AMPL Book Examples#

AMPL: A Modeling Language for Mathematical Programming
by Robert Fourer, David M. Gay, and Brian W. Kernighan
Second edition
517 + xxi pp., ISBN 0-534-38809-4
The AMPL Book PDF

Use this page to download all the model, data, and script files that appear as examples in the AMPL book (second edition). You can download everything in one file:

zipfile with line endings characteristic of Windows
gzipped tar file with line endings characteristic of Unix

Or download individual files from our listings below, by filename or by figure number.

Book examples by filename#

Model file	Figure	Data file	Figure	Script file	Figure
		assign.dat	3-2
blend.mod	2-4
cut.mod	14-2	cut.dat	14-4	cut.run	14-3
cut2.mod	14-6			cut2.run
diet.mod	2-1	diet.dat	2-2
		diet2.dat	2-3	diet.run	14-1
		diet2a.dat	2-3
dietu.mod	5-1	dietu.dat	5-2
econ.mod	19-3	econ.dat	19-2
econmin.mod	19-1
econ2.mod	19-4	econ2.dat
econ2min.mod
econnl.mod	19-5	econnl.dat
iocol1.mod	16-2
iocol2.mod	16-3
iorow.mod	16-1
minmax.mod	17-9
multi.mod	4-1	multi.dat	4-2
multic.mod	6-5
multmip1.mod	20-1a	multmip1.dat	20-1b
multmip2.mod		multmip2.dat
multmip3.mod	20-2a	multmip3.dat	20-2b
net1.mod	15-2a	net1.dat	15-2b
net1node.mod	15-10
net2.mod	15-3a	net2.dat	15-3b
net3.mod	15-4a	net3.dat	15-4b
net3node.mod	15-11
netasgn.mod	15-9
netfeeds.mod	15-14
netmax.mod	15-6
netmax3.mod	15-12
netmcol.mod	16-6
netmulti.mod	15-13
netshort.mod	15-7
netthru.mod	15-15
nltrans.mod	18-4	nltrans.dat	18-5
nltransb.mod
nltransc.mod
nltransd.mod
prod0.mod
prod.mod	1-2	prod.dat	1-3
sched.mod	16-4	sched.dat	16-5
steel.mod	1-4a	steel.dat	1-4b
		steel2.dat
steel3.mod	1-5a	steel3.dat	1-5b
steel4.mod	1-6a	steel4.dat	1-6b
steel4r.mod		steel4r.dat
steelP.mod	4-6	steelP.dat	4-7
steelpl1.mod	17-5a	steelpl1.dat	17-5b
steelpl2.mod		steelpl2.dat
steelpl3.mod
steelpl4.mod		steelpl4.dat
steelpl5.mod		steelpl5.dat
steelT0.mod	4-3
				steelT.sa1
				steelT.sa2
steelT.mod	4-4	steelT.dat	4-5	steelT.sa3	13-1
				steelT.sa4	13-3
				steelT.sa5	13-4
				steelT.sa7	13-5
				steelT.tab1	13-2
steelT2.mod	5-3	steelT2.dat	5-4
steelT3.mod	6-3	steelT3.dat	6-4
transp.mod	3-1a	transp.dat	3-1b
transp2.mod	6-1
transp3.mod	6-2a	transp3.dat	6-2b

Book examples by figure number#

Figure	File
1-2	prod.mod
1-3	prod.dat
1-4a	steel.mod
1-4b	steel.dat
1-5a	steel3.mod
1-5b	steel3.dat
1-6a	steel4.mod
1-6b	steel4.dat
2-1	diet.mod
2-2	diet.dat
2-3	diet2.dat
2-4	blend.mod
3-1a	transp.mod
3-1b	transp.dat
3-2	assign.dat
4-1	multi.mod
4-2	multi.dat
4-3	steelT0.mod
4-4	steelT.mod
4-5	steelT.dat
4-6	steelP.mod
4-7	steelP.dat
5-1	dietu.mod
5-2	dietu.dat
5-3	steelT2.mod
5-4	steelT2.dat
6-1	transp2.mod
6-2a	transp3.mod
6-2b	transp3.dat
6-3	steelT3.mod
6-4	steelT3.dat
6-5	multic.mod
13-1	steelT.sa3
13-2	steelT.tab1
13-3	steelT.sa4
13-4	steelT.sa5
13-5	steelT.sa7
14-1	diet.run
14-2	cut.mod
14-3	cut.run
14-4	cut.dat
14-6	cut2.mod
15-2a	net1.mod
15-2b	net1.dat
15-3a	net2.mod
15-3b	net2.dat
15-4a	net3.mod
15-4b	net3.dat
15-6	netmax.mod
15-7	netshort.mod
15-9	netasgn.mod
15-10	net1node.mod
15-11	net3node.mod
15-12	netmax3.mod
15-13	netmulti.mod
15-14	netfeeds.mod
15-15	netthru.mod
16-1	iorow.mod
16-2	iocol1.mod
16-3	iocol2.mod
16-4	sched.mod
16-5	sched.dat
16-6	netmcol.mod
17-2	transpl1.mod
17-3a	transpl2.mod
17-3b	transpl2.dat
17-5a	steelpl1.mod
17-5b	steelpl1.dat
17-9	minmax.mod
18-4	nltrans.mod
18-5	nltrans.dat
19-1	econmin.mod
19-2	econ.dat
19-3	econ.mod
19-4	econ2.mod
19-5	econnl.mod
20-1a	multmip1.mod
20-1b	multmip1.dat
20-2a	multmip3.mod
20-2b	multmip3.dat

Additional Scripts: Looping and Testing - 1#

Writing “scripts” in the AMPL command language

All examples use steelT.mod as the model file and the steelT.dat as the data file.

Implements	Script file
One pass of sensitivity analysis on avail[3] in the multi-period production problem (Section 4.2)	steelT.sa1
Same as steelT.sa1 but split into two scripts, using the commands command	steelT.sa1a steelT.sa1b
Iterated sensitivity analysis, using contents of steelT.sa1a and steelT.sa1b within a for loop	steelT.sa2
… also building a table of results	steelT.sa3
… also using repeat until, and building up an index set for the table	steelT.sa4
… also using if to record only those parameter values at which the dual price changes	steelT.sa5
Binary search to find a breakpoint in the dual value	steelT.sa6
… also handling the case of more than one breakpoint in the initial interval	steelT.sa6a
… also counting the extra breakpoints detected	steelT.sa6b
Sensitivity analysis using break and continue	steelT.sa7
… also for each t, printing at the end of each pass	steelT.sa7a
… also for each t, storing all values until the end	steelT.sa7b
… also incorporating a break criterion requiring a named loop	steelT.sa7c
Data-specific script for a formatted table	steelT.tab0
General script for formatted table, using a for loop	steelT.tab1
… also with if statement to suppress printing of 100%	steelT.tab2

Additional Scripts: Looping and Testing - 2#

Implementing algorithms through AMPL scripts

Implements	Script	Uses
Gilmore-Gomory column generation procedure for the cutting-stock (roll trim) problem	cut1.run	cut1.mod cut.dat
Same as cut1.run, but using an alternative arrangement wherein problems are defined immediately before before their members are declared	cut2.run	cut2.mod cut.dat
Same as cut1.run, but with better formatting of output	cut3.run	cut1.mod cut.dat
Dantzig-Wolfe decomposition for a multi-commodity transportation problem, using a single subproblem	multi1.run	multi1.mod multi1.dat
Same as multi1.run, but using the same repeat loop for both phase I (infeasible) and phase II (feasible).	multi1a.run	multi1.mod multi1.dat
Same as multi1.run, but using a separate subproblem for each product; subproblems are represented in AMPL by an indexed collection of named problems	multi2.run	multi2.mod multi2.dat
Same as multi2.run, except that the separate subproblems are realized by changing the data to a single AMPL named problem	multi3.run	multi3.mod multi3.dat
Benders decomposition for a stochastic programming variant of a multi-period production problem (see Exercise 4-5)	stoch1.run	stoch1.mod stoch.dat
Same as stoch1.run, but using a separate subproblem for each scenario; subproblems are represented in AMPL by an indexed collection of named problems	stoch2.run	stoch2.mod stoch.dat
Same as stoch2.run, except that the separate subproblems are realized by changing the data to a single AMPL named problem	stoch3.run	stoch3.mod stoch.dat
Benders decomposition for a location-transportation problem (original model in trnloc.mod)	trnloc1.run	trnloc1.mod trnloc.dat
Same as trnloc2a.run, but model has upper limits on the Ship variables: LP relaxation bound is still poor, but subproblem does not have the integrality property and considerable improvement is made	trnloc2b.run	trnloc2b.mod trnloc2.dat
Same as trnloc2b.run, but model has 0-1 constraints disaggregated: LP relaxation bound is good, but subproblem has the integrality property and no further improvement can be made	trnloc2c.run	trnloc2c.mod trnloc2.dat
Lagrangian relaxation for a location-transportation problem: LP relaxation bound is poor, and subproblem has the integrality property so no improvement can be made	trnloc2a.run	trnloc2a.mod trnloc2.dat

Logic & Constraint Programming Examples#

This is a preliminary set of examples to offer some starting points for experimenting with AMPL’s “logic” and constraint programming interfaces. We welcome comments for improvements or other examples.

Represents	Model File
Integer program for assignment of people of groups so that each group is “balanced” by various criteria.	balAssign0.mod balAssign0.dat
Constraint program for balanced assignment of people to groups, using the numberof operator.	balAssign1.mod balAssign1.dat
Model for map coloring using not-equals constraint.	mapColoring.mod
Integer program for the n queens problem: put n queens on an n-by-n chessboard so that no queen is attacking any other queen.	nQueens0.mod
Constraint program for the n queens problem, using the alldiff operator.	nQueens.mod
Open-shop scheduling model using disjunctive constraints to avoid job and machine conflicts.	openShop.mod openShop.dat
Line up a group of people to maximize the number of adjacency preferences satisfied, modeling with alldiff, equivalence (<==>), and strict inequality (<)	photo.mod photo9.dat photo11.dat
Hoist scheduling model using disjunctive constraints	singleHoist.mod singleHoist3.dat singleHoist4.dat
Model to solve instances of the 3 x 3 Sudoku problem, using the alldiff operator	sudoku.mod sudokuVeryEasy.dat sudokuHard.dat