SET Command
The SET command allows you to override LINGO's default tolerances and settings. All user configurable options in LINGO are available through the SET command. The syntax for the SET command is:
SET parameter_name | parameter_index [parameter_value]
where,
parameter_name | is the name of the parameter to set, |
parameter_index | is the index of the parameter to set, and |
parameter_value | is the new value for the parameter that, if omitted, will cause LINGO to display the current value for the specified parameter. |
Use the FREEZE command to save any tolerances modified with the SET command to the configuration file, so they will be automatically restored the next time LINGO starts. You may also enter SET DEFAULT to return all parameters to their default values.
Some examples of the SET command follow:
Example 1: |
SET MXMEMB 128 |
|
FREEZE |
sets the generator memory limit to 128MB and saves parameter settings to
the configuration file,
Example 2: SET 5 1.E-7
sets the relative integrality tolerance (RELINT) to 1.e-7,
Example 3: SET DEFAULT
restores all parameters to their default values, and
Example 4: HELP SET
causes LINGO to display all parameter settings.
The parameters accessible through the SET command are:
Index |
Parameter Name |
Default |
Description |
---|---|---|---|
1 |
ILFTOL |
0.3e-5 |
Initial linear feasibility tolerance |
2 |
FLFTOL |
0.1e-6 |
Final linear feasibility tolerance |
3 |
INFTOL |
0.1e-2 |
Initial nonlinear feasibility tolerance |
4 |
FNFTOL |
0.1e-5 |
Final nonlinear feasibility tolerance |
5 |
RELINT |
0.8e-5 |
Relative integrality tolerance |
6 |
NOPTOL |
0.1e-6 |
NLP optimality tolerance |
7 |
ITRSLW |
5 |
Iteration limit for slow progress |
8 |
DERCMP |
0 |
Derivatives (0:LINGO chooses, 1:backward analytical, 2:forward analytical, 3:central differences, 4:forward differences) |
9 |
ITRLIM |
0 |
Iteration limit (0:no limit) |
10 |
TIMLIM |
0 |
Solver time limit in seconds (0:no limit) |
11 |
OBJCTS |
1 |
Objective cuts (1:yes, 0:no) |
12 |
MXMEMB |
32 |
Memory limit in megabytes for LINGO’s model generator (N/A on some machines) |
13 |
CUTAPP |
2 |
Cuts application (0:root, 1:all, 2:solver chooses) |
14 |
ABSINT |
.000001 |
Absolute integrality tolerance |
15 |
HEURIS |
3 |
IP heuristics (0:none, 100:advanced) |
16 |
HURDLE |
0 |
Use an IP hurdle value (1:yes, 0:no) |
17 |
IPTOLA |
0 |
IP absolute optimality tolerance |
18 |
IPTOLR |
.1e-4 |
IP relative optimality tolerance |
19 |
TIM2RL |
100 |
Seconds before switching to IP relative optimality tolerance |
20 |
NODESL |
0 |
0:LINGO decides, 1:depth first, 2:worst bound, 3:best bound |
21 |
LENPAG |
0 |
Terminal page length limit (0:none) |
22 |
LINLEN |
85 |
Terminal page width (0:none) |
23 |
TERSEO |
0 |
Output level (0:verbose, 1:terse, 2:errors only, 3:no output) |
24 |
STAWIN |
1 |
Post status window (1:yes, 0:no) |
25 |
SPLASH |
1 |
Display splash screen (1:yes, 0:no) |
26 |
OROUTE |
0 |
Route output to command window (1:yes, 0:no) |
27 |
WNLINE |
800 |
Max command window lines |
28 |
WNTRIM |
400 |
Min command window lines |
29 |
STABAR |
1 |
Display status bar (1:yes, 0:no) |
30 |
FILFMT |
1 |
File format (0:lng, 1:lg4, 2:ltx) |
31 |
TOOLBR |
1 |
Display toolbar (1:yes, 0:no) |
32 |
CHKDUP |
0 |
Check for duplicate model names in data (1:yes, 0:no) |
33 |
ECHOIN |
0 |
Echo command input to terminal (1:yes, 0:no) |
34 |
ERRDLG |
1 |
Route error messages to a dialog box (1:yes, 0:no) |
35 |
USEPNM |
0 |
Allow for unrestricted use of primitive set names (1:yes, 0:no) |
36 |
NSTEEP |
0 |
Use steepest edge variable selection in nonlinear solver (1:yes, 0:no) |
37 |
NCRASH |
0 |
Run crash procedure 1 to get an initial starting point in nonlinear models (1:yes, 0:no). See parameter 91, LCRASH, below for alternative procedure. |
38 |
NSLPDR |
1 |
Compute search directions in nonlinear solver using successive linear programming (1:yes, 0:no) |
39 |
SELCON |
0 |
Use selective constraint evaluation in nonlinear solver (1:yes, 0:no) |
40 |
PRBLVL |
0 |
Specify probing level on MILPs (0:LINGO chooses, 1:none, 7:high) |
41 |
SOLVEL |
0 |
Specify linear solver (0:LINGO chooses, 1:primal, 2:dual, 3:barrier) |
42 |
REDUCE |
2 |
Perform model reduction (2:LINGO chooses, 1:yes, 0:no) |
43 |
SCALEM |
1 |
Scale the model (1:yes, 0:no) |
44 |
PRIMPR |
0 |
Select primal pricing method (0:LINGO chooses, 1:partial, 2:devex) |
45 |
DUALPR |
0 |
Select dual pricing method (0:LINGO chooses, 1:Dantzig, 2:steepest edge) |
46 |
DUALCO |
1 |
Specify dual computations (0:none, 1:prices only, 2:prices and ranges, 3:prices only on optimizable rows) |
47 |
RCMPSN |
0 |
Use RC format names for MPS I/O (1:yes, 0:no) |
48 |
MREGEN |
2 |
Select model regeneration (0:only on modifications to model, 1:same as 0 plus whenever model has external references, 2:always) |
49 |
BRANDR |
0 |
Select branch direction (0:both, 1:up, 2:down) |
50 |
BRANPR |
0 |
Select branch priority (0:LINGO decides, 1:binary) |
51 |
CUTOFF |
.1e-8 |
Cutoff solution values smaller than this |
52 |
STRONG |
10 |
Specify strong branch level |
53 |
REOPTB |
0 |
IP warm start LP (0:LINGO, 1:primal, 2:dual, 3:barrier) |
54 |
REOPTX |
0 |
IP cold start LP (0:LINGO, 1:primal, 2:dual, 3:barrier) |
55 |
MAXCTP |
100 |
Max top cuts passes |
56 |
RCTLIM |
.75 |
Relative cuts limit |
57 |
GUBCTS |
1 |
GUB cuts (1:yes, 0:no) |
58 |
FLWCTS |
1 |
Flow cuts (1:yes, 0:no) |
59 |
LFTCTS |
1 |
Lift cuts (1:yes, 0:no) |
60 |
PLOCTS |
1 |
Plant location cuts (1:yes, 0:no) |
61 |
DISCTS |
1 |
Disaggregation cuts (1:yes, 0:no) |
62 |
KNPCTS |
1 |
Knapsack cover cuts (1:yes, 0:no) |
63 |
LATCTS |
1 |
Lattice cuts (1:yes, 0:no) |
64 |
GOMCTS |
1 |
Gomory cuts (1:yes, 0:no) |
65 |
COFCTS |
1 |
Coefficient reduction cuts (1:yes, 0:no) |
66 |
GCDCTS |
1 |
Greatest common divisor cuts (1:yes, 0:no) |
67 |
SCLRLM |
1,000 |
Syntax coloring line limit |
68 |
SCLRDL |
0 |
Syntax coloring delay in seconds |
69 |
PRNCLR |
1 |
Matching parenthesis coloring (1:yes, 0:no) |
70 |
MULTIS |
0 |
NLP multistart attempts (0:LINGO, n:number of attempts) |
71 |
USEQPR |
1 |
Use quadratic recognition (1:yes, 0:no) |
72 |
GLOBAL |
0 |
Use global solver on NLPs (1:yes, 0:no) |
73 |
LNRISE |
0 |
Linearization (0:LINGO, 1:none, 2:low, 3:high) |
74 |
LNBIGM |
100,000 |
Linearization BigM coefficient |
75 |
LNDLTA |
.1e-5 |
Linearization Delta coefficient |
76 |
BASCTS |
1 |
Basis cuts (1:yes, 0:no) |
77 |
MAXCTR |
2 |
Max tree cuts passes |
78 |
HUMNTM |
0 |
Minimum heuristic time limit (seconds) |
79 |
DECOMP |
0 |
Matrix decomposition (1:yes, 0:no) |
80 |
GLBOPT |
.1e-4 |
Global solver optimality tolerance |
81 |
GLBDLT |
.1e-6 |
Global solver delta tolerance |
82 |
GLBVBD |
.1e+11 |
Global solver variable bound limit |
83 |
GLBUBD |
2 |
Global solver bound use (0:no, 1:all, 2:some) |
84 |
GLBBRN |
5 |
Global solver branch selection (see below) |
85 |
GLBBXS |
1 |
Global solver box selection (0:depth first, 1:worst bound) |
86 |
GLBREF |
3 |
Global solver reformulation level (0:none, 3:high) |
87 |
SUBOUT |
2 |
Fixed variable reduction (0:none, 1:max, 2:not when using global or multistart solvers, 3:linear variables only) |
88 |
NLPVER |
N/A |
No longer supported |
89 |
DBGCLD |
0 |
Debugging cold start solver (0:LINGO, 1:primal, 2:dual, 3:barrier) |
90 |
DBGWRM |
0 |
Debug warm start solver (0:LINGO, 1:primal, 2:dual, 3:barrier) |
91 |
LCRASH |
0 |
Run crash procedure 2 to get an initial starting point in nonlinear models (1:yes, 0:no). See parameter 37, NCRASH, above for alternative procedure. |
92 |
BCROSS |
1 |
Perform a basis crossover on LPs when using barrier solver (0:no, 1:yes) |
93 |
LOWMEM |
0 |
Opt for less memory usage (0:no, 1:yes) |
94 |
FILOUT |
0 |
Fill out workbook output ranges (0:no, 1:yes) |
95 |
DBGLVL |
15 |
Debugger output level (1:low, 15:high) |
96 |
UNARYM |
1 |
Unary minus priority (0:low, 1:high) |
97 |
LINEAR |
0 |
Assume model is linear to reduce memory consumption(0:no, 1:yes) |
98 |
LOPTOL |
.1e-6 |
Linear optimality tolerance |
99 |
SECORD |
0 |
Use second order derivatives for NLPs (0:no, 1:yes) |
100 |
NONNEG |
1 |
Variables default to being non-negative (0:no, 1:yes) |
101 |
BIGMVL |
1.e8 |
BigM coefficient threshold value |
102 |
KILLSC |
0 |
Kill scripts on interrupts (0:no, 1:yes) |
103 |
TATSLV |
0 |
@SOLVE time limit in seconds |
104 |
KBESTS |
1 |
Number of K-Best MIP solutions to generate |
105 |
LCORES |
1 |
Number of concurrent LP solvers to run |
106 |
LCORE1 |
1 |
LP Solver in Core 1 (1:prm1,2:dual,3:barrier,4:prim2) |
107 |
LCORE2 |
2 |
LP Solver in Core 2 |
108 |
LCORE3 |
3 |
LP Solver in Core 3 |
109 |
LCORE4 |
4 |
LP Solver in Core 4 |
110 |
SCALEW |
1.e12 |
Scaling warning threshold |
111 |
REFRAQ |
0 |
Basis refactor frequency (0:LINGO chooses, iteration count) |
112 |
SPSMSZ |
2 |
Default SP sample size per stage |
113 |
SPSCON |
1 |
Apply SP sampling to continuous random variables only (0:no, 1:yes) |
114 |
SPSEED |
1,031 |
SP sampling random number seed |
115 |
SPMTHD |
0 |
SP solution method (0:LINGO chooses, 1:deterministic equivalent, 2:nested Benders, 3:augmented Lagrangian) |
116 |
SPXVAL |
1 |
Calculate all SP expected value statistics (0:no, 1:yes) |
117 |
SPMAXS |
40,000 |
Max scenarios allowed in an SP before auto sampling takes effect |
118 |
PRECIS |
7 |
Precision in digits for standard solution reports |
119 |
LOOPOP |
0 |
Perform loop optimization during model compiles (0:no, 1:yes) |
120 |
HEUCUT |
0 |
Heuristics cutoff criterion (0:LINGO chooses, 1:time, 2:iterations) |
121 |
NUMBIN |
0 |
Number of bins in histogram charts (0:LINGO chooses, >0: number of bins) |
122 |
SPBIGM |
1.E8 |
Stochastic solver Big M coefficient |
123 |
NSLPSV |
0 |
Use SLP solver for nonlinear models (0:no, 1:yes) |
124 |
FORCEB |
0 |
Enforce variable bounds in calc and data sections (0:no, 1:yes) |
125 |
NTHRDS |
1 |
Max number of executions threads (0:use all cores, 1:single-threaded, N>1:use up to N threads |
126 |
MTMODE |
-1 |
Multithreading mode( -1:LINGO chooses,0:off in solver, 1:prefer parallel, 2:parallel exclusively, 3:prefer concurrent, 4:concurrent exclusively) |
127 |
BNPBLK |
2 |
Branch-and-price (BNP) blocks (0:use row names, 1:user specified, 2:off, >2:max number of blocks) |
128 |
BNPHEU |
1 |
BNP block-finding heuristic (1:GP1, 2:GP2) |
129 |
REPROD |
1 |
Favor reproducibility (0:no, 1:yes) |