Skip to main content

Home Visits Routing

Vehicle Routing with Time Windows

A home healthcare agency dispatches nurses to visit patients at home. Each patient has a required time window and visit duration. The goal is to minimize total travel while ensuring every patient is seen within their window — the Vehicle Routing Problem with Time Windows (VRPTW), an NP-hard extension of the classic CVRP.

Where This Decision Fits

Healthcare planning chain — the highlighted step is what this page optimizes

Strategic Facility location
Staff Workforce scheduling
Patient Flow Home visits routing
Supply Chain Medical supply delivery

The Problem

Why route optimization matters for home healthcare

Over 5 million Americans receive home health services annually. A home healthcare agency employs teams of nurses who travel between patients' homes, providing services ranging from wound care and IV infusions to vital sign checks and medication management. Each patient requires a visit within a specific time window (e.g., 8:00–10:00 AM for morning insulin management), and each visit has a known service duration tied to clinical acuity.

Route optimization can reduce driving time by 20–30%, increasing direct patient care by 15–25% (Fikar & Hirsch, 2017). With each nurse carrying a limited shift capacity (measured in care-minutes), the agency must decide which patients each nurse visits and in what order — balancing travel efficiency against time window feasibility.

This is the Vehicle Routing Problem with Time Windows (VRPTW) — an NP-hard combinatorial optimization problem. Unlike the basic TSP, VRPTW must simultaneously determine the number of vehicles (nurses), assign customers (patients) to vehicles, and sequence visits so that every time window constraint [ei, li] is respected.

Healthcare Domain VRPTW Model
Patient Customer node
Nurse Vehicle
Appointment window Time window [ei, li]
Visit duration Service time si
Care-minutes per shift Vehicle capacity Q
Agency office Depot
VRPTW Formulation minimize   Σk Σi Σj cij · xijk   // total travel distance
subject to
  Σk Σj xijk = 1    // each patient visited exactly once
  Σi di · Σj xijk ≤ Q    // nurse capacity not exceeded
  ei ≤ tik ≤ li    // arrival within time window
  tik + si + cij ≤ tjk + M(1 - xijk)    // time consistency
  xijk ∈ {0, 1}   // binary routing decision

Where cij is the travel cost between locations i and j, xijk indicates whether nurse k travels from i to j, tik is the arrival time of nurse k at patient i, and Q is the shift capacity in care-minutes.

See full Routing theory and all algorithms

Try It Yourself

Route nurses to patient homes with time windows

Home Visits Route Optimizer

15 Patients · 200 min capacity
Metropolitan morning round: 15 patients across the metro area with standard 200 care-minute nurse capacity.
# Patient Coords Visit (min) Window Acuity

Ready. Click “Solve & Compare All Algorithms” to run.

Algorithm Distance (km) Nurses Care % Time (ms)
Click Solve & Compare All Algorithms
Route details will appear here after solving.

The Algorithms

Construction and improvement heuristics for VRPTW

Key Insight
VRPTW combines two hard problems: partitioning patients into feasible routes (bin packing with time windows) and sequencing visits within each route (TSP with constraints). No polynomial-time algorithm guarantees optimality, so practitioners rely on construction heuristics to build initial solutions and improvement metaheuristics to refine them.
Construction Heuristic

Solomon I1 Insertion

O(n² · m) per route  |  Sequential route building

Solomon’s I1 insertion heuristic (1987) builds routes one at a time by repeatedly inserting the unrouted customer that causes the least combined increase in distance and time penalty. For each unrouted customer, every feasible insertion position in the current route is evaluated. The customer with the best insertion cost is selected. When no more customers can be feasibly inserted, a new route begins. This method produces compact, time-window-respecting routes and is the standard benchmark constructor for VRPTW instances.

Construction Heuristic

Nearest Neighbor with Time Windows

O(n²)  |  Greedy sequential

Starting from the depot, greedily select the nearest unvisited patient whose time window is still feasible (i.e., the nurse can arrive before the window closes). If the nurse arrives early, she waits until the window opens. When no feasible patient remains or the nurse’s capacity is exhausted, return to depot and start a new route. Simple and fast, but myopic — it ignores future consequences and may strand hard-to-reach patients for later routes.

Improvement Metaheuristic

Simulated Annealing

O(iterations × n)  |  Swap & relocate moves

Starting from a construction heuristic solution, Simulated Annealing explores the neighborhood by randomly applying swap (exchange two patients between routes) and relocate (move one patient to a different route position) moves. Improving moves are always accepted; worsening moves are accepted with probability e−Δ/T, where T is the temperature that gradually cools. This allows escaping local optima and finding better solutions that greedy methods miss, at the cost of additional computation time.

Real-World Complexity

Why home healthcare routing goes beyond the VRPTW model

Time Windows

Patients require visits within strict clinical windows — insulin must be given before meals, wound dressings changed on schedule. Late arrivals compromise care quality and safety.

Skill Matching

Not every nurse can handle every case. IV infusions, psychiatric assessments, and palliative care each require specialized certifications, constraining which nurse can visit which patient.

Continuity of Care

Patients benefit from seeing the same nurse across visits. Continuity improves outcomes and satisfaction but limits routing flexibility and can lead to workload imbalances.

Travel Uncertainty

Traffic conditions, weather, and road closures make travel times stochastic. Robust routes must build in buffers, and real-time rerouting is often necessary during the day.

Workload Equity

Fair distribution of caseload and travel burden across nurses prevents burnout, improves retention, and ensures no single nurse is consistently overloaded or underutilized.

Key References

Foundational works in home healthcare routing

  • Fikar, C. & Hirsch, P. (2017). “Home health care routing and scheduling: A review.” Computers & Operations Research, 77, 86–95. DOI: 10.1016/j.cor.2016.07.019
  • Begur, S.V., Miller, D.M. & Weaver, J.R. (1997). “An integrated spatial DSS for scheduling and routing home-health-care nurses.” Interfaces, 27(4), 35–48. DOI: 10.1287/inte.27.4.35
  • Benzarti, E., Sahin, E. & Dallery, Y. (2013). “Operations management applied to home care services.” Decision Support Systems, 55(2), 587–598. DOI: 10.1016/j.dss.2012.10.015

Need to optimize your home healthcare operations?

From nurse routing to patient scheduling and resource allocation, mathematical modeling can transform your healthcare delivery. Let’s discuss how Operations Research can work for you.

Disclaimer
Data shown is illustrative. Patient names, locations, acuity levels, and time windows are representative scenarios for educational purposes and do not reflect any specific home healthcare operation or real patient information.
Back