How to get solver time from CPLEX when using the NEOS server through Pyomo?

Question

I've been using CPLEX on the NEOS server, via Pyomo, to solve a binary program I'm working on.

NEOS is amazing, but the documentation is somewhat lacking on the Pyomo side, so I haven't been able to find out how I can access the solve time for an optimisation run. If I use a local solver on my machine (CBC) I can access the solve time in the results, and I can see it from the solver output, neither of which I get with CPLEX on NEOS.

Does anyone know how to do this?

Thanks

Answer 1

I found it hard, using Pyomo, to get that kind of information since Pyomo's results object has a different structure depending on the solver that has been used.

To get around that, I consider that the ugly solution is still the most reliable one up to now, until someone shows me a better way.

Convert the results object that was returned by the solve() method into a String using str(results). Find the text 'Time: ' in this string. Add all characters after the 'Time: ' substring that corresponds to a number and stop when you reach the \n character. You now have the value you wanted.

Here is a function that I created and that I still use to get this kind of informations:

def get_info_from_results(results, info_string):
    i = str(results).lower().find(info_string.lower()) + len(info_string)
    value = ''
    while str(results)[i] != '\n':
        value = value + str(results)[i]
        i += 1
    return value

Using this function, you should get your solve time from your (let's call it) neos_results object doing this:

solver_solve_time = get_info_from_results(neos_results, 'Time: ')

Keep in mind at this point that the value returned by this function is still considered by Python as a String, so you might want to convert it to a number if you want to make calculations with it.