Motor carriers lose millions annually to empty miles. In regional operations, the gap between loaded and empty mileage often exceeds 22 percent. Integer Linear Programming (ILP) provides a mathematically rigorous method to reposition tractors across a multi-terminal network while respecting driver domicile rules, hours-of-service limits, maintenance windows, and customer time requirements.
Strong Tower Consulting has helped carriers integrate ILP-based repositioning engines with Truckmate, McLeod, and TMW environments. The solver consumes daily extracts of open loads (EDI 204), current tractor and driver positions, and terminal balance targets, then returns an optimal set of deadhead moves that can be imported back as suggested dispatch actions.
Problem Formulation
Let T be the set of terminals (domicile and satellite yards), K the set of tractors, and D the set of days in the planning horizon (typically 3-7 days). Decision variables are defined as follows:
- xijt = 1 if tractor i travels empty from terminal j to terminal k departing on day t, 0 otherwise.
- yld = 1 if load l is covered by a tractor departing on day d.
Objective Function
Minimize total empty miles:
minimize Σi,j,k,t (distance(j,k) * xijt)
Core Constraints
- Loaded move coverage: Every accepted load must be assigned exactly one tractor within its pickup window.
- Flow conservation: For each tractor and terminal, the number of arrivals must equal departures plus any net change required to meet domicile balance targets.
- Driver domicile balance: Σ arrivals - Σ departures = 0 for each driver domicile over the horizon. This prevents long-term accumulation or depletion of drivers at any terminal.
- Time-window constraints: Departure and arrival times must respect both customer appointment windows and Hours of Service (HOS) rules derived from ELD data.
- Equipment availability: No more tractors may depart a terminal than are physically present after accounting for inbound loaded and empty arrivals.
- Maintenance and inspection windows: Tractors due for service cannot be assigned to loads or repositioning moves during required downtime.
Branch-and-Bound Implementation Notes
Modern solvers such as Gurobi, CPLEX, or open-source HiGHS can solve realistic regional networks (15-40 terminals, 300-800 tractors) to optimality within 30-90 seconds on standard server hardware. Key performance techniques include:
- Pre-processing to fix variables for loads that have only one feasible terminal of origin.
- Valid inequalities that tighten driver domicile balance equations.
- Warm-starting from the previous day's solution shifted forward by 24 hours.
- Custom branching priorities that first fix high-cost deadhead lanes.
| Scenario | Deadhead % | Loaded Miles | Empty Miles | Annual Savings (at $1.85/mi) |
|---|---|---|---|---|
| Current manual dispatch | 23.4% | 14,872,000 | 4,542,000 | - |
| ILP optimized (3-day horizon) | 14.1% | 14,872,000 | 2,445,000 | $3,880,000 |
| ILP optimized (5-day horizon) | 12.8% | 14,872,000 | 2,184,000 | $4,370,000 |
Integration With Existing TMS Platforms
The ILP engine does not replace your TMS. Instead, it augments it. A nightly SQL extraction pulls open loads, current tractor positions, driver settlement domiciles, and fuel surcharge lanes from Truckmate, McLeod, or TMW. The solver returns a ranked list of recommended empty repositioning moves that appear as suggested dispatches inside the native system. EDI 214 status updates and driver settlement files remain unchanged.
Because Strong Tower Consulting, LLC and Exspeedite LLC are separately organized sister companies under common ownership, carriers receive independent mathematical guidance without pressure to purchase proprietary software. Our only product is better decisions.
Practical Considerations for Motor Carriers
Successful ILP repositioning projects require clean terminal master data, accurate driver domicile mapping in the settlement system, and disciplined use of EDI 204 tender acceptance windows. When these foundations exist, the model reliably reduces empty miles by 8-12 percentage points while improving on-time performance and driver satisfaction through more predictable home-time cycles.
Frequently Asked Questions
What is the objective function in an ILP tractor repositioning model?
The objective is to minimize total empty (deadhead) miles across the network while satisfying all loaded move requirements, driver domicile balance constraints, and equipment availability limits.
How are time-window constraints modeled in fleet repositioning ILP?
Time windows are enforced through flow conservation constraints that link arrival and departure times between terminals, ensuring tractors and drivers arrive within allowable windows for subsequent loads or repositioning moves.
What commercial TMS platforms benefit from an external ILP repositioning solver?
Truckmate, McLeod, TMW, and similar systems can export daily load and equipment position data via SQL or EDI 204/214 transactions. The ILP solver returns optimized repositioning recommendations that dispatchers can action within the native TMS.
Can an ILP model simultaneously balance driver domiciles and minimize empty miles?
Yes. Driver domicile balance equations are added as equality constraints that force the net flow of drivers into and out of each domicile terminal to equal zero over the planning horizon, while the primary objective still minimizes total deadhead.
Ready to quantify the financial impact an ILP-based repositioning engine could have on your specific network? Strong Tower Consulting delivers independent, vendor-neutral systems architecture and technology advisory services to motor carriers, freight brokers, and private fleets.