Issue with callback best objective and heuristic solution (versions 9.1.2 and 9.0.0)

Hello,

I am applying Benders decomposition via callback to my problem. I noticed that I can generate heuristic solution from integer solution which is possibly better than the best objective found by Gurobi. However, when I query MIPSOL_OBJBST via callback, the value is higher than the value shown from the MILP Gurobi log. Here is my code that obtain best objective bound and heuristic solution (I insset the heuristic solution at the MIPNODE due to the limitation of my Gurobi version):

if (where == GRB.Callback.MIPSOL) :

    z_sol = model.cbGetSolution(model._z)
    a_sol = model.cbGetSolution(model._a)
    r_sol = model.cbGetSolution(model._r)
    e_sol = model.cbGetSolution(model._e)
    eta_sol = model.cbGetSolution(model._eta)

best_obj = model.cbGet(GRB.Callback.MIPSOL_OBJBST)

true_z_value,r_sol, a_sol,e_sol = BendersMPnew.a_r_true_value_mip_z(z_sol, r_sol, a_sol,e_sol)
BendersMPnew.set_SP_rhs_and_solve(z_sol,r_sol,a_sol) # obtain true value of the subproblem
sp_obj = BendersMPnew.SP_optimal_obj
true_obj = true_z_value - sp_obj # obtain true value of objective function for current solution
eta_sol =sp_obj 

if((true_obj>best_obj) and ( true_obj > model._betterobj ) ):
print("best obj",best_obj)
print("new true value", true_obj)
model._betterobj = true_obj
model._betterobj_eta = eta_sol
model._betterobj_r = r_sol
model._betterobj_a = a_sol
model._betterobj_z = z_sol
model._betterobj_e = e_sol
model._betterobj_flag = 1

if(model._betterobj_flag ==1 and where == GRB.Callback.MIPNODE ):
    model.cbSetSolution(model._r, model._betterobj_r)
    model.cbSetSolution(model._a, model._betterobj_a)
    model.cbSetSolution(model._eta, model._betterobj_eta)
    model.cbSetSolution(model._e, model._betterobj_e)
    model.cbSetSolution(model._z, model._betterobj_z)
    model.cbUseSolution()

But in the MILP log, you can see the best objective bound is higher than the one shown in the log (Incumbent) and the heuristic solution is also not incorporated:

Changed value of parameter LazyConstraints to 1
   Prev: 0  Min: 0  Max: 1  Default: 0
Changed value of parameter Cuts to 0
   Prev: -1  Min: -1  Max: 3  Default: -1
Changed value of parameter PreCrush to 1
   Prev: 0  Min: 0  Max: 1  Default: 0
Changed value of parameter Threads to 1
   Prev: 0  Min: 0  Max: 1024  Default: 0
Changed value of parameter TimeLimit to 14400.0
   Prev: inf  Min: 0.0  Max: inf  Default: inf
Changed value of parameter Threads to 1
   Prev: 0  Min: 0  Max: 1024  Default: 0
Parameter LazyConstraints unchanged
   Value: 1  Min: 0  Max: 1  Default: 0
Parameter PreCrush unchanged
   Value: 1  Min: 0  Max: 1  Default: 0
Parameter OutputFlag unchanged
   Value: 1  Min: 0  Max: 1  Default: 1
Gurobi Optimizer version 9.1.2 build v9.1.2rc0 (mac64)
Thread count: 4 physical cores, 8 logical processors, using up to 1 threads
Optimize a model with 7832 rows, 7831 columns and 36261 nonzeros
Model fingerprint: 0x986f14aa
Variable types: 175 continuous, 7656 integer (7656 binary)
Coefficient statistics:
  Matrix range     [1e-02, 2e+01]
  Objective range  [1e+00, 9e+01]
  Bounds range     [4e-01, 2e+00]
  RHS range        [1e+00, 1e+04]
Presolve removed 1 rows and 0 columns
Presolve time: 0.04s
Presolved: 7831 rows, 7831 columns, 36174 nonzeros
Variable types: 1 continuous, 7830 integer (7656 binary)
Found heuristic solution: objective -11904.80000
Found heuristic solution: objective -5000.900000

Root relaxation: objective 1.323948e+03, 2014 iterations, 0.02 seconds

    Nodes    |    Current Node    |     Objective Bounds      |     Work
 Expl Unexpl |  Obj  Depth IntInf | Incumbent    BestBd   Gap | It/Node Time

     0     0 1323.36701    0   93 -5000.9000 1323.36701   126%     -    2s
best obj 959.1003024652846
new true value 959.1003024652857
H    0     0                     959.1003025 1323.36701  38.0%     -    3s
     0     0 1323.01674    0  349  959.10030 1323.01674  37.9%     -    3s
     0     2 1320.75216    0  261  959.10030 1320.75216  37.7%     -    8s
     5     7 1317.88206    4  242  959.10030 1319.82349  37.6%  70.2   10s
    27    29 1256.36028   20   33  959.10030 1319.82349  37.6%  42.4   15s
   105   107 1163.57470   82   41  959.10030 1319.82349  37.6%  22.7   20s
   269   271 1153.82935  232   75  959.10030 1319.82349  37.6%  13.0   25s
best obj 1028.9437666391068
new true value 1046.9773049115258
   388   390 1099.08559  343   14  959.10030 1319.82349  37.6%  10.5   30s
*  416   401             368     962.1459479 1319.82349  37.2%  10.3   31s
   425   410 1297.60517    9   78  962.14595 1319.05778  37.1%  11.9   35s
   437   422 1274.53888   18   87  962.14595 1319.05778  37.1%  13.0   40s
   451   436 1223.80204   28  100  962.14595 1319.05778  37.1%  13.7   46s
   459   444 1210.32124   33   29  962.14595 1319.05778  37.1%  13.9   50s
   485   470 1232.14514   56  150  962.14595 1319.05778  37.1%  13.5   55s
   513   496 1191.31886   92  351  962.14595 1319.05778  37.1%  13.2   60s
   514   497 1133.02900  217  261  962.14595 1319.05778  37.1%  13.2   66s
   516   498 1188.53743  106  371  962.14595 1319.05778  37.1%  13.1   70s
   518   500 1316.72038    5  371  962.14595 1319.05778  37.1%  13.1   77s
   519   502 1319.05778   14  376  962.14595 1319.05778  37.1%  29.3   82s
   528   508 1290.80276   18   28  962.14595 1319.05778  37.1%  30.5   85s
   540   516 1261.07762   24   50  962.14595 1319.05778  37.1%  31.2   91s
   546   520 1248.39942   27   50  962.14595 1319.05778  37.1%  31.3   95s
   564   532 1204.12531   36   29  962.14595 1319.05778  37.1%  30.8  100s
   576   540 1180.07751   42   21  962.14595 1319.05778  37.1%  30.8  106s
   588   548 1186.74079   48   38  962.14595 1319.05778  37.1%  30.4  111s
H  596   526                     962.7274479 1319.05778  37.0%  30.2  115s
   622   545 1213.47184   65  200  962.72745 1319.05778  37.0%  29.7  120s
   664   573 1160.13732   86   32  962.72745 1319.05778  37.0%  28.8  125s
H  700   572                     966.2164479 1319.05778  36.5%  28.3  128s
   726   591 1167.75969  117   33  966.21645 1319.05778  36.5%  27.5  130s
   778   626 1192.73648  143   30  966.21645 1319.05778  36.5%  26.8  135s
   828   659 1129.35955  168   29  966.21645 1319.05778  36.5%  25.8  140s
best obj 1107.4475000000011
new true value 1118.879898831533
H  856   653                     979.0094479 1319.05778  34.7%  25.3  142s
H  882   648                    1013.8994479 1319.05778  30.1%  24.8  144s
   892   657 1159.05916  200   60 1013.89945 1319.05778  30.1%  24.6  145s
   928   667     cutoff  214      1013.89945 1319.05778  30.1%  23.9  150s
   936   668 1306.38452   18   29 1013.89945 1319.05778  30.1%  24.5  155s
   947   676 1275.93971   23   41 1013.89945 1319.05778  30.1%  24.4  160s
   961   685 1272.69480   30   41 1013.89945 1319.05778  30.1%  24.3  165s
   983   700 1262.98588   41   41 1013.89945 1319.05778  30.1%  24.1  170s
  1005   714 1248.63478   52   34 1013.89945 1319.05778  30.1%  24.0  175s
  1016   722 1242.73775   58   66 1013.89945 1319.05778  30.1%  24.1  180s
  1031   732 1192.46979   65   29 1013.89945 1319.05778  30.1%  23.9  185s
  1039   737 1204.70664   69   29 1013.89945 1319.05778  30.1%  23.8  190s
  1053   746 1187.03236   76   29 1013.89945 1319.05778  30.1%  23.8  195s

I faced this issue in both 9.1.2 and 9.0.0 Gurobi versions. I am aware that Gurobi 9.5 accept the heuristic solution in MIPSOL as well, but I cannot move to other version due to limitation in installing the new version in our computer server.

Please let me know if you spot any issue in my code or have any suggestion. 

Thank you very much.

Regards,

Kianoush