Problem with callback function

Awaiting user input

MARIO BASALLO

May 23, 2024 15:33 Edited

• 
• 

Hello

I am implementing a branch and Benders cut algorithm by using a callback function with lazy constraints. When I run my code an integer solution is found at the root node, and I add a (lazy) constraint at this node. However, after adding the constraint, the optimization is finished reporting an infeasible problem. This cannot be the case for the instance I am solving since I can solve the .lp file in which the lazy constraint is added manually, and I get an optimal solution from there.

A simplified version of my callback function is below. I am wondering if something is wrong with the implementation of the callback function. I would appreciate it if you could provide me with any suggestions regarding this.

///////////////////////////////////////////////////////////////////////////////////////////////////////

int __stdcall

mycallback(GRBmodel* lp,

void* cbdata,

int   where,

void* usrdata) {

 

double* xsol;

int* matind;

double* matval;

double DualObj;

int status = 0;

 

if (where == GRB_CB_MIPSOL) {

 

status = GRBgetintattr(lp, GRB_INT_ATTR_NUMVARS, &numcols);

if (status)

printf("Error code: %d\n", status);

 

xsol = (double*)malloc(numcols * sizeof(double));

status = GRBcbget(cbdata, where, GRB_CB_MIPSOL_SOL, xsol);

if (status)

printf("Error code: %d\n\n", status);

 

// Solve the dual sub problem

DualObj = dualSPBC(xsol);

 

// Introduce Benders' cut

int nzind, numnz;

numnz = FeasHubs.dim + yvars + 1;

matind = (int*)malloc(numnz * sizeof(int));

matval = (double*)malloc(numnz * sizeof(double));

nzind = 0;

for (int i = 0; i < FeasHubs.dim; i++) {

matind[nzind] = i;

matval[nzind] = -coefz[i];

nzind++;

}

for (int i = 0; i < yvars; i++) {

matind[nzind] = FeasHubs.dim + i;

matval[nzind] = -coefy[i];

nzind++;

}

matind[nzind] = FeasHubs.dim + yvars;

matval[nzind] = 1.0;

nzind++;

 

status = GRBcblazy(cbdata, numnz, matind, matval, GRB_LESS_EQUAL, sumk);

if (status)

printf("Error: %d\n", status);

free(matind);

free(matval);

free(xsol);

}

}

///////////////////////////////////////////////////////////////////////////////////////////////////////

 

Here is additional information regarding the solution obtained using the callback function:

Set parameter Username
Set parameter LogFile to value "mip1.log"
Academic license - for non-commercial use only - expires 2025-04-29
Set parameter TimeLimit to value 86000
Set parameter Threads to value 1
Set parameter MIPGap to value 1e-05
Set parameter IntFeasTol to value 1e-07
Set parameter LazyConstraints to value 1
Gurobi Optimizer version 11.0.1 build v11.0.1rc0 (win64 - Windows 11.0 (22631.2))

CPU model: Intel(R) Core(TM) i7-1065G7 CPU @ 1.30GHz, instruction set [SSE2|AVX|AVX2|AVX512]
Thread count: 4 physical cores, 8 logical processors, using up to 1 threads

Optimize a model with 1308 rows, 1274 columns and 2616 nonzeros
Model fingerprint: 0x9a5fa3e7
Variable types: 1 continuous, 1273 integer (1273 binary)
Coefficient statistics:
  Matrix range     [1e+00, 1e+00]
  Objective range  [1e+00, 2e+04]
  Bounds range     [1e+00, 4e+06]
  RHS range        [0e+00, 0e+00]

Explored 0 nodes (0 simplex iterations) in 0.03 seconds (0.00 work units)
Thread count was 1 (of 8 available processors)

Solution count 0

Model is infeasible
Best objective -, best bound -, gap -

User-callback calls 23, time in user-callback 0.03 sec
Optimization status: 3

 

Let me know if you need further information regarding my code or instances.

Thank you!

Regards,

Mario

Related to

• C
• Callbacks
• Bender's

0

• Official comment

  

  Simranjit Kaur

  • Gurobi Staff

  November 13, 2026 01:24

  This post is more than two years old. Some information may not be up to date. For current information, please check the Gurobi Documentation or Knowledge Base. If you need more help, please create a new post in the community forum, or try Gurobot, our chatbot interface offering instant, expert-level support.
• 

  Jaromił Najman

  • Gurobi Staff

  May 28, 2024 07:33

  Hi Mario,

  You are saying

  This cannot be the case for the instance I am solving since I can solve the .lp file in which the lazy constraint is added manually, and I get an optimal solution from there.

  But I see that your callback has been called 23 times, so there has been probably more lazy constraints added than only one. Could you please double check that?

  Moreover, could you please share the way how you generate the lazy constraint you added to the model manually? If possible, could you please share the log output of the run of the model with manually added lazy constraint?

  Best regards, 
  Jaromił

  0

Post is closed for comments.