at an extreme point, not in the interior region.

Uploaded by

Mayank Apte
Copyright
© © All Rights Reserved
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 16

Linear Programming Problems

OBJECTIVE FUNCTION VALUE: A VALUE OF Z OBTAINED BY SUBSTITUTING


ANY SOLUTION (FEASIBLE, INFEASIBLE OR OPTIMAL) IN THE OBJECTIVE
FUNCTION
OPTIMAL OBJECTIVE FUNCTION VALUE: THE VALUE OF Z OBTAINED BY
SUBSTITUTING ANY OPTIMAL SOLUTION IN THE OBJECTIVE FUNCTION
4 assumptions/ properties of LPP
PROPORTIONALITY (LINEARITY)
ADDITIVITY
CERTAINTY/ DETERMINISTIC
DIVISIBILITY (WE ALLOW NON-INTEGER VALUES)
PROPORTIONALITY EXPLANATION
IN THE OBJECTIVE FUNCTION, CONTRIBUTION OF A PRODUCT/ACTIVITY
(REPRESENTED BY THE DECISION VARIABLE) TOWARD Z IS DIRECTLY PROPORTIONAL
TO THE QUANTITY OF THE ACTIVITY.
SIMILARLY IN A CONSTRAINT, CONTRIBUTION OF A PRODUCT/ACTIVITY
(REPRESENTED BY THE DECISION VARIABLE) TOWARD THE RHS OF THE CONSTRAINT
IS DIRECTLY PROPORTIONAL TO THE QUANTITY OF THE ACTIVITY.
ANY CHANGE IN THE QUANTITY OF THE PRODUCT/ACTIVITY (REPRESENTED BY THE
DECISION VARIABLE) LEADS TO A PROPORTIONAL CHANGE IN THE Z VALUE
ANY CHANGE IN THE QUANTITY OF THE PRODUCT/ACTIVITY (REPRESENTED BY THE
DECISION VARIABLE) LEADS TO A PROPORTIONAL CHANGE IN RHS OF CONSTRAINT
ALL THIS IS POSSIBLE ONLY BECAUSE OF THE FACT THAT THE DECISION VARIABLE
CORRESPONDING TO THAT ACTIVITY IS OF POWER 1 OR 0.

ADDITIVITY
WE TAKE THE MINE EXAMPLE TO SEE WHAT ADDITIVITY PROPERTY MEANS
THE TOTAL COST (Z) IS THE SUM OF THE INDIVIDUAL MINE COSTS
CHANGE IN Z = SUM OF CHANGE IN INDIVIDUAL MINE COSTS
RHS OF ANY CONSTRAINT (54 TONS) = SUM OF INDIVIDUAL RESOURCE
CONSUMPTIONS (I.E. INDIVIDUAL PRODUCTION OF THE MINES)
CHANGE IN RHS OF A CONSTRAINT = SUM OF INDIVIDUAL CHANGES OF THE
MINES
ALL THESE DEMONSTRATE THE PROPERTY OF ADDITIVITY
WE CAN ACTUALLY ADD THE INDIVIDUAL VALUES TO GET THE FINAL VALUE
SINCE THE INDIVIDUAL VALUES ARE INDEPENDENT
DIVISIBILITY: WE ALLOW NON-INTEGER VALUES
CERTAINTY: WE ARE CERTAIN ABOUT THE VARIOUS COSTS, VARIOUS
CONTRIBUTIONS, VARIOUS PROFITS, THE PROTEIN AND FIBER CONTENT %
PRESENT IN CORN AND SOYA, THE QUANTITY OF RAW MATERIAL THAT INTERIOR
AND EXTERIOR PAINTS TAKE, THE TIME THAT EACH CHOCOLATE VARIETY TAKES
TO BE PRODUCED, ETC.
WE DON'T ATTACH ANY PROBABILITY
Slack & Surplus
SLACK APPEARS IN AN UNDERUSED CONSTRAINT (USED LESS THAN ITS MAXIMUM ALLOWED LEVEL). IT
RELATES TO <= CONSTRAINTS.
TAKE THE CEREAL PROBLEM AS AN E.G.
0.02X1 +0.06X2 <= 40 WAS A CONSTRAINT THERE
THE OPTIMUM SOLUTION WAS 470.588KG OF CORN AND 329.412KGOF SOYA. BY PRODUCING THESE
QUANTITIES, THE FIBRE CONTENT IN THE 800KG OF DISH ENDED UP BEING ONLY 29.176KG.
NOW THIS CONSTRAINT SAYS WE CAN HAVE UPTO 40KG, BUT WE ENDED UP WITH LESS THAN THAT. WE
DID NOT MAKE USE OF THIS CONSTRAINT TO ITS FULLEST.
HERE SLACK = 40 29.176 = 10.824KG. HAD THE OTHER CONSTRAINTS BEEN LESS RESTRICTING, WE
COULD HAVE USED MORE OF THE FIBER ALLOWANCE (CONSTRAINT)
THOUGH THE FIBER CONSTRAINT WAS PRESENT, IT DID NOT REALLY ACTIVELY LIMIT OR BIND US HERE IN
THIS PROBLEM. SO IT IS CALLED A NON-BINDING CONSTRAINT.
SO, ANY CONSTRAINT WITH SLACK > 0 IS A NON-BINDING CONSTRAINT
IN PROBLEM AB1, THE OPTIMAL / OPTIMUM SOLUTION IS TO MAKE 85.714 UNITS OF 30G CHOCOLATES
AND 114.286 UNITS OF 50G CHOCOLATES. SO WE SEE THAT THE CONSTRAINT X1+X2 <= 200
CONSTRAINT, IS A BINDING CONSTRAINT SINCE IT WAS NOT ONLY PRESENT BUT ALSO ACTIVELY LIMITED
THE NUMBER OF UNITS OF CHOCOLATES WE COULD PRODUCE.
THE TERM SLACK DOES NOT APPLY TO PROTEIN CONSTRAINT AND THE
QUANTITY CONSTRAINT AS THEY ARE >= CONSTRAINTS. WE USE THE
TERM 'SURPLUS' FOR >= CONSTRAINTS
SURPLUS APPEARS WHEN A CONSTRAINT IS USED MORE THAN ITS
MINIMUM ALLOWED LEVEL.
IN THE CEREAL PROBLEM, BOTH , BOTH QUANTITY AND PROTEIN
CONSTRAINTS HAD A SURPLUS OF 0. FOR E.G. THE MINIMUM ALLOWED
LEVEL OF PRODUCTION OF THE DISH WAS 800KG, AND ACCORDING TO THE
OPTIMAL SOLUTION, WE ENDED UP PRODUCING EXACTLY 800KG. THERE
WAS NO SURPLUS. SURPLUS = 0 FOR THESE CONSTRAINTS. SO THEY
BOTH ARE BINDING CONSTRAINTS.
WHEN YOU COME ACROSS A >= CONSTRAINT WITH SURPLUS >0, YOU
CAN CALL IT A NON-BINDING CONSTRAINT. ACTUALLY, YOU HAVE TO CALL IT
A NON-BINDING CONSTRAINT.

FOR A >= CONSTRAINT, IT IS A BINDING CONSTRAINT WHEN SURPLUS = 0


AT OPTIMAL POINT . ELSE, IT IS A NON-BINDING CONSTRAINT
FOR A <= CONSTRAINT, IT IS A BINDING CONSTRAINT WHEN SLACK = 0 AT
OPTIMAL POINT . ELSE, IT IS A NON-BINDING CONSTRAINT
WE KNOW THAT LPPS CAN HAVE MULTIPLE SOLUTIONS
IF AN LPP HAS MULTIPLE SOLUTIONS, THEN SURELY IT HAS INFINITE SET OF
SOLUTIONS
SO LPPS HAVE EITHER 0 OR 1 OR INFINITE SOLUTIONS
IN OTHER WORDS, IF EXTREME POINT METHOD GIVES 2 OPTIMAL SOLUTIONS
FOR AN LPP, IT MEANS THAT THAT LPP HAS INFINITE SOLUTIONS

INFEASIBLE SOLUTION, MULTIPLE OPTIMAL, UNBOUNDED FEASIBLE REGION AND
ABSENCE OF FEASIBLE REGION HAVE BEEN DISCUSSED IN CLASS AND ALSO
EXPLAINED IN BOOK IN A SIMPLE MANNER
A WORD ON REDUNDANT AND BINDING CONSTRAINT:
IT IS POSSIBLE TO HAVE A CONSTRAINT THAT IS NON-REDUNDANT AND NON-
BINDING
E.G. 4X + 10Y <= 100 ASSEMBLY CONSTRAINT SHOWN IN FIG 3-5 IN TEXT
BOOK
TO SEE WHICH SIDE OF THE GRAPH SATISFIES A PARTICULAR CONSTRAINT:
PICK ON POINT ON EACH SIDE
SUBSTITUTE BOTH POINTS IN THE CONSTRAINT
THE SIDE THAT HAS THE COORDINATE THAT SATISFIES THE CONSTRAINT IS THE
SIDE THAT SATISFIES THE CONSTRAINT. WE 'ELIMINATE' THE OTHER SIDE.
Extreme Point Theorem:
If an LPP has a feasible region, then the optimal
solution lies at one or more of the extreme points of the LPP.
It never lies in the interior region.

You might also like