Prototype for linear programming is done. We only need more examples …

…of different CPLEX formats, specifically MOLP, big M, Q and ILP methods.
psmgeelen · Feb 10, 2018 · 071d1ee · 071d1ee
1 parent b8af08e
commit 071d1ee
Show file tree

Hide file tree

Showing 9 changed files with 167 additions and 127 deletions.
diff --git a/figures/advanced_encoding.png b/figures/advanced_encoding.png
diff --git a/figures/cplexformat.png b/figures/cplexformat.png
diff --git a/figures/openfile.png b/figures/openfile.png
diff --git a/figures/savefile.png b/figures/savefile.png
diff --git a/glpkAPI.pdf b/glpkAPI.pdf
diff --git a/linearprogramming.Rmd b/linearprogramming.Rmd
@@ -4,60 +4,62 @@ output: html_notebook
 ---
 
 This notebook with give you a start in Linear Programming in R. Arseny and I have been working many hours to develop a workflow by which you can truly write down the functions and constraints, as you will for the exam. It is our belief that this will prepare more adequately for the exam and we hope to improve the overall "feel" of the course. 
-
+<p></p><p></p>
 <h3> CPLEX </h3>
 In order to "just write down" the functions, we need to work with a particular format which is the "CPLEX". Its based on a more complicated engine that is used by big companies like IBM. R, of course, also provides a free alternative. The overal format is as follows.
 
-<p>--Start-File--</p>
-<p>Maximize</p>
-<p>&nbsp; obj: x1 + 2 x2 + 3 x3 + x4</p>
-<p>Subject To</p>
-<p>&nbsp; c1: - x1 + x2 + x3 + 10 x4 <= 20</p>
-<p>&nbsp; c2: x1 - 3 x2 + x3 <= 30</p>
-<p>&nbsp; c3: x2 - 3.5 x4 = 0</p>
-<p>Bounds</p>
-<p>&nbsp; 0 <= x1 <= 40</p>
-<p>&nbsp; 2 <= x4 <= 3</p>
-<p>General</p>
-<p>&nbsp; x4</p>
-<p>End</p>
-<p>--End-File--</p>
+![](figures/cplexformat.png)
 
 This link provides a concrete guide for the use of this format:
 http://lpsolve.sourceforge.net/5.0/CPLEX-format.htm
-Note that this package also contains a free book from Fernandez that provide excellent examples for this method. The first template is provided on page 34. 
+Note that this package also contains a free book from Fernandez that provide excellent examples for this method. The first template is provided on page 34, which is excerpted above.
+
+<h4> Workflow CPLEX </h4>
+
+The CPLEX format is based on ASCII coding. This coding is relates to how a file is stored. The workflow of creating a CPLEX file is a follows. You can create a new file in RStudio by clicking on the new file button. This will open a menu, select txt file. 
+![](figures/openfile.png)
+
+When everything has been entered in the correct CPLEX format, you can save the file by going on the menubar and clicking on the file button. 
+![](figures/savefile.png)
+
+From there you select the "Save file with encoding..." option. This will open another small screen with a few options of encoding, including ASCII. Select ASCII. After that you can select a path name for the file. Please note that if you save it as linear.programm.txt, you need to refer to it as such. In the examples above we used no extension (.txt, .doc, .docx, etc..) and therefore do not refer to one either in our commands.
+
 
 <h3> Required Packages </h3>
 There are 3 packages required to run these linear programms
 
 <ul>
-<li>lpSolve</li>
-<li>lpSolveAPI</li>
+<li>glpkAPI</li>
 <li>Rglpk</il>
 </ul>
 
 So lets install those by executing the code below:
 ```{r}
-install.packages("lpSolve")
-install.packages("lpSolveAPI")
+install.packages("glpkAPI")
 install.packages("Rglpk")
 ```
 
 Having installed the requried packages, we can run the generic code for linear programming which is:
 ```{r}
-library(lpSolveAPI)
+library(glpkAPI)
 library(Rglpk)
 
-x <- Rglpk_read_file("example_lp", type = "CPLEX_LP", ignore_first_row = FALSE, verbose = FALSE)
+problem <- initProbGLPK(ptrtype = "glpk_prob")
 
-solve <- Rglpk_solve_LP(x$objective, x$constraints[[1]], x$constraints[[2]], x$constraints[[3]], x$bounds, x$types, x$maximum)
-```
+readLPGLPK(problem, "example_lp")
+
+solveSimplexGLPK(problem)
+
+printRangesGLPK(problem, fname = "report.txt")
 
-Here the CPLEX object is 1) imported into R, 2) An optimum is calculated and 3) the results are retrieved. 
+file.show("report.txt")
+
+```
 
+In order to undestand the code above, I want to elaborate on what an API is. An API is an "Application Programming Interface". More simply put, its a way of communicating to other software. In order to use GLPK (GNU Linear Programming Kit), we need to create an object within this package. To do this we use the initProbGLPK, which initiates an empty object within the GLPK package. Note that this object is connected to R through an object in R, called "problem". More simply put: the "problem" object in R is the connection to the "glpk_prob" object in GLPK. These objects are in effect mirrors and connections. The read, solve and print commands dont require an object within R, since they are executed in the GLPK package.  Due to the fact that the real calculations are not run in R, the sensitivty report cannot be displayed in the terminal either. For this reason a file.show command has been used, to open the file when the calculations are done. 
 
-<h3> Sensitivity Report </h3>
+<h3> Examples </h3>
 
-Tornado graph would be kewl too
+There are alot of concrete examples of how to write Linear Programs in CPLEX in the book that is also added to this zip file. The book is called "Linear Programming.pdf" There 
 
-To do: Sensitivity report, CPLEX formats for MOLP, ILP and big M methods, we need to test and formulate the ASCII issue
+To do:  Are there a couple of formats that we can steal from the book to cover the MOLP, ILP and LP principles? Maybe even Big M and stuff. 
diff --git a/linearprogramming.nb.html b/linearprogramming.nb.html
diff --git a/report.txt b/report.txt
@@ -0,0 +1,38 @@
+GLPK 4.47 - SENSITIVITY ANALYSIS REPORT                                                                         Page   1
+
+Problem:    
+Objective:  obj = 125.2083333 (MAXimum)
+
+   No. Row name     St      Activity         Slack   Lower bound       Activity      Obj coef  Obj value at Limiting
+                                          Marginal   Upper bound          range         range   break point variable
+------ ------------ -- ------------- ------------- -------------  ------------- ------------- ------------- ------------
+     1 c1           NU      20.00000        .               -Inf       -2.00000      -1.64583      89.00000 x4
+                                           1.64583      20.00000       22.00000          +Inf     128.50000 x4
+
+     2 c2           NU      30.00000        .               -Inf       28.00000      -1.35417     122.50000 x4
+                                           1.35417      30.00000       52.00000          +Inf     155.00000 x4
+
+     3 c3           NS        .             .             .             -.50000          -Inf     123.00000 x4
+                                           4.41667        .             5.50000          +Inf     149.50000 x4
+
+GLPK 4.47 - SENSITIVITY ANALYSIS REPORT                                                                         Page   2
+
+Problem:    
+Objective:  obj = 125.2083333 (MAXimum)
+
+   No. Column name  St      Activity      Obj coef   Lower bound       Activity      Obj coef  Obj value at Limiting
+                                          Marginal   Upper bound          range         range   break point variable
+------ ------------ -- ------------- ------------- -------------  ------------- ------------- ------------- ------------
+     1 x1           NU      40.00000       1.00000        .            29.00000       -.29167     111.00000 x4
+                                           1.29167      40.00000       41.00000          +Inf     126.50000 x4
+
+     2 x2           BS      10.20833       2.00000        .             7.00000      -2.42857      80.00000 x1
+                                            .               +Inf       10.50000      11.28571     220.00000 c2
+
+     3 x3           BS      20.62500       3.00000        .            19.50000        .59259      75.55556 c2
+                                            .               +Inf       22.00000      13.33333     338.33333 x1
+
+     4 x4           BS       2.91667       1.00000       2.00000         .          -14.50000      80.00000 x1
+                                            .            3.00000        4.44444      33.50000     220.00000 c2
+
+End of report
diff --git a/sar.txt b/sar.txt
@@ -0,0 +1,38 @@
+GLPK 4.47 - SENSITIVITY ANALYSIS REPORT                                                                         Page   1
+
+Problem:    
+Objective:  obj = 125.2083333 (MAXimum)
+
+   No. Row name     St      Activity         Slack   Lower bound       Activity      Obj coef  Obj value at Limiting
+                                          Marginal   Upper bound          range         range   break point variable
+------ ------------ -- ------------- ------------- -------------  ------------- ------------- ------------- ------------
+     1 c1           NU      20.00000        .               -Inf       -2.00000      -1.64583      89.00000 x4
+                                           1.64583      20.00000       22.00000          +Inf     128.50000 x4
+
+     2 c2           NU      30.00000        .               -Inf       28.00000      -1.35417     122.50000 x4
+                                           1.35417      30.00000       52.00000          +Inf     155.00000 x4
+
+     3 c3           NS        .             .             .             -.50000          -Inf     123.00000 x4
+                                           4.41667        .             5.50000          +Inf     149.50000 x4
+
+GLPK 4.47 - SENSITIVITY ANALYSIS REPORT                                                                         Page   2
+
+Problem:    
+Objective:  obj = 125.2083333 (MAXimum)
+
+   No. Column name  St      Activity      Obj coef   Lower bound       Activity      Obj coef  Obj value at Limiting
+                                          Marginal   Upper bound          range         range   break point variable
+------ ------------ -- ------------- ------------- -------------  ------------- ------------- ------------- ------------
+     1 x1           NU      40.00000       1.00000        .            29.00000       -.29167     111.00000 x4
+                                           1.29167      40.00000       41.00000          +Inf     126.50000 x4
+
+     2 x2           BS      10.20833       2.00000        .             7.00000      -2.42857      80.00000 x1
+                                            .               +Inf       10.50000      11.28571     220.00000 c2
+
+     3 x3           BS      20.62500       3.00000        .            19.50000        .59259      75.55556 c2
+                                            .               +Inf       22.00000      13.33333     338.33333 x1
+
+     4 x4           BS       2.91667       1.00000       2.00000         .          -14.50000      80.00000 x1
+                                            .            3.00000        4.44444      33.50000     220.00000 c2
+
+End of report