Open 4er4er4er opened 1 year ago
Changing as follows:
ampl: include multmip_pool__mp.run
Gurobi 10.0.2: sol:stub = multmip
sol:poolmode = 2
sol:poollimit = 5
Gurobi 10.0.2: optimal solution; objective 235625
331 simplex iterations
68 branching nodes
5 alternative solution(s)
with objective values 242025..235625
written to multmip1.sol..multmip5.sol.
suffix nsol OUT;
suffix npool OUT;
solve_result = solved
solve_result_num = 0
Gurobi 10.0.2: Alternative solution 1, objective 235625
solve_result = solved
solve_result_num = 0
Gurobi 10.0.2: Alternative solution 2, objective 237125
solve_result = solved
solve_result_num = 0
Gurobi 10.0.2: Alternative solution 3, objective 238225
solve_result = solved
solve_result_num = 0
Gurobi 10.0.2: Alternative solution 4, objective 239900
solve_result = solved
solve_result_num = 0
Gurobi 10.0.2: Alternative solution 5, objective 242025
solve_result = solved
solve_result_num = 0
The solve status of the alternative solutions is going to be that of the final solution.
The attached run file uses gurobiasl to find 5 MIP solutions. Then it reads the resulting .sol files one at a time. The
solve_result
andsolve_result_num
values are allsolved
and 0 as expected:When
option solver gurobiasl;
is replaced byoption solver gurobi;
the solutions are the same, and thesolve_result
andsolve_result_num
values reported by the solver are the same. But the values taken from the .sol files are incorrectly reported assolved?
and 100:(Note also that the messages about the solution pool from
gurobiasl
are longer -- but also more informative -- than the messages fromgurobi
.)