If we are solving a linear program and the problem is unbounded, then is it possible to get a primal ray which leads to the unboundedness? I found this pyscipopt.lp.getPrimalRay. But I did not find any examples on how to create a pyscipopt.lp object. Can someone suggest how to get a primal ray for a pyscipopt.scip.Model object.
An example of a model which leads to unboundedness -
from pyscipopt import *
model = Model("Example")
x = model.addVar("x")
y = model.addVar("y")
model.setObjective(-x -y)
model.addCons(2*x - y*y >= 0)
model.optimize()
sol = model.getBestSol()
print("x: {}".format(sol[x]))
print("y: {}".format(sol[y]))
There is a method in SCIP that gives you the values of the primal ray (if one exists) called SCIPgetPrimalRayVal which will give you the value for a certain variable. However, that method is not yet wrapped inside PySCIPopt.
You can easily do this yourself though (just do the same as all the methods in scip.pyx) for both SCIPgetPrimalRayVal and SCIPhasPrimalRay. And then it would be great if you could create a pull request over on the PySCIPopt github to extend the interface for everyone.