Software

Python tools for modeling and solving mixed-integer linear programs. Write models using natural Python expressions, with access to advanced features such as cut generation, MIP starts, and solution pools.

Cbc is an extensible, open-source mixed-integer linear programming solver. It combines branch-and-cut search, cutting planes, and heuristics, and can be used as a library or from the command line.

I have contributed to the solver and written a short guide to its command-line interface. The guide is an archived reference; consult the current project documentation for up-to-date options.

A collection of routines for building dense conflict graphs and separating clique, odd-hole, and wheel inequalities. Much of the cut-separation code is solver-independent, with implementations in C and C++.

The project explores exact and heuristic clique separation, including lifting to strengthen the generated inequalities.

Our team's school timetabling solver won ITC 2011. It uses hybrid local-search methods to build schedules under complex resource and availability constraints.

An archived research project, preserved for reference.