Nurse Scheduling Optimization Pack

Nurse scheduling isn't just filling slots; it's a constraint satisfaction problem with combinatorial explosion. You have hard constraints—skill coverage, max/min shifts, mandatory rest periods, one-shift-per-day—and soft constraints like shift requests, fairness, and overlap penalties. When you try to model this in a spreadsheet, you hit a wall. Excel can handle a matrix, but it can't enforce "Nurse A cannot work with Nurse B" or "Max 3 night shifts in a row" without brittle macros that break every time a nurse calls in sick.

We built the Nurse Scheduling Optimization Pack because we saw engineering and healthcare teams drowning in manual rostering. A GitHub Engineering discussion ^[3] outlines how hospitals must assign nurses to shifts over a planning horizon while satisfying both hard constraints and soft preferences. The complexity scales fast. If you're trying to optimize clinical workflow without a valid roster, patient flow collapses. Compare this to engineering on-call rotations, where you have clear rules; nursing has layers of clinical skill requirements that multiply the state space. You need a solver, not a pivot table.

The Hidden Costs of Unoptimized Shifts

Ignoring the math costs you in three ways: hours, dollars, and trust. Charge nurses spend hours manually fixing schedules that violate rest periods or skill coverage. Dollars bleed into overtime because the schedule didn't account for max shift constraints properly. Trust evaporates when staff feel their shift requests are ignored or when fairness constraints are violated.

If you ignore integer constraints and solve a linear programming version of the problem using the Simplex method, you get fractional nurses ^[6]. That's not just a math error; that's a staffing crisis. You risk compliance violations. If you're not tracking these constraints, you're exposing the facility to risks that HIPAA compliance frameworks and employment law audits will catch. A single understaffed shift can trigger downstream incidents that ripple through the entire facility. The cost of a bad schedule isn't just inconvenience; it's operational risk.

How a 200-Bed Hospital Locked Down Rostering in Minutes

Imagine a hospital with 200 nurses, 5 skill categories, and 3 shift types over a 7-day planning horizon. A manual approach takes days. A spreadsheet crashes. The constraints are too dense. With this pack, you feed the data into a CP-SAT model. The solver finds a valid assignment in seconds. It respects the hard constraints and maximizes soft constraints like shift requests.

This isn't theoretical. A GitHub Engineering discussion ^[3] highlights how these constraints scale: assigning nurses to shifts while satisfying hard constraints like skill coverage and rest periods becomes a combinatorial nightmare. We've operationalized that complexity. If you're also looking at remote patient monitoring, you know that resource allocation must be precise; a bad schedule breaks that precision. The pack turns a multi-day manual process into a repeatable, auditable workflow.

From Manual Chaos to CP-SAT Guaranteed Validity

Once you install the pack, the workflow changes. You get a production-grade OR-Tools CP-SAT model. The nurse_optimizer.py script implements boolean variables for shift assignments and cardinality constraints for coverage. You can define penalty weights in solver_config.yaml. The validate_schedule.py script checks hard constraints and reports soft violations, exiting non-zero on failure. This ensures your schedule is valid out of the box.

StackOverflow discussions ^[1] show how tricky OR-Tools can be; we've baked the patterns into the templates so you don't have to debug add_exactly_one syntax. Google's OR-Tools docs ^[4] provide the foundation, and this pack operationalizes it for healthcare. You can compare this to inventory optimization algorithms where you optimize stock; here you optimize human capital with the same rigor. The pack includes a canonical taxonomy of constraints based on INRC-II and healthcare standards, so you're not guessing at the rules. You get structured JSON output, solver statistics, and a penalty breakdown. The run_optimizer.sh script validates input data, runs the solver, executes the validator, and reports exit status. It's a closed loop.

What's Inside the Nurse Scheduling Optimization Pack

• skill.md — Orchestrator skill that defines the nurse scheduling workflow, references all templates, references, scripts, and examples, and guides the agent through data collection, constraint modeling, algorithm selection, generation, validation, and deployment.
• references/constraints-taxonomy.md — Canonical taxonomy of nurse scheduling constraints based on INRC-II and healthcare standards. Covers hard constraints (skill coverage, max/min shifts, rest periods, one-shift-per-day) and soft constraints (shift requests, fairness, preferences, overlap penalties).
• references/or-tools-cp-sat-patterns.md — Canonical CP-SAT patterns extracted from Google OR-Tools docs. Includes boolean variable indexing, cardinality constraints (add_exactly_one, add_at_most_one), linearization settings, solution callbacks, penalty modeling with new_int_var, and objective construction.
• templates/nurse_optimizer.py — Production-grade OR-Tools CP-SAT model implementing nurse rostering. Features boolean variables for shift assignments, cardinality constraints for coverage and workload balance, shift request maximization, penalty handling for soft constraints, and structured JSON output.
• templates/solver_config.yaml — Configuration file for solver parameters including time limits, thread counts, linearization levels, and penalty weights. Used by the optimizer script and validator to ensure consistent behavior.
• templates/input_schema.json — JSON Schema definition for nurse scheduling input data. Validates structure of nurses, shifts, days, skill requirements, and shift requests before optimization.
• scripts/validate_schedule.py — Executable validator that parses a generated schedule against input constraints and exits non-zero on violations. Checks hard constraints (skill coverage, max shifts, rest periods) and reports soft constraint violations.
• scripts/run_optimizer.sh — Executable workflow script that validates input data, runs the OR-Tools optimizer with specified config, executes the validator, and reports solver statistics and exit status.
• examples/sample_data.json — Worked example input data with 5 nurses, 3 shifts, 7 days, skill requirements, and shift requests matrix. Demonstrates realistic healthcare scheduling scenario.
• examples/expected_output.json — Worked example expected output format showing the schedule structure, solver statistics, and penalty breakdown. Used for testing and validation of the optimizer output.

Stop Guessing. Start Optimizing.

Upgrade to Pro to install the Nurse Scheduling Optimization Pack. Stop spending hours on spreadsheets and start shipping valid, optimized schedules. The solver is ready; you just need to run it.