Fleet Design, Sizing and Vehicle Selection

Introduction

Fleet design is the structured process of deciding how many vehicles an organisation needs, what types and capacities they should have, where they should be based, and whether they should be owned, leased, or outsourced. In logistics, fleet decisions influence delivery reliability, transport cost, asset utilisation, carbon emissions, driver requirements, customer service, and operational resilience. They are therefore strategic decisions rather than simple purchasing exercises.

A fleet that is too small may create missed deliveries, excessive overtime, subcontracting costs, and poor service. A fleet that is too large may produce idle capacity, unnecessary capital investment, high fixed costs, and asset obsolescence. The central managerial task is to balance capacity and flexibility against cost and risk under uncertain demand.

Fleet sizing must consider demand profiles, vehicle payload, volumetric capacity, route duration, loading and unloading time, regulatory restrictions, driver availability, maintenance downtime, safety requirements, and service-level commitments. Vehicle selection must also reflect the characteristics of the freight and operating environment rather than relying only on purchase price or nominal payload.

Learning Objectives

By the end of this chapter, the reader should be able to:

• distinguish fleet design, fleet sizing, and vehicle selection;
• translate demand into required vehicle capacity and productive hours;
• calculate fleet requirements using payload, utilisation, availability, and service constraints;
• evaluate fleet mix and the trade-off between standardisation and flexibility;
• compare ownership, leasing, subcontracting, and hybrid acquisition strategies;
• apply total-cost-of-ownership and lifecycle reasoning;
• assess technological, environmental, regulatory, and ethical considerations in fleet decisions.

1. The Fleet Design Problem

Fleet design has four linked decisions:

1. Capacity decision: How much transport capacity is required?
2. Composition decision: What combination of vehicle types and capacities is appropriate?
3. Deployment decision: Where and when should vehicles be positioned and scheduled?
4. Governance decision: Which activities should be performed with owned, leased, or outsourced resources?

These decisions are interdependent. A larger vehicle may reduce the number of trips but be unsuitable for narrow urban streets. A highly standardised fleet can reduce training and maintenance complexity but may be inefficient for specialised loads. Outsourcing may provide peak flexibility but can reduce direct control over service quality and safety.

Fleet planning normally occurs at three horizons:

A robust design separates the base fleet, required for predictable demand, from flex capacity, obtained through overtime, temporary leases, spot-market carriers, or contracted vehicles.

2. Understanding Demand and Service Requirements

Fleet sizing begins with a demand profile, not an average volume alone. Relevant dimensions include:

• number of consignments, stops, pallets, cages, or parcels;
• weight and volume by product category;
• geographic distribution and route density;
• time windows and delivery frequency;
• weekday, seasonal, and promotional variation;
• collection as well as delivery demand;
• returns, failed deliveries, and reverse logistics;
• demand uncertainty and required service probability.

The planning period must be explicit. A vehicle requirement calculated for a month cannot automatically be used as a daily requirement because vehicles operate through shifts, route cycles, maintenance intervals, and legal driving limits.

Capacity dimensions

Vehicle capacity is multidimensional. Important measures include:

• Payload capacity: maximum legal mass of freight, usually in kilograms or tonnes.
• Volumetric capacity: available cubic space, usually in cubic metres.
• Pallet or roll-cage capacity: number of standard handling units.
• Floor or loading length: useful for long or awkward freight.
• Operational capacity: capacity remaining after safety, segregation, temperature, hazardous-material, and loading constraints.

A vehicle may reach its weight limit before its volume limit, or vice versa. For low-density products, volume is often the binding constraint; for dense products, payload is more likely to bind.

3. Core Fleet-Sizing Models

3.1 Capacity-based sizing

For a homogeneous demand stream, the minimum number of vehicles based on payload is:

where:

• D_w = demand weight during the planning period;
• Q_w = legal payload per vehicle;
• u_w = achievable payload utilisation, expressed as a proportion;
• a = fleet availability proportion;
• ⌈x⌉ = round upward to the next whole vehicle.

All weight terms must use the same unit. Availability accounts for vehicles unavailable because of maintenance, inspection, accident, cleaning, or other planned downtime. This formula is appropriate when demand is divisible and vehicle trips are comparable. It is only a first approximation because it ignores route duration, stop time, geography, time windows, and the fact that availability is not necessarily a capacity multiplier in every planning context.

For volume:

where D_v is demand volume, Q_v is vehicle volume capacity, and u_v is volumetric utilisation.

When both weight and volume constrain operations, a conservative preliminary estimate is:

This is not always sufficient for mixed loads. A vehicle can have unused weight capacity and unused volume simultaneously because the remaining orders do not fit together. Mixed-load problems require route-level packing or an optimisation model.

3.2 Time-based sizing

A fleet can also be sized from productive vehicle hours. Required vehicles are estimated by:

where:

• H_d = total required vehicle-hours in the planning period;
• H_s = scheduled hours available per vehicle;
• u_t = productive-time utilisation;
• a = availability.

Productive hours exclude loading, waiting, refuelling or charging, administrative activity, and other nonproductive time unless those activities are already included in the required-hours estimate. The definition must be consistent: double-counting nonproductive time causes over-sizing.

For a route, total time may be represented as:

where T_r is route duration. Service time may be estimated as:

where n_s is the number of stops and t_s is average service time per stop. In practice, stop time varies with delivery type, building access, congestion, paperwork, and unloading equipment.

3.3 Service-level and peak sizing

Average demand is inadequate when customers require reliable service during peaks. If daily demand has mean μ and standard deviation σ, a simple capacity buffer can be expressed as:

where z is a planning factor associated with the desired protection against demand variation. This relationship assumes that the selected demand measure is approximately stable and that the use of a normal approximation is reasonable. It should not be applied mechanically to highly skewed, intermittent, or strongly seasonal demand.

The resulting fleet requirement is:

where C_effective is usable capacity per vehicle after utilisation and operating restrictions. A service-level buffer may instead be represented through scenario analysis, simulation, or a prescribed percentile of demand. The choice should reflect the cost of under-capacity versus the cost of idle assets.

Worked Example 1: Payload and availability

A distributor must deliver 1,080 tonnes during a monthly planning period. Each truck has a legal payload of 12 tonnes, expected average payload utilisation of 75%, and fleet availability of 90%. The payload-based requirement is:

The effective monthly capacity per truck is 8.1 tonnes under the stated simplified assumptions. Therefore:

This result is intentionally a capacity-period calculation. It does not mean that 134 trucks are physically required at one moment; the result depends on how many trips each truck makes during the month. If each vehicle can complete 10 equivalent trips and the demand is 1,080 tonnes per month, the denominator must include trip frequency. A planner must therefore avoid applying a monthly demand figure to a single-trip capacity formula.

Worked Example 2: Daily fleet requirement using trips

Daily demand is 420 pallets. A vehicle carries 18 pallets on average and can complete two planned trips per day. Operational utilisation is 85%, and availability is 92%. Effective daily capacity per vehicle is:

Thus:

If a contractual service requirement demands capacity for a 15% peak, design demand becomes 483 pallets. The requirement is then:

Management must decide whether three additional vehicles should be owned, leased seasonally, or supplied by a carrier.

4. Fleet Mix and Vehicle Selection

Fleet mix concerns the number of vehicles in each class rather than the total number alone. Common classes include motorcycles, vans, rigid trucks, tractor-trailer combinations, refrigerated vehicles, tankers, specialist vehicles, and electric or alternative-fuel vehicles.

Selection criteria should be translated into operational requirements:

Standardisation versus specialisation

Standardisation reduces driver training, spare-parts variety, maintenance complexity, and scheduling difficulty. It may also improve resale value and procurement leverage. However, a standard vehicle may be poorly matched to some routes, causing low utilisation or unsafe loading.

Specialisation is justified when cargo, regulation, geography, or service requirements are materially different. Examples include refrigerated vehicles for temperature-sensitive products, smaller vehicles for urban access, and high-cube vehicles for lightweight bulky freight. The correct objective is not maximum standardisation but economically justified commonality.

Weight-volume compatibility

A useful diagnostic is freight density:

where ρ_f is freight density, W_f is freight weight, and V_f is freight volume. If freight density is below the vehicle's weight-to-volume ratio, volume is likely to fill first; if it is above that ratio, payload is likely to fill first. This is a screening measure, not a substitute for actual load planning.

5. Route Structure and Fleet Sizing

Fleet requirements depend on route topology. A dense urban operation may serve many stops per kilometre but experience long service and congestion times. A rural operation may have few stops but high travel time. Identical daily volume can therefore require different fleet sizes.

Important route constraints include:

• maximum legal driving and working hours;
• customer time windows;
• depot opening hours;
• loading-bay appointment systems;
• vehicle access restrictions;
• driver breaks and shift rules;
• charging or refuelling time;
• return-to-depot requirements;
• delivery sequence and compartment restrictions.

A practical workflow is to estimate demand, convert demand into orders or handling units, construct representative routes, calculate route duration and load feasibility, and then test the design under peak and disruption scenarios. Vehicle routing software can improve accuracy, but poor demand data or unrealistic service-time assumptions will still produce poor decisions.

6. Ownership, Leasing, Outsourcing and Acquisition

Ownership

Owned vehicles provide control over specifications, branding, schedules, data, and driver practices. They are often suitable for stable, predictable, strategically important operations. Disadvantages include capital commitment, residual-value risk, maintenance responsibility, and exposure to technological obsolescence.

Finance and operating leases

A finance lease generally transfers much of the economic use and risk of the asset to the lessee, while an operating lease traditionally provides use for a defined period with different residual-value and return arrangements. Accounting treatment depends on the applicable reporting standard and contract terms; managers should not assume that a lease is economically “off balance sheet.”

Leasing can preserve capital, provide access to newer vehicles, and improve flexibility. It may, however, cost more over a long stable life and impose mileage, condition, or termination constraints.

Outsourcing and subcontracting

Outsourcing converts some fixed costs into variable or semi-variable costs and provides access to capacity, geographic coverage, and specialist expertise. Risks include loss of control, inconsistent service, carrier financial failure, safety noncompliance, data limitations, and reputational exposure. Outsourcing does not eliminate the buying organisation's responsibility for due diligence and governance.

Hybrid models

A common design is a core owned or long-term leased fleet for predictable base demand, supplemented by contracted carriers or short-term rentals for peaks, abnormal loads, or remote regions. Hybrid models require clear rules for allocation, tendering, performance measurement, escalation, and emergency capacity.

7. Total Cost of Ownership and Economic Evaluation

Vehicle selection should use lifecycle cost rather than acquisition price alone. A simplified annual total-cost model is:

where each C represents an annual cost category and R is the annualised value of residual value or other recoveries. The exact treatment of depreciation, financing, taxes, and residual value must follow the organisation's accounting and investment methodology. Avoid adding depreciation and capital repayment indiscriminately if this double-counts the same economic cost.

For comparing alternatives over several years, discounted cash flow may be used:

where I₀ is initial investment, CF_t is net cash flow in year t, r is the discount rate, n is the evaluation horizon, and RV_n is residual value at the end. Cash flows must be stated consistently as nominal or real and before or after tax.

Cost per useful unit is often more informative than cost per vehicle:

where U may be deliveries, tonnes, pallet movements, kilometres, or productive vehicle-hours. The chosen denominator must reflect the service being purchased; cost per kilometre can be misleading when vehicles have very different payloads or stop productivity.

8. Risk, Resilience and Sustainability

Fleet design should test disruption scenarios such as demand surges, fuel shortages, extreme weather, labour absence, vehicle recalls, charging failure, border delays, and carrier insolvency. Resilience may come from spare capacity, vehicle interchangeability, multiple suppliers, geographic dispersion, repair agreements, and emergency contracts. However, resilience has a cost, and excess redundancy should be justified by risk exposure.

Sustainability assessment should include direct and indirect impacts. Relevant measures include fuel or electricity consumption, greenhouse-gas emissions, local air pollutants, noise, tyre and brake particulates, battery production, end-of-life treatment, and infrastructure requirements. A simplified operational emissions estimate is:

where E is emissions, A is activity such as vehicle-kilometres, f is energy or fuel consumed per activity unit, and e is the emissions factor per energy or fuel unit. Boundaries and emissions factors must be stated because results differ between tailpipe and lifecycle accounting.

Electric vehicles may reduce local emissions and operating energy cost but introduce range, payload, charging, battery degradation, grid-capacity, and acquisition challenges. The appropriate decision depends on route length, dwell time, duty cycle, climate, charging access, and electricity source—not on technology labels alone.

9. Technology and Analytical Support

Telematics can provide actual mileage, idling, harsh braking, route duration, fuel or energy consumption, and utilisation data. Transport-management systems support order consolidation, routing, dispatch, proof of delivery, and carrier management. Digital twins and simulation can test alternative fleet mixes before investment.

Artificial intelligence can assist demand forecasting, travel-time prediction, vehicle replacement analysis, and dynamic capacity allocation. Nevertheless, AI outputs depend on data quality and model assumptions. Governance should address explainability, privacy, cybersecurity, driver monitoring, bias, human oversight, and the consequences of automated decisions. Technology should improve managerial judgement, not replace validation of legal, safety, and operational constraints.

10. Illustrative Case: MetroFresh Distribution

Background

MetroFresh is an illustrative regional grocery distributor serving supermarkets and convenience stores. It currently operates 22 diesel rigid trucks. Demand is growing, urban deliveries are becoming more frequent, and several cities are introducing low-emission zones. Orders vary from dense chilled products to bulky low-density packaging.

Problem

The existing fleet experiences 88% availability, average payload utilisation of 68%, and frequent late deliveries during promotional periods. Management must choose between purchasing additional diesel vehicles, acquiring electric vehicles, leasing capacity, or outsourcing peak demand.

Analysis

The team separates demand into three route families: dense urban, suburban mixed, and long-distance regional. Urban routes are constrained primarily by stop time and access, while regional routes are constrained by driving hours and payload. A single average vehicle is therefore rejected as the planning unit.

For the next peak period, forecast daily demand is 520 pallet-equivalents. The proposed urban vehicle carries 14 pallets and completes 2.1 trips per day. With 82% productive utilisation and 90% availability, effective capacity is:

The preliminary requirement is:

Route simulation shows that four routes exceed legal working time unless service appointments are changed. Consequently, the capacity calculation is increased to 26 vehicles for the peak design. The base requirement is 22 vehicles; the additional four vehicles are treated as flex capacity rather than automatically purchased.

Approach and solution

MetroFresh adopts a hybrid design:

• retain 18 owned vehicles for stable base routes;
• lease four smaller vehicles for urban access and seasonal flexibility;
• contract two vetted carriers for peak and disruption coverage;
• pilot four electric vehicles on predictable urban routes with depot charging;
• standardise body dimensions and telematics across owned and leased vehicles;
• include safety, on-time delivery, temperature compliance, emissions, and data reporting in carrier contracts.

The decision is evaluated using cost per completed delivery, not vehicle acquisition cost alone. A six-month pilot is required before wider electric adoption, with control routes, energy records, maintenance data, range incidents, and customer-service outcomes.

Outcome

The design reduces the number of late urban deliveries during the peak scenario while avoiding permanent ownership of all peak capacity. It also generates evidence on whether electric vehicles can meet the actual duty cycle. The case illustrates that fleet sizing is an integrated capacity, route, contract, and risk decision.

Learning points

1. Average demand concealed route-specific constraints.
2. Vehicle count alone did not reveal the required fleet mix.
3. Flex capacity was economically preferable to sizing the owned fleet for the absolute peak.
4. Vehicle technology was evaluated through duty-cycle evidence rather than assumptions.
5. Outsourcing required measurable standards and governance controls.

11. Implementation Framework and Best Practices

A professional fleet-design process should:

1. define the service promise and planning horizon;
2. collect segmented demand, load, route, and time data;
3. identify legal, safety, environmental, and access constraints;
4. estimate base, peak, and disruption capacity;
5. test payload, volume, time, and route feasibility;
6. construct alternative fleet mixes;
7. compare lifecycle cost, service, risk, and sustainability;
8. assess ownership, leasing, and outsourcing options;
9. pilot uncertain technologies or operating models;
10. establish review triggers for demand, regulation, cost, and performance changes.

Useful performance indicators include availability, utilisation, empty kilometres, payload fill, on-time delivery, cost per delivery, fuel or energy per kilometre, accident frequency, maintenance compliance, emissions, and subcontractor service quality. Indicators should be interpreted together: high utilisation with poor service may signal overloading of the fleet, while low empty kilometres with excessive waiting may still indicate weak productivity.

Challenges and Limitations

Fleet models simplify reality. Demand may be nonstationary, service times may be correlated with congestion, and vehicle availability may decline during severe peaks when it is most needed. Nominal capacity may be unusable because of axle limits, load sequence, product segregation, or customer handling conditions. Financial comparisons are sensitive to fuel prices, interest rates, residual values, tax treatment, and technology learning curves.

Models should therefore be accompanied by sensitivity analysis. Test changes in demand, utilisation, availability, route time, energy price, maintenance cost, and service penalties. A recommendation that remains attractive across plausible scenarios is more credible than one based on a single precise forecast.

Chapter Summary

Fleet design integrates capacity planning, vehicle selection, deployment, and sourcing governance. Fleet size should be based on usable capacity and productive time rather than nominal vehicle specifications. Payload, volume, utilisation, availability, route duration, service windows, and demand variability must be examined together. Fleet mix decisions balance standardisation against specialisation and should reflect freight density, geography, handling, regulation, and sustainability requirements.

Ownership, leasing, outsourcing, and hybrid models distribute cost, control, flexibility, and risk differently. Total cost of ownership and discounted cash-flow analysis provide stronger economic comparisons than purchase price alone. Resilience requires deliberate treatment of peak demand and disruption scenarios. Telematics, routing systems, simulation, and AI can improve analysis, but data quality, human oversight, privacy, safety, and regulatory compliance remain essential. The best fleet is not necessarily the largest or newest; it is the fleet that reliably delivers the required service at acceptable lifecycle cost and risk.

Key Terms

Fleet design: The strategic configuration of vehicle capacity, type, location, ownership, and operating rules.

Fleet sizing: Determining the number of vehicles required for a specified service and planning horizon.

Fleet mix: The combination of vehicle classes and capacities in a fleet.

Payload: Maximum permitted mass of freight carried by a vehicle.

Volumetric capacity: Usable cubic space available for freight.

Utilisation: Proportion of available capacity or time used productively.

Availability: Proportion of the fleet operationally ready for use.

Productive vehicle-hours: Vehicle time contributing directly to required transport service.

Base fleet: Capacity required for predictable, recurring demand.

Flex capacity: Additional capacity obtained for peaks, disruptions, or unusual demand.

Fleet density: Freight weight divided by freight volume.

Total cost of ownership: Lifecycle cost of acquiring, operating, maintaining, and disposing of an asset.

Residual value: Expected value recovered when an asset is sold or retired.

Outsourcing: Obtaining transport service from an external provider.

Operating lease: Contractual use of a vehicle for a defined period under specified terms.

Service level: Required performance standard, such as on-time delivery probability.

Duty cycle: Actual pattern of distance, load, stops, energy use, and operating time.

Telematics: Connected collection and transmission of vehicle and driving data.

Resilience: Capacity to maintain or recover service during disruptions.

Fleet standardisation: Use of common vehicle specifications to reduce operational complexity.

Review Questions

Short Answer Questions

1. Why is average demand insufficient for fleet sizing?
2. Distinguish payload capacity from volumetric capacity.
3. What is the purpose of separating base fleet from flex capacity?
4. Identify two advantages and two risks of outsourcing transport.
5. Why should vehicle selection use total cost of ownership?

Long Answer Questions

1. Explain how payload, volume, utilisation, availability, and route time interact in fleet-sizing decisions.
2. Compare owned vehicles, leasing, outsourcing, and hybrid fleet models.
3. Discuss standardisation and specialisation as competing fleet-design strategies.
4. Explain how service levels and demand variability affect fleet capacity decisions.
5. Evaluate the operational and governance implications of using electric vehicles and AI-enabled fleet systems.

Analytical / Application Questions

1. A fleet must deliver 600 pallets per day. Each vehicle carries 20 pallets, completes 1.8 trips per day, has 80% productive utilisation, and fleet availability is 95%. Calculate the preliminary number of vehicles required and state two limitations of the calculation.
2. Design a fleet-mix methodology for a distributor serving dense urban, suburban, and rural customers with both chilled and ambient products.
3. Compare purchasing four vehicles permanently with leasing or outsourcing them for a three-month seasonal peak. Identify the data required for a defensible recommendation.

Multiple Choice Questions with Answers

Question 1

Which factor most directly accounts for vehicles unavailable because of maintenance or inspection?

A. Payload utilisation

B. Fleet availability

C. Freight density

D. Stop time

Correct Answer: B. Fleet availability

Question 2

A vehicle carries 18 pallets and completes two trips daily. With 85% utilisation and 92% availability, what is its effective daily capacity?

A. 28.15 pallets

B. 33.12 pallets

C. 30.60 pallets

D. 25.20 pallets

Correct Answer: A. 28.15 pallets

Question 3

When low-density freight fills the vehicle's cubic space before its legal mass limit, which constraint is binding?

A. Payload constraint

B. Time-window constraint

C. Volumetric constraint

D. Availability constraint

Correct Answer: C. Volumetric constraint

Question 4

What is a principal advantage of a hybrid fleet model?

A. It eliminates all fixed costs

B. It combines base control with peak flexibility

C. It guarantees lower cost in every situation

D. It removes the need for carrier governance

Correct Answer: B. It combines base control with peak flexibility

Question 5

Which measure is generally more meaningful than purchase price when comparing vehicle alternatives over their useful lives?

A. Vehicle colour

B. Total cost of ownership

C. Nominal engine size alone

D. Number of available cup holders

Correct Answer: B. Total cost of ownership

Question 6

In the equation ρ_f = W_f ÷ V_f, what does ρ_f represent?

A. Fleet availability

B. Freight density

C. Route duration

D. Residual value

Correct Answer: B. Freight density

Question 7

Why may a route with modest freight volume still require a large fleet?

A. Volume always determines fleet size

B. Route time, stops, access, and time windows consume capacity

C. Vehicle payload is irrelevant

D. Availability is always 100%

Correct Answer: B. Route time, stops, access, and time windows consume capacity

Question 8

Which is an appropriate governance concern when using AI for fleet decisions?

A. Whether the model uses any data at all

B. Explainability, privacy, cybersecurity, and human oversight

C. Whether drivers can ignore all safety rules

D. Whether regulation becomes unnecessary

Correct Answer: B. Explainability, privacy, cybersecurity, and human oversight

Question 9

What does fleet standardisation typically reduce?

A. Training and spare-parts complexity

B. All route constraints

C. All demand variability

D. The need for maintenance

Correct Answer: A. Training and spare-parts complexity

Question 10

If design demand is 483 pallets and effective vehicle capacity is 28.152 pallets per day, how many vehicles are required after rounding upward?

A. 15

B. 16

C. 17

D. 18

Correct Answer: D. 18