Scheduling
Series
1
Part 1 The World of Scheduling Problems
2
Part 2 Scheduling with CP-SAT
3
Part 3 When Scheduling Meets Reality
4
Part 4 Scheduling in the Wild
Scheduling Series · Part 4

Scheduling in the Wild

The mathematics is the same. The constraints are not. Six industries, six versions of the same problem — and what each one teaches us.

The pattern: every domain frames scheduling differently — different objectives, different constraints, different failure costs. But underneath each one sits the same combinatorial core: allocate limited resources to tasks across time. Understanding each domain means understanding what they traded away to make the problem tractable.
scroll to explore

Six domains, one problem

Scheduling is everywhere — but each industry has its own language for it, its own constraints that make it hard, and its own catastrophic failure mode when things go wrong. Click any industry to dive in.

🏭
Manufacturing
Jobs through machines. Makespan. Setup times. The original scheduling problem.
Job Shop · Flow Shop
🏥
Healthcare
Nurses, surgeries, and beds. Fairness, safety, and the working time directive.
Rostering · OR Scheduling
✈️
Aviation
Crew pairings, tail assignments, and tight regulations. Cascading delays cost millions.
Crew Pairing · Tail Assignment
🚚
Logistics
Thousands of deliveries, time windows, vehicle capacity, and traffic. The VRP family.
VRP · VRPTW
🛒
Retail & Workforce
Demand-driven shift planning. Cover the peaks, honour preferences, don't over-spend.
Shift Planning · Rostering
📐
Project Management
Activity networks, precedence constraints, resource pools. The RCPSP family.
RCPSP · Critical Path

What makes each one hard

Select an industry above — or click any tab below — to explore the constraints, scale, and failure modes that define scheduling in each domain.

Manufacturing Scheduling

Manufacturing scheduling is where the field was born — and it remains one of the hardest instances in practice. The classical Job Shop Problem (JSP) asks: given n jobs, each requiring a sequence of operations across m machines, find a schedule that minimises total completion time (makespan). Real factories layer on top: sequence-dependent setup times (cleaning a machine between products), maintenance windows, shift patterns, material dependencies, and multi-plant routing.

Machine capacity Operation sequences Setup times Due dates / tardiness Shift patterns Material availability Maintenance windows Worker skills
10–500
Jobs typical
5–50
Machines
NP-hard
Complexity
When it fails

Bottleneck machines sit idle waiting for upstream operations while urgent jobs queue behind lower-priority work. A single late delivery cascades into penalty costs, customer complaints, and stock outs downstream. In automotive plants, a 1-hour unplanned stoppage can cost £50,000 or more.

Job Shop — Gantt Chart
M1
J1·Op1
J2·Op2
J3·Op3
M2
J2·Op1
J1·Op2
J4·Op1
M3
J3·Op1
J1·Op3
J2·Op3
M4
J4·Op2
J3·Op2
J1·Op4
0255075100%

Each row is a machine; each coloured block is one operation from a job. The solver must find an ordering that avoids machine conflicts while minimising the end of the last block.

Typical Approach

Constraint Programming (CP-SAT) or MILP for medium instances. Metaheuristics (tabu search, genetic algorithms) for large plants where exact solvers time out. Rolling-horizon re-scheduling when new urgent jobs arrive.

Healthcare Scheduling

Healthcare scheduling is uniquely human — the constraints aren't just mathematical, they're ethical. Nurse rostering must balance legal requirements (the Working Time Directive caps duty at 48 hrs/week with minimum 11-hour rests), clinical safety (minimum nurse-to-patient ratios, right skill mix), and fairness (equitable weekend load, respecting preferences). Operating room (OR) scheduling must sequence surgeries across limited theatres while accounting for surgeon availability, equipment sterilisation, and unpredictable overruns.

48-hr weekly limit 11-hr rest between shifts Skill-mix requirements Minimum cover per shift Weekend fairness Staff preferences Annual leave Consecutive shift limits
20–500
Nurses per unit
4–13 wk
Planning horizon
2–3 days
Manual time saved
When it fails

Understaffed night shifts create patient safety risks. Over-reliance on agency staff costs 3–4× base salary. Unfair rosters drive burnout and resignations — a particular crisis in post-pandemic healthcare systems where every nurse retained matters.

Nurse Roster — One Week
Mon
Tue
Wed
Thu
Fri
Sat
Sun
N1
D
D
·
·
E
E
·
N2
N
N
N
·
·
D
D
N3
E
·
D
D
D
·
N
N4
·
E
E
N
N
·
·
N5
D
D
·
E
·
N
N
D — Day N — Night E — Evening

The solver must ensure every shift has enough cover, no nurse works consecutive night+day shifts, and weekends are distributed fairly across the 4-13 week horizon.

Typical Approach

ILP or CP-SAT with a soft/hard constraint hierarchy. Hard constraints encode legal requirements; soft constraints encode preferences and fairness objectives. Commercial solutions include Optima Nursing, Allocate, and custom CP-SAT models.

Aviation Scheduling

Aviation scheduling is among the most complex operational planning problems in existence — and one of the most consequential. It happens in three linked stages: fleet assignment (which aircraft type flies each route), crew pairing (building multi-day trip sequences that start and end at a crew base, respecting FAA/EASA duty-time rules), and crew rostering (assigning pairings to individual crew members for the monthly period). A single scheduling error can propagate through an entire network — grounding aircraft, stranding crew, and triggering millions in recovery costs.

FAA/EASA duty time limits Minimum rest periods Aircraft type ratings Crew base positioning Turnaround times Maintenance checks Seniority / bidding Disruption recovery
1,000s
Flights / day
100s
Crew per airline
$M/hr
Cost of disruption
When it fails

A single delayed flight cascades: crew go "out of position," aircraft sit at the wrong airport, and the disruption ripples through the network for hours. Southwest Airlines' December 2022 meltdown — partly attributed to scheduling system failures — cost over $800M and stranded 2 million passengers.

Crew Pairing — 3-Day Trip
Pilot A
LHR→CDG
CDG→FCO
FCO→LHR
F/O B
LHR→CDG
CDG→FCO
FCO→LHR
A320
G-EUXY
G-EUXY
G-EUXY
Day 1Day 2Day 3
Required rest break (min. 10 hrs) — hard constraint, cannot be violated
Typical Approach

Set-partitioning formulation solved with column generation + branch-and-price. MILP cores are standard. Commercial platforms: Sabre AirCentre, Jeppesen, Carmen. Real-time disruption recovery uses heuristics — exact methods are too slow for the 20-minute decision window.

Logistics Scheduling

Logistics scheduling is fundamentally the Vehicle Routing Problem (VRP) — assign a set of deliveries to a fleet of vehicles and find the optimal route for each vehicle. In practice, the "base" VRP is almost never enough. Most real problems are VRPTW (with time windows), VRPC (with vehicle capacity), or both. Add multiple depots, mixed fleet types, driver shift hours, real-time traffic, and same-day delivery windows — and you have the problem that drives billions in annual software investment. Last-mile delivery alone accounts for 53% of total shipping costs.

Vehicle capacity Delivery time windows Driver hours (WTD) Fleet type mix Depot constraints Real-time traffic Failed delivery re-routing Sustainability targets
100s–M
Deliveries / day
53%
Of shipping cost
NP-hard
Complexity
When it fails

Drivers return over capacity or run out of time mid-route. Deliveries miss time windows, triggering failed-delivery fees and redelivery costs. Poor routing inflates fuel spend and driver overtime. At scale, even a 1% routing improvement translates to millions in annual savings.

VRP — Three Vehicle Routes
DEPOT A B C D E F G H I J

Three vehicles, each with a capacity limit and shift-end deadline, collectively cover 9 stops. The solver must partition stops across vehicles and sequence each route to minimise total distance while respecting time windows.

Typical Approach

OR-Tools' routing library (Google), NVIDIA cuOpt for GPU acceleration at scale, or commercial platforms like Routific and OptimoRoute. Large instances use LNS (Large Neighbourhood Search) or guided local search — exact methods don't scale past ~30 stops.

Retail & Workforce Scheduling

Workforce scheduling in retail and service industries is demand-driven: the number of staff needed at any hour depends on forecasted footfall, call volume, or throughput — and that forecast changes weekly. The problem is to build a set of shifts, assign employees to them, and ensure coverage meets demand at every time interval, while honouring labour law (maximum weekly hours, minimum rest, break entitlements), union agreements, and individual preferences. Unlike manufacturing, the "product" is service — and the cost of failure is invisible until a customer gives up and leaves.

Demand coverage per hour Max weekly hours Minimum rest between shifts Break entitlements Contract type (FT/PT) Employee availability Role / skill requirements Overtime cost control
20–500
Staff per store
4–13 wk
Planning window
15–30%
Labour cost saved
When it fails

Over-scheduling during quiet periods burns payroll unnecessarily. Under-staffing during peaks creates queues, abandoned carts, and poor customer experience scores. Unfair schedules drive turnover — and in low-margin retail, replacing a trained employee costs 50–200% of their annual salary.

Demand Coverage — Mon–Sun
gap M T W Th F Sa Su
Demand Scheduled staff Undercoverage
Typical Approach

ILP for medium instances; decomposition (first build shifts, then assign) for large ones. Commercial platforms dominate: Kronos (UKG), Deputy, Rotageek, When I Work. Many use ML to forecast demand, then OR to solve the scheduling problem.

Project Scheduling

Project scheduling formalises as the Resource-Constrained Project Scheduling Problem (RCPSP): given a set of activities with precedence relations and resource requirements, find a start time for each activity that minimises the project makespan (or meets a deadline), subject to resource availability. Real projects add: activity durations that are uncertain, multi-mode activities (you can allocate more resources to speed up), time-varying resource availability, multi-project shared resource pools, and the ever-present scope change.

Precedence relations (AON) Resource availability Milestone deadlines Activity mode choices Uncertainty in durations Multi-project sharing Budget constraints External dependencies
50–1k
Activities
10–100
Resource types
NP-hard
Complexity
When it fails

Critical path activities slip, compressing buffers downstream until deadlines are missed. Resource conflicts stall multiple workstreams simultaneously. The CHAOS Report consistently shows ~70% of software projects delivered late or over budget — with poor scheduling cited as a leading cause.

RCPSP — Precedence + Resources
A1
Design
A2
Spec
A3
Develop A
A4
Develop B
A5
Develop C
A6
Integration
A7
Deploy
Wk 0Wk 3Wk 6Wk 9Wk 12
A5 delayed by resource conflict — developers still occupied on A4
Typical Approach

Critical Path Method (CPM) and PERT for deterministic baselines. CP-SAT or MILP for resource-constrained variants. Monte Carlo simulation for stochastic duration modelling. Tools: Microsoft Project, Primavera P6, or custom Python + OR-Tools models.

When it actually shipped

Two fully documented, award-winning deployments — UPS in last-mile logistics and Netherlands Railways in crew scheduling — that show what happens when scheduling theory meets operational scale, institutional resistance, and the messy constraints of real infrastructure.

UPS · Logistics · 2012–present
ORION: 55,000 Drivers, One Algorithm

UPS's On-Road Integrated Optimization and Navigation system is arguably the largest-scale VRP deployment in history. It re-routes 55,000 US drivers every morning using a blend of network optimisation, machine learning, and constraint satisfaction — in time for the first shift.

UPS operates ~55,000 drivers making 16 million deliveries per day across the US. Before ORION, routes were planned by experienced dispatchers using local knowledge and fixed templates. The question wasn't whether routes could be improved — it was whether you could prove it, systematically, and deploy it at scale across every UPS terminal simultaneously.

1Model each driver's day as a VRPTW: time-windowed stops, a capacity-limited vehicle, and a shift-end constraint.
2Feed in historical traffic data and telemetry to build real-world travel time matrices between every stop pair.
3Solve each driver's route using a combination of savings algorithms and large neighbourhood search — fast enough to replan overnight as new pickups are added.
4Layer in the "right-turn bias" heuristic: prefer right turns (in US road networks) to avoid crossing traffic — a domain insight that no pure algorithm would discover alone.
100M
Miles saved / year
10M
Gallons fuel saved
$300–400M
Annual savings

Domain knowledge still wins. UPS's engineers spent years encoding knowledge that experienced dispatchers carry implicitly — right-turn biases, local access restrictions, building-specific delivery norms. The solver provided optimality guarantees; the domain experts provided the constraints that made those guarantees meaningful. Ten years on, ORION continues to evolve, integrating real-time data and dynamically re-optimising routes mid-day.

Netherlands Railways (NS) · Transportation · Edelman Award 2008
The OR Revolution: Rescheduling an Entire National Railway

In the early 2000s, NS faced a crisis: a new timetable had created an unworkable crew scheduling problem that planners could no longer solve by hand. What followed became one of the most celebrated OR deployments in history — a complete re-architecture of how Dutch railway operations are planned, from timetable to crew.

NS operates roughly 400,000 train kilometres per day across the Dutch network, staffed by around 2,800 train drivers and 3,000 conductors. When NS introduced a new demand-driven timetable in 2007, the resulting crew scheduling problem broke every existing manual and semi-automated planning process. The timetable created complex transfer patterns — crew had to be repositioned across the network constantly — and the scale made the combinatorial explosion intractable with prior methods.

Crucially, NS treated timetabling, rolling stock scheduling, and crew scheduling not as three independent problems solved sequentially, but as a single integrated planning challenge. Each stage's output fed the next, and the whole pipeline was redesigned around OR from the ground up.

1CADANS — a cyclic timetabling system using integer programming to build a timetable that was both passenger-friendly and operationally feasible for rolling stock and crew.
2ROSA — rolling stock scheduling, assigning train units to services to minimise fleet size while meeting all capacity requirements.
3TURNI — crew scheduling, using a set-partitioning formulation solved with column generation and branch-and-price. Generated feasible duty sequences for all 2,800 drivers across all depots and shift patterns.
4Real-time disruption management — a separate layer that re-optimises crew and rolling stock assignments within minutes when delays or cancellations occur, replacing the previous human dispatcher model.
€80M
Annual savings
+3%
Punctuality improvement
2008
Edelman Award winner

The biggest insight from NS wasn't a solver breakthrough — it was an architectural one. Previous railway planning systems solved timetabling, rolling stock, and crew scheduling separately, in sequence, with each handoff introducing infeasibilities the next stage had to absorb. NS's team, working with researchers at CWI and Erasmus University Rotterdam, proved that integrating all three stages — even imperfectly — produced solutions that were fundamentally better than any sequential approach could reach. The timetable that emerged was one no human planner had imagined, because it was shaped from the start by what was actually feasible for crews and rolling stock to execute.

The published paper — Kroon et al. (2009), Interfaces — remains one of the clearest accounts of what it takes to move OR from theory into live national infrastructure.

Side by side

The same five dimensions across all six industries. The patterns that emerge reveal why some domains converged on CP-SAT, others on metaheuristics, and a few on decomposition methods that mix both.

Industry Problem Family Optimisation Goal Primary Solver Approach Typical Horizon Failure Cost
🏭 Manufacturing Job Shop / Flow Shop Makespan, tardiness CP-SAT MILP Tabu search Shift / day £50k+/hr stoppage
🏥 Healthcare Nurse rostering / OR sched. Fairness, cost, coverage CP-SAT ILP 4–13 weeks Patient safety, burnout
✈️ Aviation Crew pairing / rostering Cost, regulatory compliance Column gen. B&P Heuristics (recovery) Monthly cycle $M per disruption
🚚 Logistics VRP / VRPTW Distance, time, fuel LNS Guided LS OR-Tools / cuOpt Daily Late fees, redelivery cost
🛒 Retail / Workforce Shift planning / rostering Coverage, cost, fairness ILP Decomposition Commercial WFM 4–8 weeks Service failures, turnover
📐 Project Mgmt. RCPSP / CPM Makespan, budget CP-SAT MILP Monte Carlo Weeks–years Deadline penalties, overruns
The Pattern Underneath

Every domain above is a variation of the same core problem — allocate limited resources to tasks across time. What makes each hard is the constraint structure: aviation has strict regulatory cliffs (violate rest rules and the whole schedule is invalid), logistics has combinatorial route explosion, healthcare has soft fairness constraints that are genuinely hard to formalise. The solver choice follows from the constraint structure: CP-SAT excels where constraints are rich and interlocking; metaheuristics dominate where scale defeats exact methods; decomposition wins where the problem has natural sub-problems that can be solved independently and then composed.

Sources & further reading

Organised by domain — industry reports, Edelman Award deployments, vendor documentation, and regulator publications. Tags indicate the type of source.

Industry-wide
[1] INFORMS. (annual). Franz Edelman Award for Achievement in Operations Research and the Management Sciences. informs.orgEdelman — The industry's most rigorous real-deployment award. Past winners span all six domains covered here: UPS (logistics), New Zealand Police (workforce), Netherlands Railways (crew scheduling), and more. Each published in Interfaces.
[2] McKinsey Global Institute. (2022). Delivering the Great Supply Chain Reset. McKinsey & Company.report — Cross-industry analysis of scheduling and planning failures exposed by COVID-19 disruptions, with quantified cost estimates by sector.
[3] Gartner. (2024). Magic Quadrant for Supply Chain Planning Solutions. Gartner Research.report — Annual vendor evaluation covering the major commercial scheduling and planning platforms across manufacturing, logistics, and retail. Useful for understanding the commercial landscape.
🏭 Manufacturing
[4] Siemens Digital Industries Software. (2023). Opcenter Advanced Planning and Scheduling: Product Overview. Siemens AG. plm.automation.siemens.comvendor — The APS market leader for discrete manufacturing. Documents how sequence-dependent setup times, multi-resource operations, and shift calendars are modelled in practice.
[5] Pinedo, M., Yen, B.P.C. (1997). On the design and development of object-oriented scheduling systems. Annals of Operations Research, 70, 359–378.deployment — Unusually practical paper: describes the architecture decisions behind building a real scheduling system for a chemical plant, including the gap between theory and what the plant floor actually needed.
[6] McKinsey & Company. (2021). Industrial Operations: Next-Generation Manufacturing Planning. McKinsey Operations Practice.report — Documents the state of advanced scheduling adoption across discrete and process manufacturing; estimates 10–20% throughput gains achievable through APS deployment.
🏥 Healthcare
[7] NHS England. (2023). NHS Long Term Workforce Plan. NHS England Publications. england.nhs.ukregulator — The official strategic framework for NHS workforce planning over a 15-year horizon, including nurse supply, rostering policy, and digital workforce tools.
[8] NHS Improvement. (2021). Developing Workforce Safeguards: Supporting NHS Providers to Deliver High Quality Care Through Sustainable Staffing. NHS Improvement. england.nhs.ukregulator — Policy document establishing minimum safe staffing levels and shift constraints that feed directly into the hard constraints of any nurse rostering model.
[9] Rotageek. (2023). NHS Workforce Scheduling: From Manual Rosters to AI-Assisted Planning. Rotageek Ltd. rotageek.comvendor — Deployment white paper from the UK workforce scheduling platform used across NHS and retail. Includes before/after metrics from acute trust rostering deployments.
✈️ Aviation
[10] Kroon, L., Huisman, D., Abbink, E., Fioole, P.J., Fischetti, M., Maróti, G., Schrijver, A., Steenbeck, A., Ybema, R. (2009). The new Dutch timetable: The OR revolution. Interfaces, 39(1), 6–17.Edelman 2008 — Primary source for the Netherlands Railways case study. Written by the NS/CWI/Erasmus team; describes CADANS (timetabling), ROSA (rolling stock), and TURNI (crew scheduling) and their integration into a single OR-driven planning pipeline. Documents approximately €80M in annual savings and the design of the 2007 Dutch national timetable.
[11] IATA. (2023). Airline Operations Control: Best Practices and Technology Guide. International Air Transport Association. iata.orgregulator — IATA's operational guidance on crew scheduling, disruption management, and recovery procedures. Documents the real-world regulatory constraints that make aviation scheduling uniquely constrained.
[12] Boeing / Jeppesen. (2022). Jeppesen Crew Pairing and Rostering: Solution Overview. Jeppesen Systems. jeppesen.comvendor — Product documentation from the market-leading airline crew scheduling platform. Used by over 100 airlines globally; describes the column generation approach in plain-language terms.
[13] US Department of Transportation / Inspector General. (2023). Southwest Airlines December 2022 Operational Meltdown: Causes and Systemic Factors. DOT OIG Report.regulator — The official post-mortem on the Southwest meltdown that stranded 2 million passengers. Identifies crew scheduling system failures and the cascade dynamics that real-time disruption recovery must handle.
🚚 Logistics
[14] Holland, A., Levis, J., Nuggehalli, R., Santilli, B., Winters, J. (2017). UPS optimizes delivery routes. Interfaces, 47(1), 8–23.Edelman 2016 — The primary source for the ORION case study. Written by the UPS operations research team; describes system architecture, right-turn heuristic, phased rollout, and verified savings across 55,000 drivers.
[15] World Economic Forum. (2020). The Future of the Last-Mile Ecosystem. WEF White Paper. weforum.orgreport — Cross-industry analysis of last-mile delivery economics, including the 53% cost-of-shipping statistic. Documents the technology and regulatory landscape shaping VRP deployments globally.
[16] McKinsey & Company. (2022). Same-Day Delivery: The Next Frontier in US Retail. McKinsey Operations Practice.report — Quantifies the scheduling complexity introduced by same-day delivery commitments: tighter time windows, dynamic re-routing, and the trade-off between route efficiency and on-time performance.
[17] NVIDIA. (2024). cuOpt: GPU-Accelerated Route Optimisation. NVIDIA Developer Documentation. developer.nvidia.com/cuoptvendor — Technical documentation for NVIDIA's GPU-native VRP solver. Includes benchmark comparisons against CPU solvers at 1,000–100,000 stop scales and integration guidance with fleet management systems.
🛒 Retail & Workforce
[18] UKG (Kronos) Workforce Institute. (2023). The State of Workforce Management. UKG Inc. ukg.com/workforce-institutereport — Annual survey of 3,000+ HR and operations leaders across retail, healthcare, and hospitality. Quantifies the cost of scheduling errors, turnover linked to unfair rosters, and adoption of automated WFM tools.
[19] Deloitte Insights. (2023). Retail Workforce Transformation: From Scheduling Burden to Competitive Advantage. Deloitte Consulting.report — Documents the shift from fixed-schedule to demand-driven rostering in large retail chains, with quantified labour cost savings and customer satisfaction improvements from automated scheduling deployments.
[20] New Zealand Police. (2011). INFORMS Edelman Award finalist. Optimizing Police Patrol Scheduling. Published in Interfaces.Edelman finalist — A rare public-sector workforce scheduling deployment study. Documented how CP-based shift scheduling reduced overtime by 20% while maintaining coverage levels — directly analogous to retail demand-driven rostering.
📐 Project Management
[21] PMI. (2024). Pulse of the Profession. Project Management Institute. pmi.orgreport — The industry's flagship annual survey of project performance. Source for statistics on schedule overruns, budget deviations, and the cost of poor planning across industries including construction, IT, and engineering.
[22] Standish Group. (2023). CHAOS Report: Project Success and Failure Rates. Standish Group International.report — Annual study tracking IT project delivery outcomes since 1994. The 70% late/over-budget figure cited in this article is drawn from this report. Provides longitudinal context for how scheduling quality affects project outcomes.
[23] Oracle. (2023). Primavera P6 Enterprise Project Portfolio Management: Technical Reference. Oracle Corporation. oracle.com/primaveravendor — The dominant project scheduling platform in construction and engineering. Documentation describes how CPM, resource levelling, and constraint-based scheduling are implemented at enterprise scale.
Tools & Open Source
[24] Google OR-Tools team. (2024). OR-Tools: Open Source Software for Combinatorial Optimization. Google LLC. developers.google.com/optimizationopen source — Documentation for the solver suite used throughout this series. The routing library (for VRP) and CP-SAT (for scheduling and rostering) are both covered. Source code at github.com/google/or-tools.
[25] Gurobi Optimization. (2024). Gurobi Optimizer: Reference Manual. Gurobi LLC. gurobi.comvendor — The commercial MILP solver most commonly used in aviation crew scheduling and large-scale manufacturing APS. Gurobi's case study library at gurobi.com/case_studies documents real deployments across all six industries covered here.
Scheduling
Series
1
Part 1 The World of Scheduling Problems
2
Part 2 Scheduling with CP-SAT
3
Part 3 When Scheduling Meets Reality
4
Part 4 Scheduling in the Wild