Originally posted by: dominic2014

I wish to solve a quadratic programming problem with binary variables.

If I use the MIP solver then it runs forever without ever showing a gap. I suspect that this is because it is unable to find a feasible solution.

I know a feasible solution soln={+1,+1,...,+1} that is not optimal.

Can I start it there and will this help it to find a better solution.

I have tried to find the Python syntax for passing this solution to the MIP Start interface but have failed. 

Can anyone tell me what I need to do. I wish to pass the vector soln to the optimiser.

Isn't the reference documentation of Cplex.MIP_starts.add() good enough? It even has a small example. This (untested) code should add a MIP start with a 1 for each variable:

c.MIP_starts.add([range(c.variables.get_num()), [1] * c.variables.get_num()], cplex.MIP_starts.effort_level.auto)