http://support.gams-software.com/ 2010-08-27T05:38:35-04:00 http://support.gams-software.com/ http://support.gams-software.com/lib/images/favicon.ico text/html 2007-10-21T00:44:25-04:00 http://support.gams-software.com/doku.php?id=solver:a_newer_version_of_the_solver_returns_a_different_solution&rev=1192941865&do=diff Q: I noticed that different versions of GAMS give different results for the same input file. In my case I am running a quadratic model with MINOS Contributed by Arne Drud: Do you know if your model can have more than one optimal solution (is it STRICTLY convex or has other nice properties)? Are the differences large or small (round-off errors)? Different versions of an NLP algorithms can give different results if the model has more than one solution. And if the optimum is very flat, then yo… text/html 2007-07-31T10:30:58-04:00 http://support.gams-software.com/doku.php?id=solver:a_reference_for_the_cplex_sifting_-method&rev=1185892258&do=diff Q: One of the options for choice of method for solving LP problems in the Cplex solver is the sifting method. Can you suggest a publication in the open mathematical literature that describes this method ? Here's the reference: Bixby, R.E., Gregory, J.W., Lustig, I.J., Marsten, R.E., and Shanno, D.F. (1992). “Very Large-Scale Linear Programming: A Case Study in Combining Interior Point and Simplex Methods”, Operations Research 40(5), 885-897. text/html 2008-09-10T04:33:03-04:00 http://support.gams-software.com/doku.php?id=solver:accessing_best_possible_objective&rev=1221035583&do=diff Q: BARON reports the best feasible point found (UBD in a minimization), as well as the “best possible” value of the objective (LBD in a minimization) in the output. For my application the latter (“best possible”) is needed. However, I did not find a way of accessing this from within a GAMS program. text/html 2008-05-08T09:48:49-04:00 http://support.gams-software.com/doku.php?id=solver:bad_rowindex_in_readat&rev=1210254529&do=diff This error pops up if you are using GAMS scaling (.scale) in connection with DICOPT and CPLEX as sub solver for MIP problems. The complete error message is: ... ILOG CPLEX May 1, 2008 22.7.1 WIN 3927.4700 VIS x86/MS Windows Cplex 11.0.1, GAMS Link 34 Reading data... Bad rowindex in readat(). --- DICOPT: MIP solver did not write a solution file DICOPT terminated prematurely because of a solver failure. The listing file will contain additional logs from the NLP and MI… text/html 2007-08-10T06:44:59-04:00 http://support.gams-software.com/doku.php?id=solver:can_i_tell_conopt_to_run_the_slp_algorithm_exclusively&rev=1186742699&do=diff Answer by Arne Stolbjerg Drud: The SLP solver inside CONOPT is used to find good search directions by approximately solving a linear approximation to the model in the current point. However, CONOPT uses a linesearch in this direction and uses the final basis from the LP to restore feasibility. This is different from 'classical SLP'. In CONOPT we start from a feasible point and feasibility is restored outside the SLP framework. In classical SLP feasibility and change in objective are handled si… text/html 2009-08-11T08:32:26-04:00 http://support.gams-software.com/doku.php?id=solver:changing_the_default_solver&rev=1249993946&do=diff Q: I would like to change the default solver (e.g. use CONOPT instead of MINOS for NLPs). Changing the default solver can be done in various ways: * Windows: Using the GAMSIDE: File -> Options -> Solvers. Click on the intersection of solver and model type that you want to be the default combination. * Unix: Run gamsinst again and change the default solver there. * Use a command line parameter, e.g. nlp=CONOPT * Add a line to your model (before the current solve statement) to switc… text/html 2009-01-22T18:12:01-04:00 http://support.gams-software.com/doku.php?id=solver:counting_the_number_of_discrete_variables&rev=1232665921&do=diff Q: How can I pass the the number of discrete variables shown in the solve summary section to a parameter? Please use the (undocumented) model attribute .numdvar. Find below an example for your reference. $call gamslib prodsch $include prodsch parameter numint; numint = prod1.numdvar; display numint; text/html 2010-03-08T06:40:41-04:00 http://support.gams-software.com/doku.php?id=solver:cplex_error_32201&rev=1268048441&do=diff This error should not occur, if you have purchased a license for GAMS/CPLEX. If you are using a GAMS/CPLEX-Link it is mandatory that the ILOG license manager is also installed and running properly. You also need to set an environment variable ILOG_LICENSE_FILE to point to the ILOG license file access.ilm. text/html 2007-06-23T10:45:14-04:00 http://support.gams-software.com/doku.php?id=solver:cplex_getting_a_list_of_best_integer_solutions_by_cplex&rev=1182609914&do=diff Q: I have an MIP which I'm solving using CPLEX. Besides the optimal solution, I'm also interested in the second,third,fourth etc. best solution. Is there a way for GAMS to tell CPLEX to return a ranked-order list of these solutions? I understand that this can be done with integer cuts... “Solve once, get first solution, append corresponding integer cut, resolve, get second best solution etc....” text/html 2007-07-31T10:23:16-04:00 http://support.gams-software.com/doku.php?id=solver:default_upper_bound_for_an_integer_variable&rev=1185891796&do=diff Q: I have a model with large integer variables but I don't manage to get a level greater than 100 For some historical reasons, GAMS sets an upper bound of 100 for all integer variables. The user has to reset the upper bound, i.e.: x.up(i) = 1000; However, formulating integer variables, which can take such huge numbers, as integer variables is in general not reasonable. text/html 2007-04-05T14:06:50-04:00 http://support.gams-software.com/doku.php?id=solver:dicopt_illegal_row_type&rev=1175796410&do=diff Q: I receive this following error, when running my MINLP with DICOPT. -------------------------------------------- S O L V E S U M M A R Y MODEL rlp OBJECTIVE zobjR TYPE MINLP DIRECTION MINIMIZE SOLVER DICOPT FROM LINE 314 **** SOLVER STATUS ERROR SYSTEM FAILURE **** MODEL STATUS ERROR NO SOLUTION **** OBJECTIVE VALUE NA RESOURCE USAGE, LIMIT NA 1000.000 ITERA… text/html 2008-09-15T10:07:55-04:00 http://support.gams-software.com/doku.php?id=solver:difference_between_gams_xpress-link_and_gams_xpress_in_the_log_file&rev=1221487675&do=diff GAMS/XPRESS Link ... Xpress-MP Aug 1, 2008 22.8.1 WIN 5247.6015 VIS x86/MS Windows Xpress-Optimizer v18.00.08 Hyper capacity, MIP, Barrier, QP/MIQP (c) Copyright Dash Associates 1984-2007 Reading data . . . done. ... GAMS/XPRESS ... Xpress-MP Aug 1, 2008 22.8.1 WIN 5247.6015 VIS x86/MS Windows Xpress-MP licensed by Dash to GAMS Development Corp. for GAMS Reading data . . . done. ... text/html 2007-09-21T08:44:53-04:00 http://support.gams-software.com/doku.php?id=solver:difference_between_nlp_and_dnlp-models&rev=1190378693&do=diff Q: Is there any disadvantage in using DNLP instead of a NLP? Submitted by Arne Drud: The distinction between NLP and DNLP is (currently) only in the GAMS compiler where certain constructs are defined as potentially dangerous and the modeler must explicitly say that she or he recognizes the danger with the clause “using dnlp”. text/html 2008-09-15T10:04:35-04:00 http://support.gams-software.com/doku.php?id=solver:difference_in_the_log_file_of_the_gams_xpress-link_and_gams_xpress&rev=1221487475&do=diff Xpress Link ... Xpress-MP Aug 1, 2008 22.8.1 WIN 5247.6015 VIS x86/MS Windows Xpress-Optimizer v18.00.08 Hyper capacity, MIP, Barrier, QP/MIQP (c) Copyright Dash Associates 1984-2007 Reading data . . . done. ... GAMS/XPRESS ... Xpress-MP Aug 1, 2008 22.8.1 WIN 5247.6015 VIS x86/MS Windows Xpress-MP licensed by Dash to GAMS Development Corp. for GAMS Reading data . . . done. ... text/html 2007-04-04T02:41:35-04:00 http://support.gams-software.com/doku.php?id=solver:different_results_under_windows_and_linux&rev=1175668895&do=diff Q: We have a large economics model running on both Windows and a Linux machine. The main development is on Windows. Recently we copied the model to Linux to run. The model gave us differences in the 5 significant digits. (XXXX.5374 was now XXXX.4157, for example). The model code is exactly the same (except for some obvious file path changes). Is this possible? text/html 2009-05-29T07:40:33-04:00 http://support.gams-software.com/doku.php?id=solver:division_by_a_variable_makes_model_infeasible&rev=1243597233&do=diff Q: Within a NLP I wrote an equation like this at first: POWER_DEMAND( I ) =E= SUBSTATION_SUPPLY( I ) + ADDTIONAL_SUPPLY( I ) and got a good solution back. But if I write that equation this way: POWER_DEMAND( I ) / SUBSTATION_SUPPLY( I )=E= 1 + ADDTIONAL_SUPPLY( I ) / SUBSTATION_SUPPLY( I ) text/html 2009-09-14T08:59:40-04:00 http://support.gams-software.com/doku.php?id=solver:error_1001_out_of_memory&rev=1252933180&do=diff This message indicates that Cplex is running out of memory while solving the model. The log file will provide more information about the size of the model and where in the solution process the memory message appears: * during preprocessing of the model and while solving the preprocessed model * during the branch & bound process * during the final solve and the calculation of the marginals. text/html 2007-11-07T10:26:46-04:00 http://support.gams-software.com/doku.php?id=solver:error_baron_has_encountered_a_licensing_error_barstatus_11_despite_of_a_valid_baron_license&rev=1194449206&do=diff Please check the log file of your model run. Most likely some lines will read: ... =========================================================================== Factorable Non-Linear Programming =========================================================================== LP Solver: CPLEX NLP Solver: MINOS .. text/html 2008-05-08T10:33:56-04:00 http://support.gams-software.com/doku.php?id=solver:execution_error_s_after_solve_and_scr-files_in_working_directory&rev=1210257236&do=diff This error pops up if you are using GAMS scaling (.scale) in connection with ALPHAECP and CPLEX as sub solver for MIP problems. Only the distributions 22.6 and 22.7.1 are affected by this bug. If your system is still under under maintenance, please download the current GAMS version from our download server. text/html 2007-07-31T07:05:54-04:00 http://support.gams-software.com/doku.php?id=solver:exit_the_current_point_cannot_be_improved_upon&rev=1185879954&do=diff Q: MINOS performs about five major iterations and exits with the message: EXIT THE CURRENT POINT CANNOT BE IMPROVED UPON. I have tried in vain for the past few days trying to fix this problem. Could you suggest what might be wrong? This is not always bad, it means that MINOS is stuck in a local point without being able to establish local optimality. This may very well be the point you want to add up. It is difficult to to say more without having a look at the model. Sometimes it helps to re… text/html 2008-06-19T16:57:01-04:00 http://support.gams-software.com/doku.php?id=solver:fatal_error_insufficient_memory_for_setup_of_optimization_vectors&rev=1213909021&do=diff The message indicates that CONOPT has ran out of memory. Since CONOPT is written in FORTRAN, it must estimate its memory requirements and then do one large allocation. It that turns out not to be large enough, it is a fatal error. You should find more information in the CONOPT log which will have been copied to the lst file right after the Solve Summary. That information will include what the estimate was that was insufficient. text/html 2007-09-21T08:31:09-04:00 http://support.gams-software.com/doku.php?id=solver:feasible_solution._the_tolerances_are_minimal_and_there_is_no_change_in_objective&rev=1190377869&do=diff Q: While optimizing a non-linear minimization problem using Conopt, I occasionally (i.e., in certain cases conditions for which are not clear to me) experience the problem that the objective value appears to increase again, before the solver exits with text/html 2007-07-31T05:01:12-04:00 http://support.gams-software.com/doku.php?id=solver:fixed_lp_of_a_mip_is_infeasible&rev=1185872472&do=diff Q: I solve a mixed-linear integer problem with Cplex and I get the following message: Final LP did not return an optimal solution. Proven optimal solution. What is the difference between this solution and the one obtained when the message is just: text/html 2010-07-13T10:37:54-04:00 http://support.gams-software.com/doku.php?id=solver:gams_cplex-link_and_osicplex-link_licenses&rev=1279031874&do=diff The GAMS/CPLEX Link and GAMS/OSICPLEX each require that you have a valid license agreement with IBM for use of the current version of the CPLEX library. To add either solver to your GAMS license, please write to sales@gams.com to confirm that you have such a license agreement and that your use of the GAMS solver will comply with the terms of that license agreement. Please reference your License ID, DCxxxx. text/html 2007-12-16T16:08:48-04:00 http://support.gams-software.com/doku.php?id=solver:gams_cplex_versus_cplex_callable_library&rev=1197839328&do=diff Q: I use GAMS/CPLEX link for my optimization problems. Some of my colleagues call the CPLEX callable library directly from C. Can there be a difference in solution quality, etc.? The GAMS/CPLEX link sets up the problem in the CPLEX and calls one of the optimization routines of the CPLEX library (e.g. CPXmipopt). Assuming your C program and the GAMS/CPLEX link setup the problem in the same way, you will get identical results. The GAMS/CPLEX link also exposes all CPLEX features that are relevant… text/html 2008-06-30T13:18:16-04:00 http://support.gams-software.com/doku.php?id=solver:getting_a_list_of_best_integer_solutions_of_my_mip_model&rev=1214846296&do=diff Q: I have an MIP which I'm solving using CPLEX. Besides the optimal solution, I'm also interested in the second,third,fourth etc. best solution. Is there a way for GAMS to tell CPLEX to return a ranked-order list of these solutions? I understand that this can be done with integer cuts... “Solve once, get first solution, append corresponding integer cut, resolve, get second best solution etc....” text/html 2009-09-30T09:39:52-04:00 http://support.gams-software.com/doku.php?id=solver:getting_a_list_of_the_default_values_for_cplex_option_within_gams&rev=1254317992&do=diff Use the Cplex option printoptions 1 to get such a list into the lst file. Below is an example. variables z, x; equation foo; foo.. z=e=2*x; x.up=1; model m /all/; m.optfile=1; option lp=cplex; $onecho>cplex.opt printoptions 1 $offecho Solve m using lp maximizing z; text/html 2009-03-26T15:54:00-04:00 http://support.gams-software.com/doku.php?id=solver:glibcxx_3.4.4_not_found&rev=1238097240&do=diff A user could not run the COIN solvers on a Linux-64 bit machine because the COIN solvers require very recent versions of GLIBC, i.e. the GNU C/C++ library. The possible workarounds are: * to install a more recent C++ library * to move to another machine with a more recent C++ library. * to contact the maintainer of the GAMS-COIN Links (Stefan Vigerke to get a fully-static link of the solver (for Linux 64bit). text/html 2009-09-02T16:14:33-04:00 http://support.gams-software.com/doku.php?id=solver:how_can_i_change_the_default_sub_solver_for_emp&rev=1251922473&do=diff You will need an option file emp.opt with this line: subsolver=coincbc (or any other [licensed] solver suitable for your problem). To pass a specific option file for the subsolver use the EMP option: SubSolverOpt (integer). For a complete example please inspect the model sequence of the EMPlib. text/html 2008-03-12T05:15:14-04:00 http://support.gams-software.com/doku.php?id=solver:how_can_i_hook_my_own_solver_to_gams&rev=1205313314&do=diff For experimental links we recommend to look into a COIN-OR project called GAMSlinks. This project has everything that is required to connect a solver to GAMS. There are example drivers for MIP (Cbc, Glpk), NLP (IpOpt), and MINLP (Bonmin) solvers. The project leader (Stefan Vigerske) is usually extremely helpful if there are any questions. text/html 2008-06-18T08:24:18-04:00 http://support.gams-software.com/doku.php?id=solver:how_do_i_automatically_re-submit_a_model_which_failed_due_to_a_solver_license_error&rev=1213791858&do=diff Q: I have a batch job consisting of a large number of gams models to be run, each of these requiring cplex. I am running this on my faculty's it system, where we share a number of licenses. It varies a lot how many is available, but sometimes none are available. This answer is returned to gams, which outputs this and terminates. That is very unfortunate for my application and I would very much prefer if it waits until a license becomes available. So my question is, can some flag be set for ei… text/html 2010-05-19T05:07:43-04:00 http://support.gams-software.com/doku.php?id=solver:how_do_i_fix_a_semiint_cont_variable&rev=1274260063&do=diff A “fixed” semicont variable has still the implicit lower bound of 0. In order to truly fix it the variable needs to be relaxed via RMIP or to add to attribute prior=INF. The example below illustrates this. variable x,z; semicont variable x; equation e; e.. z =e= x; model m /all/; * A "fixed" semicont variable has still the implicit lower bound of 0 * The domain for x is {0,1} x.fx=1; solve m us mip min z; abort$(abs(z.l-0)>1e-6) 'wrong obj'; * In order to truly fix it the variable needs to… text/html 2007-05-03T05:34:49-04:00 http://support.gams-software.com/doku.php?id=solver:how_do_i_interrupt_a_solver&rev=1178184889&do=diff I want to interrupt my solver and have it return the current status, in much the same way as would be done if an iteration or resource limit were exceeded. How can I do this? The Unix and PC systems now support a user interrupt that can be triggered by the user to stop a solver job at the next convenient point. Typically, solvers will check for a user interrupt in the same places where the resource limit is checked. A user might wish to trigger an interrupt in order to stop a MIP job that has … text/html 2007-11-20T13:22:50-04:00 http://support.gams-software.com/doku.php?id=solver:how_is_resusd_being_calculated_when_cplex_runs_sequentially_parallel&rev=1195582970&do=diff For GAMS on Unix (We did our experiments on Linux but are pretty sure that the other UNIX systems behave the same. If there is still doubt, please let us know what Unix you are running) CPLEX returns CPU seconds in resusd if Cplex runs sequentially (which is the default unless you specify option threads, mipthreads, and/or choose the concurrent optimizer). text/html 2007-11-07T08:45:36-04:00 http://support.gams-software.com/doku.php?id=solver:how_much_memory_ram_does_my_gams_run_require&rev=1194443136&do=diff Memory is required by GAMS for model compilation and generation and by the solver during the optimization. Memory required by GAMS The log file (within the GAMS IDE the process window) will tell you how much memory was required by GAMS during compilation, execution and model generation: text/html 2007-08-08T03:33:16-04:00 http://support.gams-software.com/doku.php?id=solver:implementing_new_branching_rules&rev=1186558396&do=diff Q: I have implemented a MIP-model in GAMS, which I would like to solve to proven optimality. The problem has a lot of degenerate optimal problems, and solutions very close to the optimum, so it is not terminating with the optimal solution, but continuing to examine nodes with a very small gap > 0.00 % I have a good idea for a branching rule, which i believe can work around the degeneracy problems. Instead of doing binary branching on fractional binary variables, I intend to branch on the s… text/html 2007-08-15T09:33:03-04:00 http://support.gams-software.com/doku.php?id=solver:improving_the_precision_of_the_interior_point_method&rev=1187184783&do=diff Q: Using Cplex to solve my QCP model returns a slightly different solution (e.g. 256.949847) than the one I get back from MINOS (e.g. 256.949789699778700). Is this just a matter of Cplex using smaller precision internally and therefore having a slightly different answer? I've noticed that it probably only uses 6 or 7 decimal places of precision whereas the same problem formulated as an nlp and solved in MINOS has a much higher precision, say 15 decimal places. text/html 2007-08-15T09:31:31-04:00 http://support.gams-software.com/doku.php?id=solver:improving_the_precision_of_the_results_of_the_interior_point_method&rev=1187184691&do=diff Q: Using Cplex to solve my QCP model returns a slightly different solution (e.g. 256.949847) than the one I get back from MINOS (e.g. 256.949789699778700). Is this just a matter of cplex using smaller precision internally and therefore having a slightly different answer? I've noticed that it probably only uses 6 or 7 decimal places of precision whereas the same problem formulated as an nlp and solved in MINOS has a much higher precision, say 15 decimal places. text/html 2007-07-31T07:11:26-04:00 http://support.gams-software.com/doku.php?id=solver:infeasibility_in_the_pre-triangular_part_of_the_model&rev=1185880286&do=diff This comes from the solver's presolve. It means that the model really is infeasible and that this can be determined before the numerical solution algorithm really starts. The presolve finds that one of your equations has just one variable in it and so can use that equation to determine the value of that variable. Once that variable's value is fixed, there may be other equations that now involve just one variable -- and so on, that's the pre-triangular part of the model. text/html 2007-08-10T03:25:37-04:00 http://support.gams-software.com/doku.php?id=solver:infeasible_nlp_s&rev=1186730737&do=diff Q: I have a MINLP problem to solve with nonlinearities at the continuous variables. The model is convex at these variables. When I use the Dicopt, none of the NLP's is feasible. I know that not all combination of binary variables gives a feasible continuous problem, so I incremented the number of Dicopt iterations (to 100), but without success. All NLP subproblems were infeasible. text/html 2009-08-05T10:59:27-04:00 http://support.gams-software.com/doku.php?id=solver:infeasible_solution._reduced_gradient_less_than_tolerance&rev=1249484367&do=diff Q: I am using “Conopt” to solve a large scale model. But the message I get back from the solver says: Infeasible solution. Reduced gradient less than tolerance How do I interpret this? The messages mean what they say. If you look at the Solve Summary in the listing file you will see that the status is Locally Infeasible. This means that there is no feasible solution near the point where CONOPT stopped. You should analyze the solution returned to get clues as to why this might be. text/html 2008-06-18T09:13:32-04:00 http://support.gams-software.com/doku.php?id=solver:intercepting_a_solver_call&rev=1213794812&do=diff The GAMS solve statement usually represents a monolithic call which can not be intercepted. There is an exception to this: the GAMS/Cplex interactive option which allows to interrupt the B&C process, to reset some options and to abort or continue the algorithms. However you can not look at the solution. Please also note that the 'interactive' option can not be used within the IDE since to input stream is attached to the process window. If you want this, you need to run it from the command line… text/html 2007-11-20T13:17:19-04:00 http://support.gams-software.com/doku.php?id=solver:is_it_possible_that_a_concurrent_execution_takes_longer_than_a_non-concurrent_execution_for_the_same_model&rev=1195582639&do=diff * Solving an LP with concurrent is always as fast as solving sequentially with one of the methods run in concurrent mode, but the resusd can be much larger for the concurrent since we report real time in the concurrent and CPU time in the sequential run. Depending on the load of the machine this can be quiet different. * If we talk about parallel MIP (parallel solve of node LPs), then the parallel run can be slower than the sequential, since the parallel can go a different route in the algori… text/html 2007-10-20T01:00:17-04:00 http://support.gams-software.com/doku.php?id=solver:miles_vs._path&rev=1192856417&do=diff Q: I have a MCP problem which is going to grow. I have both MILES and PATH,and since PATH costs extra, I am assuming it is in some sense better. I have not been able to find anything clear in the documentation about their relative merits. Running the small model I have now produces exactly the same solution, and in approximately the same time.The model is soon to be shared with other users, and I don’t know if they should buy PATH. Any experience or pointers to relevant literature would be appre… text/html 2008-04-26T00:51:01-04:00 http://support.gams-software.com/doku.php?id=solver:minos_declares_infeasible_model_as_feasible&rev=1209185461&do=diff Q: My question is, how come that GAMS (MINOS) declares the result as being feasible whereas in reality the solution is not feasible? GAMS/MINOS by default scales the model, solves the scaled model to optimality and then unscales the model and reports the solution. It can (rarely) happen that the unscaled model is infeasible. If you use anther LP solver (like CPLEX, XPRESS, or one of the free COIN solvers) for your model you will not have this problem. text/html 2007-08-14T05:35:37-04:00 http://support.gams-software.com/doku.php?id=solver:model_is_unbounded_model_has_an_unbounded_ray_model_status_3_unbounded&rev=1187084137&do=diff Q: I got an error shown as above (output from different solver). What is the possible reason for this error? Your model contains at least one variable with a coefficient in the objective but otherwise it does not show up in the constraint matrix (at least after preprocessing). The sign of the coefficient in the objective and the type of the variable (positive, negative, free) allows the model to become unbounded (or dual infeasible). Check the output for variables, which have a very large leve… text/html 2007-10-11T04:01:29-04:00 http://support.gams-software.com/doku.php?id=solver:model_which_worked_with_an_old_version_of_conopt_fails_after_upgrading_to_the_current_version&rev=1192089689&do=diff Q: I discovered that my model runs fine under the old GAMS 22.2, but won't run under the new 22.5 Are there 22.5 defaults I should set to make it run? Here is the log from 22.2: C O N O P T 3 x86/MS Windows version 3.14P-016-057 Copyright (C) ARKI Consulting and Development A/S Bagsvaerdvej 246 A DK-2880 Bagsvaerd, Denmark Using default options. Reading Data Iter Phase Ninf Infeasibility RGmax NSB Step InItr MX… text/html 2008-03-27T01:59:56-04:00 http://support.gams-software.com/doku.php?id=solver:multiple_cplex_threads&rev=1206597596&do=diff Q: I use CPLEX as my solver and have heard that it is possible to use multiple threads to decrease the solution time and use the full capacity of the computer. How is this implemented in GAMS? A GAMS run involves * the GAMS language compiler for processing the source code and data of the model and generating the model * and a solver for doing the optimization. text/html 2007-07-31T10:21:04-04:00 http://support.gams-software.com/doku.php?id=solver:multiple_solutions_of_a_mip&rev=1185891664&do=diff I was wondering how you would code a Mixed Integer Linear Program, so that you would be able to obtain all solutions with an objective value less than some predetermined value for a minimisation problem. Obtaining all solutions (with objective in a particular interval) of a MIP is not an easy task. If you have continous variables in the MIP there may be even infinite many optimal solutions. In this case there is no efficient general way. You may check books on 'sensitivity analysis' of LPs an… text/html 2007-05-18T08:25:21-04:00 http://support.gams-software.com/doku.php?id=solver:nlp_different_results_depending_on_the_order_of_data&rev=1179491121&do=diff Q: I am using a non-linear solver and I am finding I get different results depending on the order of my data. Can you suggest any reason why this would be? I assume that you've checked the model and solver status values in the Solve Summary to be sure that the model is stopping with “Normal Completion” and “Locally Optimal.” If not, you are just looking at the point where the solver gave up looking for a solution. Unless your model is convex, the objective function will have various hills and… text/html 2008-09-10T05:44:14-04:00 http://support.gams-software.com/doku.php?id=solver:nodetable_limit_exceeded&rev=1221039854&do=diff Q: I'm using BDMLP as MIP solver. The solver stops and can't find the solution and returns this message: NODETABLE limit exceeded You run against an internal limit of BDMLP. There is not much you can do about it other than switching to a better MIP solver. Recent GAMS systems for Windows, Solaris, Darwin and Linux include some solver from the Coin project. (you don't need an extra solver license for this). Please check the COIN section of the solver manuals for details. text/html 2007-11-16T08:10:29-04:00 http://support.gams-software.com/doku.php?id=solver:non_integer_results_of_a_mip&rev=1195218629&do=diff Q: I am getting results back from my MIP model, which I don’t expect. Although I define a variable to be integer, the output file shows values that they aren’t (eg, 0.194, etc). Please check the solution status. It must read something like: **** SOLVER STATUS 1 NORMAL COMPLETION **** MODEL STATUS 8 INTEGER SOLUTION text/html 2007-05-14T06:16:39-04:00 http://support.gams-software.com/doku.php?id=solver:not_enough_memory_to_solve_my_problem_with_osl&rev=1179137799&do=diff GAMS/OSL complains that there is not enough memory to solve my problem. What do I do ? The solver reports its memory estimate and availability in both the listing file and the log as follows: Work space requested by solver --- 23.45 MB Work space allocated --- 0.00 MB Maximum work space available --- 16.31 MB This indicates that GAMS/OSL estimated that 23.45 MB would be required to solve the problem, and since only 16.31 MB were available on the machine, no me… text/html 2009-12-01T13:52:18-04:00 http://support.gams-software.com/doku.php?id=solver:osicplex_osigurobi_osimosek_osixpress&rev=1259693538&do=diff GAMS/CoinCplex, GAMS/CoinGurobi, GAMS/CoinMosek, GAMS/CoinXpress bring the open source Open Solver Interface (OSI) to the broad audience of GAMS users. These bare bone solver links allow users to solve their GAMS models with a standalone license of Cplex, Gurobi, Mosek, or Xpress. The links use the COIN-OR Open Solver Interface (OSI) to communicate with these solvers. text/html 2007-11-16T04:38:32-04:00 http://support.gams-software.com/doku.php?id=solver:parallel_option_is_on_gams_license_but_gamscplex.ilm_is_missing_or&rev=1195205912&do=diff GAMS/Cplex Aug 1, 2005 WIN.CP.CP 22.0 029.032.041.VIS For Cplex 9.1* * Cplex 9.1.2, GAMS Link 29 * * Cplex licensed for 1 use of lp, qp, mip and barrier, with 4 parallel threads.* * Parallel option is on GAMS license but gamscplex.ilm is missing or ... text/html 2007-08-24T11:30:18-04:00 http://support.gams-software.com/doku.php?id=solver:presolve_found_the_problem_infeasible_or_unbounded&rev=1187969418&do=diff The message indicates that the CPLEX presolve procedures has declared a model infeasible before the selected optimization algorithm begins. This early declaration saves considerable execution time in most cases. The log file will give you more information: text/html 2008-05-16T09:15:14-04:00 http://support.gams-software.com/doku.php?id=solver:providing_additional_debugging_information_if_a_model_is_integer_infeasible&rev=1210943714&do=diff Q: Is there any way to trace the infeasiblility cause by which equations/constraints ( with =e=, =l= or =g= ), which set by the user, from GAMS when the modelstate = 10, ie 'No feasible integer solution could be found', has occurred? Most MIP solver do not return a solution at all, if a problem is integer infeasible. If you check the lst file, you will see a messages like: text/html 2007-11-08T07:58:14-04:00 http://support.gams-software.com/doku.php?id=solver:references_for_lp_qp-presolve_techniques&rev=1194526694&do=diff Find below a few references: E. D. Andersen and K. D. Andersen, Presolving in linear programming, Math. Programming 71(2):221--245, 1995 C. Mészáros and U.H. Suhl, Advanced preprocessing techniques for linear and quadratic programming, OR Spectrum, 25(4), Springer, 2003, 575-595 Jacek, Gondzio, Presolve Analysis of Linear Programs Prior to Applying an Interior Point Method, Logilab, HEC Geneva, Section of Management Studies, University of Geneva, 102 Bd Carl Vogt, CH-1211 Geneva 4, Sw… text/html 2007-10-20T12:18:17-04:00 http://support.gams-software.com/doku.php?id=solver:reliability_of_model_if_changing_rtmaxv&rev=1192897097&do=diff Q: We use CONPOT3 to solve our model. Although we tried to scale it properly, we failed, getting always following message: ** Infeasible solution. A variable has reached 'infinity'. Largest legal value (Rtmaxv) is 3.16E+07 Modifying the option file into: Rtmaxv = 5.e+08 resulted in a feasible solution with a warning that the variance of the derivatives in the initial point is large = 6.1 and should be optimized. text/html 2009-01-16T00:10:32-04:00 http://support.gams-software.com/doku.php?id=solver:sbb_could_not_rename_could_not_open_control_file&rev=1232082632&do=diff Q: I'm using gams 22.0 with sbb full version to solve a MINLP.My problem is that if i run the program in a desktop pc with XP operating system there is no problem but if i run gams with sbb solver in my laptop with vista operating system gams stops after running for a while and gives some error messages like: text/html 2007-03-20T15:43:21-04:00 http://support.gams-software.com/doku.php?id=solver:scaling_the_hessian&rev=1174419801&do=diff Q: I am working on a model with a non-linear objective function (2-level nested CES) and over 1600 non-linear variables and 160 linear constraints. I have been using GAMS/MINOS and CONOPT to solve it. All the variables are scaled to 1. But, both MINOS and CONOPT algorithms still stop short of optimal solution due to slow convergence. The problem is, I think, despite the scaling, the Hessian is still ill-conditioned as evidenced by a condition number of around 38000. There is a discussion of how … text/html 2010-03-29T09:36:11-04:00 http://support.gams-software.com/doku.php?id=solver:sensitivity_analysis_parameter_ranges_for_lp&rev=1269869771&do=diff How can I get sensitivity information (i.e. “parameter ranges”) for the rhs and objective function coefficients in a linear program? There are two types of sensitivity information that one can extract from a solved GAMS model, information available within GAMS and information available only within the LP solver. text/html 2010-03-29T08:55:36-04:00 http://support.gams-software.com/doku.php?id=solver:sensitivity_analysis_with_gams_cplex&rev=1269867336&do=diff Introduction Sensitivity analysis (post-optimality analysis) in linear programming allows one to find out more about an optimal solution for a problem. In particular, objective ranging and right-hand-side ranging give information about how much an objective coefficient or a right-hand-side coefficient can change without changing the optimal basis. In other words, they give information about how sensitive the optimal basis is to a change in the objective function or a right-hand side. text/html 2007-09-26T09:29:58-04:00 http://support.gams-software.com/doku.php?id=solver:setting_bounds_on_all_variables&rev=1190813398&do=diff Q: I have 4 variables. x (binary), t (continous), p (continuous) and Z (objective function) and set: t.lo = 0; t.up = 14; p.lo=0; p.up=1; //but BARON says I still did not give bounds for all variables. Why do I get this message even I bound my all variables (I get this message even I set bound for binary variables). text/html 2009-06-17T02:19:13-04:00 http://support.gams-software.com/doku.php?id=solver:setting_the_sub_solvers&rev=1245219553&do=diff Q: I would like to set sub-solvers (LP and NLP) for BARON solver when I tried to solve MINLP problem. I define solvers I want to use as follows: option minlp = baron; option lp = minos; option nlp = snopt; However, it seemed to me that BARON always use cplex for LP solver and minos for NLP solver all the times no matter how I defined solvers. text/html 2010-08-04T23:33:42-04:00 http://support.gams-software.com/doku.php?id=solver:solver&rev=1280979222&do=diff General * How can I hook my own solver to GAMS? * Solver ignores the iterlim/reslim settings in my model * How do I automatically re-submit a model, which failed due to a solver license error? * How much Memory (RAM) does my GAMS Run require? * Model is unbounded / Model has an unbounded ray / Model Status 3 unbounded * Changing the default Solver * Intercepting a Solver Call * The Solver Option File * How do I interrupt a solver? * Solver option file not found (wrong file e… text/html 2007-12-08T02:13:35-04:00 http://support.gams-software.com/doku.php?id=solver:solver_ignores_the_iterlim_reslim_settings_in_my_model&rev=1197098015&do=diff Q:I am getting an error “Iteration limit exceeded ...”, although I am using “option iterlim = 100000000;” in my model. There are various ways of setting the iteration limit the solver works with and that overwrite “option iterlim = 1000000”. For example: modelname.iterlim or using a solver option file (e.g. itlim for Cplex). The parameter settings of an active solver option file will always overwrite the limits you have set in your model or as a command line parameter. text/html 2008-06-19T01:43:11-04:00 http://support.gams-software.com/doku.php?id=solver:solver_option_file_not_found_wrong_file_extension&rev=1213854191&do=diff Sometimes a solver option file is not found by GAMS because the file extension is wrong, even if it appears the extension is .opt. The solver option file must be a text file and must have the .opt extension. When saving text files under Windows certain programs may in fact add a .txt extension to the options file so that the option file is (solver).opt.txt instead of (solver).txt. text/html 2007-10-20T01:16:46-04:00 http://support.gams-software.com/doku.php?id=solver:some_notes_on_scaling&rev=1192857406&do=diff Contributed by Arne Drud to the GAMS-User List: If you look at the equation listing for your model you will see: ---- E_DHEDGE =E= E_DHEDGE.. - (2.581140E+7)*X(1) - D_HDG =E= 0 ; (LHS = 4.1541556E+9, INFES = 4.1541556E+9 ***) Nonlinear models can be hard to solve if terms and derivatives are very large. And 2.8e7 = 28 million is very large. 4.15e9 is very very large. text/html 2010-07-09T03:49:27-04:00 http://support.gams-software.com/doku.php?id=solver:supported_sub_solvers&rev=1278661767&do=diff In order to use GAMS/BARON, users will need to have a GAMS/BARON license as well as a licensed GAMS linear programming (LP) solver. A licensed GAMS nonlinear programming (NLP) solver is optional and usually expedites convergence. By default, GAMS/BARON will attempt to use CPLEX as the LP solver and MINOS as the NLP solver. If the user does not have licenses for these solvers, then the user must use the options LPSol and NLPSol to specify another LP or NLP solver. text/html 2008-02-02T02:05:31-04:00 http://support.gams-software.com/doku.php?id=solver:terminate_the_optimization_after_gams_cplex_reaches_a_certain_value_of_the_objective&rev=1201935931&do=diff The example model below shows how to interact with CPLEX that it stops pretty much after CPLEX found the solution value 21 (or larger). This is done in using the same mechanism as described in ”The GAMS Branch-and-Cut-and-Heuristic Facility”. First we have userincbicall which is called every time we find a new incumbent. We also are also in ”interactive” mode, meaning that we accept new options on the fly. The trigger for reading in new options is a trigger file called readop2. This is created… text/html 2008-10-10T20:51:47-04:00 http://support.gams-software.com/doku.php?id=solver:terminating_the_optimization_after_the_first_feasible_solution&rev=1223686307&do=diff Q: Which GAMS or Cplex option can I use to make the solver stop quickly once a feasible solution has been found? On the GAMS level you can set optcr and optca to large numbers. If the bound and the value are on opposite sides of zero, this might result in larger values than one for optcr. For setting optca large enough you need some idea about the range of objective values. In GAMS/CPLEX there is also an option called intsollim. Setting intsollim to 1 will terminate the optimization after t… text/html 2008-05-08T10:36:30-04:00 http://support.gams-software.com/doku.php?id=solver:the_input_file_...logdisj.scr_could_not_be_opened._lmlboa&rev=1210257390&do=diff This error pops up if you are using GAMS scaling (.scale) in connection with LMLBOA and CPLEX as sub solver for MIP problems. Only the distributions 22.6 and 22.7.1 are affected by this bug. If your system is still under under maintenance, please download the current GAMS version from our download server. text/html 2010-02-23T06:36:06-04:00 http://support.gams-software.com/doku.php?id=solver:the_solver_option_file&rev=1266924966&do=diff If you’ve not used an option file before, create a new file of that name and save it in the same directory as your GAMS source file. If you want the options to appear in your GAMS source code, you can use the following code fragment as an example (with an option file for baron as example): text/html 2007-07-31T07:14:16-04:00 http://support.gams-software.com/doku.php?id=solver:too_many_iterations&rev=1185880456&do=diff Q: I recently ran a small program with two nonlinear constraints, but it terminated before it converged at 200 major iterations, even though I set options iterlim = 2000. The reason for termination was ”major iterations terminated before convergence/ Exit too many iterations”. The problem was infeasible when it exited. text/html 2010-03-09T07:36:02-04:00 http://support.gams-software.com/doku.php?id=solver:using_lazy_constraints_or_user_cuts&rev=1268138162&do=diff GAMS/CPLEX does currently not support lazy constraint or user cuts. We have not seen convincing examples where these Cplex features make a significant difference in performance. We understand that this might be a chicken and egg problem, therefore we introduced some time ago a way to experiment with these Cplex features. This is not for production models but to see if these features improve performance for your model. text/html 2008-10-24T09:22:23-04:00 http://support.gams-software.com/doku.php?id=solver:using_the_mipstart_-_option&rev=1224854543&do=diff Q:How can I use an initial vector of binary variables in a MIP problem with CPLEX (providing a good start point)? The CPLEX manual says about “mipstart 1” but doesn't say how you provide the initial binary solution but only the capability of using this option. Should I use the ”.L” field before solve command (e.g. X.L(J1) = 1) or is there another way? text/html 2008-10-29T10:34:42-04:00 http://support.gams-software.com/doku.php?id=solver:warning_700&rev=1225290882&do=diff Warning 700 means: One or more almost zero elements are specified in sparse input data. GAMS never passes a 0 in the constrain matrix down to the solver, but MOSEK (as other solvers) projects values smaller than some tolerance to zero. The default tolerance is 1e-12 (range from 1e-6 to 1e-16). So any matrix coefficient smaller than 1e-12 will be treated as 0 and MOSEK will issue a warning 700 that this happened. text/html 2007-07-31T05:42:52-04:00 http://support.gams-software.com/doku.php?id=solver:what_do_i_do_if_osl_crashes_without_warning&rev=1185874972&do=diff Q:GAMS/OSL runs my MIP problem for a while. But once the branch and bound search has gone on for a while, GAMS/OSL suddenly crashes and GAMS cannot recover the solution. What can I do ? The most likely cause of the GAMS/OSL crash is due to its memory estimate being wrong. In such a case, you will probably find lines in the log similar to the following: text/html 2007-07-31T07:04:26-04:00 http://support.gams-software.com/doku.php?id=solver:what_does_a_in_the_first_position_of_a_log_line_mean&rev=1185879866&do=diff A plus (+) following the iteration number denotes reporting while solving a probing (as opposed to a relaxation) subproblem of the corresponding node. A star (*) in the first position of a line indicates that a better feasible solution was found. text/html 2007-09-18T09:59:03-04:00 http://support.gams-software.com/doku.php?id=solver:what_is_optca_optcr&rev=1190123943&do=diff All algorithms in GAMS for solving MIPs are LP based branch-and-bound algorithms. While the algorithm runs it maintains two very important numbers: “best estimate” and “best integer”. The “best integer” is the best solution that satisfies all integer requirements found so far. The “best estimate” provides a bound for the optimal integer solution. text/html 2007-11-06T05:44:04-04:00 http://support.gams-software.com/doku.php?id=solver:what_is_the_meaning_of_the_marginal_prices_in_the_gams_output_for_a_mip&rev=1194345844&do=diff GAMS reports consistent marginals across any MIP code used by GAMS. The idea is very simple. Fix all the discrete variables at the integer values and solve the resulting LP. Now you have a clear definition and can use it for your analysis. If you pick a different integer solution, you will, of course, get a different dual solution. text/html 2007-07-31T10:17:17-04:00 http://support.gams-software.com/doku.php?id=solver:why_do_various_mip_solvers_give_differing_answers_to_my_problem&rev=1185891437&do=diff The MIP solver xxx gives an answer different to MIP solver yyy for my problem. Why does this happen ? The default value of the relative optimality tolerance (optcr)is 0.1. This means that, unless otherwise specified, the solver will stop when it finds a feasible integer solution within 10% of the global optimum. Various solvers can find different solutions that fall in this 10% gap. The user can override this and force the solver to find the global optimum by setting the optcr option to 0.0 usi…