ojdo / pyomo-example

"Hello World!" example for combining Pyomo (Coopr 3.5 or Pyomo 4) and Pandas for mathematical optimisation

Hello Pandas + Pyomo examples

This is a minimum example that demonstrates how Pandas and Pyomo (both version 3.5 [within Coopr] and version 4) are combined to yield a mathematical optimisation model.

The example is a Minimum-cost flow problem formulation of the task to find the Bacon number of a given actor/actress within the given social graph of Hollywood actors.

What is it about?

Technically, the example shows how to convert the conceptually simpler, self-contained bacon.mod (GNU MathProg problem) into a more complex triplet of input data (input.xlsx), model formulation and helper functions (bacon.py) and a run script (runme.py).

While for this example, the scaffolding code seems unnecessarily complex, it scales much better to larger (i.e. more variables, more constraints) models like urbs.

Requirements

• Python 2.7 or 3.5 (Anaconda)
• Pandas (included in Anaconda)
• xlrd (included in Anaconda)
• Either (old): Coopr 3.5.8787 (pip install coopr=3.5.8787) or (new) Pyomo 4 (pip install pyomo)
• GLPK

How to run

Setup

git clone https://github.com/ojdo/pyomo-example.git
cd pyomo-example

MathProg example (bacon.mod)

glpsol -m bacon.mod

Pyomo example (everything else in the folder)

python runme.py

For a better learning experience, one can also launch ipython and execute the lines of runme.py step by step, inspecting the created objects with one eye on the Pyomo Online Documentation:

import coopr.environ
from coopr.opt.base import SolverFactory
import bacon

data = bacon.read_excel('input.xlsx')
model = bacon.create_model(data)
prob = model.create()
optim = SolverFactory('glpk')
result = optim.solve(prob, tee=True)
prob.load(result)

All the work happens in the functions defined within bacon.py.

Copyright

Copyright (C) 2015 ojdo

This program is free software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.

You should have received a copy of the GNU General Public License along with this program. If not, see http://www.gnu.org/licenses/

pyomo-example's Issues

Performance bottleneck

pyomo-example/bacon.py

Line 94 in 6c29c4c

flow_balance += m.data['actor'].loc[a, 'sink-source']

As I had to learn the hard way in the larger model tum-ens/urbs#140, accessing single values using pandas .loc[...] in a loop is wasteful. Instead convert each parameter to a dictionary with fast access. It won't be noticeable here, but with millions of accesses, accessing DataFrames for single parameter values should be avoided.