Debugging the Schedule

When a course isn’t being scheduled where you want it, finding the cause can seem daunting given all the interacting constraints and objectives. This guide provides a systematic process for identifying and resolving scheduling issues.

Example Problem

Suppose you want C-101 scheduled before noon, but it keeps getting placed at 2:40 PM. How do you fix this?

Step-by-Step Process

1. Add an Incentive Objective

First, ensure there’s an objective that encourages the optimizer to place C-101 before noon:

objectives = [
    # ... other objectives ...
    MinimizeClassesAfter("12:00", courses=["C-101"]),
]

Run the optimizer. If C-101 moves before noon, you’re done!

2. Identify the Blocking Constraint or Objective

If C-101 is still scheduled after noon, something with higher priority is preventing it. The blocker must be a constraint or an objective that appears before your incentive objective in the list.

To find it:

  • Review only the constraints and objectives above your incentive objective

  • Try moving the incentive objective higher in the priority list

  • Use bisection search to efficiently narrow down which objective is blocking: move your incentive to the middle of the list, check if it’s satisfied, then search the appropriate half

3. Resolve the Conflict

Once you’ve identified the blocking constraint or objective, consider:

  • Remove it if it’s no longer needed

  • Reduce its scope (e.g., exempt certain instructors or courses from the constraint)

  • Lower its priority by moving it after other objectives

Common Pitfall: Adding Constraints to “Make Room”

When a course won’t go where you want, it’s tempting to add constraints or force other courses into specific slots to “make room.”

This approach cannot help. Adding constraints to an optimization problem can never improve the optimal objective value. If there’s a feasible way to place C-101 before noon, the optimizer will find it automatically.

Instead, improve the schedule by:

  • Refining objectives to better match your goals

  • Removing constraints that are blocking better solutions