Integer Programming

In the last post, we used Gurobi’s hierarchical optimization features to compute the Pareto front for primary and secondary objectives in an assignment problem. This relied on Gurobi’s setObjectiveN method and its internal code for managing hierarchical problems. Some practitioners may need to do this without access to a commercial license. This post adapts the previous example to use HiGHS and its native Python interface, highspy. It’s also useful to see what the procedure is in order to understand it better. This isn’t exactly what I’d call hard, but it is easy to mess up.1 ...

One of the first technology choices to make when setting up an optimization stack is which modeling interface to use. Even if we restrict our choices to Python interfaces for MIP modeling, there are lots of options to consider. If you use a specific solver, you can opt for its native Python interface. Examples include libraries like gurobipy, Fusion, highspy, or PySCIPOpt. This approach provides access to important solver-specific features such as lazy constraints, heuristics, and various solver settings. However, it can also lock you into a solver before ready for that. ...

At a Nextmv tech talk a couple weeks ago, I showed a least absolute deviations (LAD) regression model using OR-Tools. This isn’t new – I pulled the formulation from Rob Vanderbei’s “Local Warming” paper, and I’ve shown similar models at conference talks in the past using other modeling APIs and solvers. There are a couple reasons I keep coming back to this problem. One is that it’s a great example of how to build a machine learning model using an optimization solver. Unless you have an optimization background, it’s probably not obvious you can do this. Building a regression or classification model with a solver directly is a great way to understand the model better. And you can customize it in interesting ways, like adding epsilon insensitivity. ...

Note: This post was updated to work with Python 3 and PySCIPOpt. The original version used Python 2 and python-zibopt. It has also been edited for clarity. As a followup to the last post, I created another SCIP example for finding Normal Magic Squares. This is similar to solving a Sudoku problem, except that here the number of binary variables depends on the square size. In the case of Sudoku, each cell has 9 binary variables – one for each potential value it might take. For a normal magic square, there are $n^2$ possible values for each cell, $n^2$ cells, and one variable representing the row, column, and diagonal sums. This makes a total of $n^4$ binary variables and one continuous variables in the model. ...

Note: This post was updated to work with Python 3 and PySCIPOpt. The original version used Python 2 and python-zibopt. It has also been edited for clarity. Back in October of 2011, I started toying with a model for finding magic squares using SCIP. This is a fun modeling exercise and a challenging problem. First one constructs a square matrix of integer-valued variables. from pyscipopt import Model # [...snip...] m = Model() matrix = [] for i in range(size): row = [m.addVar(vtype="I", lb=1) for _ in range(size)] for x in row: m.addCons(x <= M) matrix.append(row) Then one adds the following constraints: ...