Mike Musty - Optimizing sales representative assignments using integer programming | PyData Vermont

Integer linear programming can efficiently solve sales rep assignment problems by maximizing expected profits while respecting assignment constraints

The problem involves assigning binary values (0 or 1) to represent whether a rep is assigned to a lead, with constraints ensuring each lead gets exactly one rep

Google’s ORtools library provides an accessible Python interface for solving these optimization problems through constraint programming

While brute force approaches become intractable with large numbers of leads/reps, specialized solvers can find optimal solutions efficiently

Two key constraints govern the problem:

• Each lead must be assigned exactly one sales rep
• Each rep can be assigned to at most one lead

The solver can handle thousands of leads/reps while finding globally optimal assignments that maximize total expected profit

Recent advances in constraint programming focus on manipulating feasible solutions rather than just optimizing objective functions

The approach helps prevent local optimization (like sending the “best” rep to each lead) in favor of better global outcomes

Integer constraints are essential since fractional assignments (splitting reps between leads) wouldn’t make practical sense

The framework can be adapted for similar assignment optimization problems beyond just sales rep scheduling