Is it possible to access dual information related to the variable bounds in Pyomo? For constraints, you can declare a Suffix, but is there an equivalent for the variable bounds?
You can declare a suffix named rc (reduced cost) to obtain this from the following interfaces:
Xpress might also be on that list, but I have no way of verifying that.
The Gurobi and Cplex solvers for AMPL do not return this information as suffixes (and I don't know why), so you can not get these through the NL-file interface to these solvers in Pyomo.
Also, for Ipopt you can get this by declaring suffixes named ipopt_zL_out and ipopt_zU_out for the duals of the lower bounds and upper bounds, respectively. See this example for a better explanation.
The above list is only what I am aware of. There are likely other AMPL solvers that provide this information through suffixes, so you would be able to access that solution information through Pyomo's NL-file interface as long is you know the name of the suffix.
Update: Here is an example with gurobi:
import pyomo.environ as aml
model = aml.ConcreteModel()
model.x = aml.Var(bounds=(0,1))
model.o = aml.Objective(expr=model.x)
model.c = aml.Constraint(expr=model.x >= -1)
model.rc = aml.Suffix(direction=aml.Suffix.IMPORT)
gurobi = aml.SolverFactory("gurobi",
solver_io="lp")
results = gurobi.solve(model)
assert str(results.solver.termination_condition) == "optimal"
print(model.rc[model.x])
As I explained above, you can set solver_io in this example to "lp", "mps", or "python" with Gurobi.