Capacitated Vehicle Routing Problem (CVRP) Guide 2026

Most routing tools answer one question — “what’s the shortest path?” That question is interesting in a math class. In a real plumbing or HVAC shop, the question is “what’s the shortest path my team can actually finish today?” Those are not the same problem. The capacitated vehicle routing problem — CVRP for short — exists to answer the second one.

This guide walks through CVRP for field service in plain English. The four capacity types every dispatcher tracks. The difference between hard and soft capacity constraints. The capacity-first architecture that produces schedules that survive past lunch. Multi-day CVRP. Equipment-based scheduling with shared bucket trucks and jetting trailers. And the real-world scenarios where capacity-aware routing earns its keep. Everything below matches what the live FieldCamp AI route optimization engine runs in production.

What the Capacitated Vehicle Routing Problem Is

The capacitated vehicle routing problem adds real-world limits to route optimization. Basic VRP finds the shortest routes. CVRP finds the shortest routes that your team can actually complete within shift hours, job caps, certifications, and equipment availability. “Capacitated” simply means “with capacity limits.”

Those limits include shift hours (a tech can work 8 hours, not 14), job count (6 jobs per day max, not 12), certifications (only EPA-certified techs handle refrigerant), equipment (one bucket truck shared across three crews), and physical capacity (a 600-gallon tank that fills up after 8 stops).

Every CVRP scenario has four components — starting location, customer demands, technician constraints, and travel times. At scale, the math gets brutal fast: a 50-job routing scenario has more possible solutions than atoms in the observable universe. Manual scheduling cannot evaluate even a fraction of these combinations while respecting capacity simultaneously. This is exactly why algorithmic routing is required — and why the broader AI dispatch scheduling stack is built around capacity-aware solvers, not pure shortest-path heuristics.

The Four Capacity Types Every Field Service Schedule Tracks

Capacity is not just how much a truck can carry. In field service, four distinct capacity types determine whether a schedule is executable — temporal, workload, skill, and equipment. All four must be evaluated simultaneously for the schedule to survive contact with the day.

1. Temporal Capacity — When a Tech Can Work

Shift hours, breaks, overtime limits, end-of-day constraints. An 8-hour shift from 7:30 AM to 5:00 PM with a 1-hour lunch means 7 productive hours, not 8. Without temporal checks, a system might schedule 12 hours of work in an 8-hour shift. The route looks efficient on paper. The last three jobs never happen.

What to configure: shift start/end, mandatory breaks, overtime allowance, end-of-day buffer for return drive, overtime cost multiplier (typically 1.5x). The control surface lives in FieldCamp AI Dispatcher docs.

2. Workload Capacity — How Much a Tech Should Handle

Max jobs per day prevents burnout and quality drift. A tech with 12 jobs rushes every appointment. A tech with 4 sits idle half the day. Different day types need different caps — a local day might allow 6, a travel-heavy day with 2-hour drives should cap at 3–4.

3. Skill Capacity — Who Can Do the Work

When only 2 out of 5 techs hold EPA 608 certification and 15 jobs require refrigerant handling, those 2 techs become a bottleneck. Skill capacity means the system can’t just assign the closest tech — it must assign a qualified one. A non-certified tech dispatched to a refrigerant job arrives, can’t legally perform the work, and leaves. The customer waits. The company faces compliance risk.

4. Equipment Capacity — What’s Available to Share

Expensive equipment shared across crews creates scheduling dependencies. Three tree crews sharing one bucket truck can’t all use it simultaneously. The CVRP solver sequences jobs with handoff buffers so Crew A finishes with the asset before Crew B needs it. Full handoff logic is covered below.

Equipment-Based Scheduling: Treat Bucket Trucks as First-Class

Equipment-based scheduling is the discipline of treating shared physical assets — bucket trucks, jetting trailers, crane arms, specialty diagnostic kits — as first-class scheduling resources, not as side notes attached to a technician. The bucket truck gets its own calendar, its own location, its own constraints.

In a non-equipment-aware schedule, the dispatcher assigns jobs to people and hopes the equipment lines up. In an equipment-aware schedule, every shared asset has its own row alongside the techs. A job that requires the bucket truck can only be assigned to a (technician, time-slot) pair where the truck is available, has been transported, and isn’t already booked.

When equipment moves between crews mid-day, three handoff patterns cover almost every realistic case:

• Direct. Tech A drives the equipment to Tech B. Best when crews are nearby. Adds drive time to Tech A’s day, which the optimizer accounts for.
• Depot. Equipment returns to base; the next tech picks it up at shift start. Best for overnight handoffs or geographically central depots.
• Meet halfway. Both techs travel to a midpoint. Best when crews are on opposite sides of the service area.

Buffer time between handoffs (typically 20–30 minutes) prevents idle waiting. Without that buffer, a 10-minute traffic delay on Tech A becomes a 10-minute idle penalty on Tech B’s clock. The system also needs an equipment registry (every shareable asset with current location), job-equipment requirements, handoff strategy preferences, buffer defaults, and certification overlays for which techs are trained on which gear. The full configuration story lives inside job submission settings.

Hard vs. Soft Capacity Constraints

Not all capacity limits are equal. Hard capacity constraints cannot be violated — the system eliminates any assignment that breaks them. Soft capacity constraints can flex with penalty scores. Misclassifying soft as hard is the single most common reason schedules feel rigid and unschedulable jobs pile up.

Hard capacity constraints include shift boundaries when overtime is off, skill certifications (EPA 608, electrical licenses, gas permits), mandatory breaks for labor compliance, and equipment availability (a bucket truck already assigned can’t be double-booked). A tech without EPA cert assigned to refrigerant work isn’t a preference violation — it’s a compliance violation.

Soft capacity constraints include preferred max jobs per day, preferred tech assignments, drive time preferences, workload balance targets. A tech getting a 7th job when the preferred max is 6 isn’t ideal — but it’s legal and feasible. The optimizer adds a penalty score and tries to avoid it.

The companion guide on AI dispatching algorithms covers the broader hard-vs-soft framework that wraps around capacity.

The Capacity-First Approach (Why Most Routing Tools Get It Backwards)

Most routing tools optimize first, then discover capacity violations. The dispatcher notices impossible assignments, manual fixes are required, the route is no longer optimal, and the cycle repeats every morning. A capacity-first architecture eliminates impossible assignments before optimization even starts. Every generated schedule is both efficient and executable.

The right pipeline runs four steps:

1. Rapid feasibility check. Eliminates capacity-violating assignments in milliseconds.
2. Constraint-based skeleton. Builds a schedule respecting every hard constraint.
3. Deep optimization. Improves routes while maintaining compliance.
4. Continuous monitoring. Reshuffles throughout the day without violating limits.

The key difference: impossible assignments are eliminated before optimization begins. No more generating beautiful routes that fall apart by 10 AM. This same capacity-first principle drives the entire AI Command Center — and is the foundation underneath accept/override workflows that keep dispatchers in control.

Multi-Day CVRP and Day-Chaining

When daily capacity is exceeded, traditional CVRP breaks. Multi-day CVRP automatically creates schedules across multiple days, with day-chaining — each day starts where the previous one ended. This eliminates geographic backtracking and lets regional teams work efficiently across territories without daily home returns.

Example: with 2 technicians, a max of 2 jobs each per day, and 28 jobs to schedule, day capacity is 4 jobs/day. The remaining 24 jobs overflow. The system extends to 7 days automatically — each tech gets daily assignments across the full period. No manual intervention. No spreadsheet planning.

Day-chaining for geographic continuity:

• Day 0: tech starts in Dallas, works Dallas jobs, ends in Fort Worth.
• Day 1: starts in Fort Worth, works toward Austin, ends in Austin.
• Day 2: starts in Austin, works San Antonio, heads back toward Dallas.

Day-chaining supports extended operations up to 14 consecutive days — enabling regional deployments where techs cross territories without daily home returns. For pricing the labor savings, the labor cost calculator is the quickest way to see the dollar impact.

Real-World CVRP Scenarios Where Capacity-Aware Routing Pays Off

Three patterns appear constantly across HVAC, plumbing, electrical, and tree service shops. In each, basic VRP produces a route that looks efficient but isn’t executable. CVRP catches the violation upstream and produces a schedule that survives the day.

Equipment bottleneck — tree service with shared bucket truck. 3 crews, 1 bucket truck, 8 jobs requiring it. Without CVRP, all 3 crews get assigned bucket-truck jobs simultaneously. Two crews arrive on-site unable to work. Failed jobs, wasted drive time, angry customers. With CVRP, jobs sequence across crews with handoff buffers: Crew A takes jobs 1–3, 30-minute handoff, Crew B takes 4–5, handoff, Crew C takes 6–8. Zero conflicts.

Skill bottleneck — HVAC with limited certifications. 5 techs, 40 jobs, 15 require EPA 608, only 2 techs certified. Without CVRP, non-certified techs get assigned refrigerant jobs because they’re closer. Compliance violations and failed appointments follow. With CVRP, the feasibility check eliminates every non-certified assignment instantly. 15 refrigerant jobs go to 2 certified techs. The remaining 25 spread across all 5 by location. If the 2 certified techs can’t finish all 15 in one day, the system creates multi-day assignments automatically. For pricing context on these calls, see the HVAC pricing guide.

Temporal violation — plumber with an impossible schedule. 1 plumber, 6 jobs, 14 hours of service time, 8-hour shift, no overtime. Without CVRP, the route looks perfect — minimal drive time, great clustering — but 14 hours doesn’t fit in 8. The last 3 jobs get missed. With CVRP, the feasibility check catches it immediately. Total time needed: 14 hours service + 2 hours travel + 1 hour lunch = 17 hours. The system redistributes excess jobs to other plumbers or creates multi-day. No impossible schedules reach the dispatch board. The plumbing pricing guide covers the upstream math for these calls.

How FieldCamp Implements CVRP

FieldCamp’s AI Dispatcher implements CVRP as a capacity-first pipeline — feasibility check first, optimization second. Shift hours, certifications, equipment availability, and territory limits are encoded as hard constraints. Preferred job counts, drive-time targets, and workload balance are encoded as soft penalties.

Equipment-based scheduling is wired into the same optimizer — the bucket truck has its own calendar, with handoff strategies and buffer times configurable per asset class. For longer operations, multi-day chaining extends automatically when daily capacity is exceeded, with day-end locations bridging into the next morning’s start. Dispatchers see every assignment alongside the capacity check that approved it through the AI dispatch scheduling view, and work order management stays in sync as capacity decisions are applied.

The result — schedules that don’t fall apart by 10 AM, because the impossible has been ruled out before the optimizer ever runs.

Continue Reading

• AI route optimization explained — how routing turns assignments into drive sequences.
• How AI dispatching algorithms work — the four-layer pipeline that wraps CVRP.
• How AI matches jobs to technicians — the assignment side of the equation.
• What is an AI dispatcher — the foundational definition.