Difference between revisions of "ApCoCoA-1:GLPK.LPSolve"

From ApCoCoAWiki
m (added version info)
 
(47 intermediate revisions by 9 users not shown)
Line 1: Line 1:
 +
{{Version|1|[[Package glpk/GLPK.LPSolve]]}}
 
<command>
 
<command>
<title>LPSolve</title>
+
<title>GLPK.LPSolve</title>
<short_description>solve linear programms</short_description>
+
<short_description>Solving linear programmes.</short_description>
 
<syntax>
 
<syntax>
GLPK.LPSolv(Objective_Function:POLYNOM, EQ_Polynomials:LIST, LE_Polynomials:LIST, GE_Polynomials:LIST, Bounds:LIST, Methode:STRING, MinMax:STRING)
+
GLPK.LPSolve(Objective_f:POLY, EQ_Poly:LIST, LE_Poly:LIST, GE_Poly:LIST, Bounds:LIST, Method:STRING, MinMax:STRING):LIST
 
</syntax>
 
</syntax>
 
<description>
 
<description>
 +
<em>Please note:</em> The function(s) explained on this page is/are using the <em>ApCoCoAServer</em>. You will have to start the ApCoCoAServer in order to use it/them.
  
The basic idea behind this package is to make the linear optimization program GLPK usable in/with ApCoCoA.
+
<itemize>
The package GLPK contains various functions that let you make use of the GLPK library, rather the stand-alone LP/MIP Solver glpsol.
+
<item>@param <em>Objective_f</em>: A linear polynomial which is equivalent to the linear objective function.</item>
 +
<item>@param <em>EQ_Poly</em>: List of linear polynomials, which are equivalent to the equality-part in the list of conditions.</item>
 +
<item>@param <em>LE_Poly</em>: List of linear polynomials, which are equivalent to the lower or equal-part in the list of conditions.</item>
 +
<item>@param <em>GE_Poly</em>: List of linear polynomials, which are equivalent to the greater or equal-part in the list of conditions.</item>
 +
<item>@param <em>Bounds</em>: List of lists with two elements. Each List contains the lower and upper bounds for each variable. You can choose between INT or RAT for the type of each bound, if you type in a (empty) string, then it means minus infinity (first place) or plus infinity (second place).</item>
 +
<item>@param <em>Method</em>: You can choose between the interior-point-method ("InterP") or the simplex-algorithm ("Simplex"). Usually you should use the simplex-algorithm.</item>
 +
<item>@param <em>MinMax</em>:  Minimization ("Min") or maximization ("Max"), that's the question.</item>
 +
<item>@return List of linear polynomials, the zeros of the polynomials are the points where the optimal value of the objective function is achieved</item>
 +
</itemize>
  
<em>Objective_Function</em>: A linear polynomial which is equivalent to the linear objective Function.
 
 
<em>EQ_Polynomials</em>: List of linear polynomials, which are equivalent to the equality-part in the list of conditions.
 
 
<em>LE_Polynomials</em>: List of linear polynomials, which are equivalent to the lower or equal-part in the list of conditions.
 
 
<em>GE_Polynomials</em>: List of linear polynomials, which are equivalent to the greater or equal-part in the list of conditions.
 
 
<em>Bounds</em>: List of lists with two elements. Each List contains the lower and upper bounds for each variable.
 
 
<em>Methode</em>: You can choose between the inner-point-method (InnerP) or the simplex-algorithm (Simplex).
 
 
<em>MinMax</em>:  Minimization (Min) or maximization (Max), that's the question.
 
 
 
When you execute the package, you had to specify the path of the LP solver glpsol, i.e. GlpsolPath:="/usr/bin/";
 
and the working path, i.e. WorkingPath:="/home/user/";
 
 
 
First we want to discuss a rather easy example.
 
 
<example>
 
<example>
We want to maximize the Function y = - 1/2x,  
+
-- We want to maximize the Function y = - 1/2x,  
with the two conditions y ≤ 6 - 3/4x and y ≥ 1 - x and the bounds 0 ≤ x ≤ 6 and 1/3 ≤ y ≤ 4.
+
-- with the two conditions y ≤ 6 - 3/4x and y ≥ 1 - x and the bounds 0 ≤ x ≤ 6 and 1/3 ≤ y ≤ 4.
  
We prename the input of GLPK.LPSol-function.
+
-- We prename the input of GLPK.LPSolve-function.
 +
Use S::=QQ[x,y];
 
OF := 1/2x + y;
 
OF := 1/2x + y;
LE := 3/4x + y - 6;
+
LE := [3/4x + y - 6];
GE := x + y - 1;
+
GE := [x + y - 1];
 
Bounds:=[[0,6], [1/3,4]];
 
Bounds:=[[0,6], [1/3,4]];
  
Then we compute the solution with
+
-- Then we compute the solution with
Use S::=Q[x,y];
+
GLPK.LPSolve(OF, [], LE, GE, Bounds, "Simplex", "Max");
$GLPK.LPSol(OF, LE, GE, Bounds, Simplex, Max);
 
  
And we achieve x = 8/3 and y = 4.
+
-- And we achieve:
 +
-------------------------------------
 +
Solution Status: OPTIMAL
 +
Value of objective function: 5333333333/1000000000
 +
[x - 266667/100000, y - 4]
 
</example>
 
</example>
 
[[Image:LinOpt.jpg]]
 
 
 
 
<example>
 
''Linear programming example 1996 MBA exam''
 
 
A cargo plane has three compartments for storing cargo: front, centre and rear.
 
These compartments have the following limits on both weight and space:
 
 
Compartment  Weight capacity (tonnes)  Space capacity (cubic metres)
 
Front        10                        6800
 
Centre        16                        8700
 
Rear          8                          5300
 
 
Furthermore, the weight of the cargo in the respective compartments must be the same
 
proportion of that compartment's weight capacity to maintain the balance of the plane.
 
 
The following four cargoes are available for shipment on the next flight:
 
 
Cargo  Weight (tonnes)  Volume (cubic metres/tonne)  Profit (£/tonne)
 
C1      18                480                          310
 
C2      15                650                          380
 
C3      23                580                          350
 
C4      12                390                          285
 
 
Any proportion of these cargoes can be accepted. The objective is to determine how much
 
(if any) of each cargo C1, C2, C3 and C4 should be accepted and how to distribute each
 
among the compartments so that the total profit for the flight is maximised.
 
 
 
To solve this problem we had to compose a linear program.
 
 
Variables
 
 
We need to decide how much of each of the four cargoes to put in each of the three compartments.
 
Hence let:
 
 
xij be the number of tonnes of cargo i (i=1,2,3,4 for C1, C2, C3 and C4 respectively) that is put into
 
compartment j (j=1 for Front, j=2 for Centre and j=3 for Rear) where xij >=0 i=1,2,3,4; j=1,2,3
 
 
Note here that we are explicitly told we can split the cargoes into any proportions (fractions) that we like.
 
 
Constraints
 
 
* cannot pack more of each of the four cargoes than we have available
 
 
x11 + x12 + x13 ≤ 18
 
x21 + x22 + x23 ≤ 15
 
x31 + x32 + x33 ≤ 23
 
x41 + x42 + x43 ≤ 12
 
 
* the weight capacity of each compartment must be respected
 
 
x11 + x21 + x31 + x41 ≤ 10
 
x12 + x22 + x32 + x42 ≤ 16
 
x13 + x23 + x33 + x43 ≤ 8
 
 
* the volume (space) capacity of each compartment must be respected
 
 
480x11 + 650x21 + 580x31 + 390x41 ≤ 6800
 
480x12 + 650x22 + 580x32 + 390x42 ≤ 8700
 
480x13 + 650x23 + 580x33 + 390x43 ≤ 5300
 
 
* the weight of the cargo in the respective compartments must be the same proportion
 
  of that compartment's weight capacity to maintain the balance of the plane
 
 
[x11 + x21 + x31 + x41]/10 = [x12 + x22 + x32 + x42]/16 = [x13 + x23 + x33 + x43]/8
 
 
Objective
 
 
The objective is to maximise total profit, i.e.
 
 
maximise 310[x11+ x12+x13] + 380[x21+ x22+x23] + 350[x31+ x32+x33] + 285[x41+ x42+x43]
 
 
The basic assumptions are:
 
 
* that each cargo can be split into whatever proportions/fractions we desire
 
* that each cargo can be split between two or more compartments if we so desire
 
* that the cargo can be packed into each compartment (for example if the cargo was spherical it would
 
  not be possible to pack a compartment to volume capacity, some free space is inevitable in sphere packing)
 
* all the data/numbers given are accurate
 
 
The advantages of using a software package to solve the above linear program, rather than a judgemental
 
approach are:
 
 
* actually maximise profit, rather than just believing that our judgemental solution maximises profit
 
  (we may have bad judgement, even if we have an MBA!)
 
* makes the cargo loading the decision one that we can solve in a routine operational manner on a computer,
 
  rather than having to exercise judgement each and every time we want to solve it
 
* problems that can be appropriately formulated as linear programs are almost always better solved by computers
 
  than by people
 
* can perform sensitivity analysis very easily using a computer
 
 
</example>
 
 
  
 
</description>
 
</description>
 
 
<types>
 
<types>
   <type>cocoaserver</type>
+
   <type>apcocoaserver</type>
 +
  <type>linear_programs</type>
 
</types>
 
</types>
 +
<see>ApCoCoA-1:Latte.Minimize|Latte.Minimize</see>
 +
<see>ApCoCoA-1:Latte.Maximize|Latte.Maximize</see>
 +
 
<key>lpsolve</key>
 
<key>lpsolve</key>
<key>solve linear programm</key>
+
<key>solve linear program</key>
 
<key>solve lp</key>
 
<key>solve lp</key>
<key>jbrandt</key>
+
<key>GLPK.LPSolve</key>
<key>skuehling</key>
+
<wiki-category>ApCoCoA-1:Package_glpk</wiki-category>
<wiki-category>Package_GLPK</wiki-category>
 
 
</command>
 
</command>

Latest revision as of 15:12, 1 November 2020

This article is about a function from ApCoCoA-1. If you are looking for the ApCoCoA-2 version of it, see Package glpk/GLPK.LPSolve.

GLPK.LPSolve

Solving linear programmes.

Syntax

GLPK.LPSolve(Objective_f:POLY, EQ_Poly:LIST, LE_Poly:LIST, GE_Poly:LIST, Bounds:LIST, Method:STRING, MinMax:STRING):LIST

Description

Please note: The function(s) explained on this page is/are using the ApCoCoAServer. You will have to start the ApCoCoAServer in order to use it/them.

  • @param Objective_f: A linear polynomial which is equivalent to the linear objective function.

  • @param EQ_Poly: List of linear polynomials, which are equivalent to the equality-part in the list of conditions.

  • @param LE_Poly: List of linear polynomials, which are equivalent to the lower or equal-part in the list of conditions.

  • @param GE_Poly: List of linear polynomials, which are equivalent to the greater or equal-part in the list of conditions.

  • @param Bounds: List of lists with two elements. Each List contains the lower and upper bounds for each variable. You can choose between INT or RAT for the type of each bound, if you type in a (empty) string, then it means minus infinity (first place) or plus infinity (second place).

  • @param Method: You can choose between the interior-point-method ("InterP") or the simplex-algorithm ("Simplex"). Usually you should use the simplex-algorithm.

  • @param MinMax: Minimization ("Min") or maximization ("Max"), that's the question.

  • @return List of linear polynomials, the zeros of the polynomials are the points where the optimal value of the objective function is achieved

Example

-- We want to maximize the Function y = - 1/2x, 
-- with the two conditions y ≤ 6 - 3/4x and y ≥ 1 - x and the bounds 0 ≤ x ≤ 6 and 1/3 ≤ y ≤ 4.

-- We prename the input of GLPK.LPSolve-function.
Use S::=QQ[x,y];
OF := 1/2x + y;
LE := [3/4x + y - 6];
GE := [x + y - 1];
Bounds:=[[0,6], [1/3,4]];

-- Then we compute the solution with
GLPK.LPSolve(OF, [], LE, GE, Bounds, "Simplex", "Max");

-- And we achieve:
------------------------------------- 
Solution Status: OPTIMAL
Value of objective function: 5333333333/1000000000
[x - 266667/100000, y - 4]


Latte.Minimize

Latte.Maximize