Fleet Fuel Optimization Without the Black Box: A Technical Look at How OptiMile Pro Actually Works

No mystery algorithms. No painful integrations. No vendor lock-in. Here's the architecture, the math, and the security model — so you can evaluate it properly.

You've seen the pitch deck. Someone from ops or the fleet owner forwarded you a link about fuel savings — 27-34% — and asked you to "take a look at this."

So here you are. And your first three questions are probably:

1. How does it actually work?
2. What does it need to integrate with?
3. What's the security and data exposure?

Fair. Let's answer all three with enough technical detail to actually be useful.

The Problem Statement (From an Engineering Perspective)

Trucking fleets negotiate contracted fuel pricing with major networks — Love's, Pilot Flying J, TA/Petro, and others. These contracts provide per-station pricing that varies significantly across locations, even within the same network.

The optimization problem: given a fixed route, a set of contracted station prices along that route, fuel tank constraints, and a fuel network topology, determine the optimal sequence of fuel stops and fill quantities that minimizes total fuel cost.

This is a constrained optimization problem with the following characteristics:

• Decision variables: which stations to stop at and how many gallons to purchase at each
• Constraints: tank capacity, minimum fuel thresholds, station availability within a defined corridor, driver MPG, fuel on board at final destination, and contracted network membership
• Objective function: minimize total fuel expenditure across the route

It's not a trivial problem. For even a 5-stop route with dozens of candidate stations per stop, the solution space is combinatorially large. Spreadsheets and manual heuristics leave significant money on the table — which is exactly what the route data shows.

How OptiMile Pro Solves It

Input: Contracted Pricing Ingestion

The system starts with your fleet's actual contracted pricing. Not retail. Not national averages. Your rates, at your stations, under your agreements.

Pricing data is ingested via direct feed from fuel networks (where available) or fleet-provided rate tables (CSV/Excel upload of your negotiated rates by station or network). The pricing model accounts for station-level price variation within the same network, time-based price fluctuations, volume-based pricing tiers where applicable, and multi-network optimization.

Processing: Route Corridor Analysis

Given a route (origin, destination, waypoints), the system defines a corridor — typically 25 miles on either side of the planned route — and identifies every contracted fuel station within that corridor. Each candidate station is geocoded and mapped to the route with distance from route centerline, position along the route, detour cost, and station attributes.

Optimization: Dynamic Programming Engine

The core solver uses a dynamic programming approach — not a greedy heuristic, not a generic algorithm, not a neural network making opaque decisions. It provides optimal substructure, a globally optimal solution (not an approximation), deterministic and explainable results, and polynomial time computation for real-time plan generation.

The solver evaluates the cost of every feasible fueling sequence, accounting for purchase price per gallon, fuel consumed between stations, detour fuel cost, and tank constraints. Output: an ordered list of fuel stops with specific gallon quantities at each, total route fuel cost, and the delta versus unoptimized baseline fueling.

Real Results From Real Routes

These are real routes with real contracted pricing, not synthetic benchmarks:

• 4,796 mi route — Savings: $737 (26.9%)
• 4,796 mi route — Savings: $754 (27.5%)
• 6,893 mi route — Savings: $1,433 (33.6%)
• 1,989 mi route — Savings: $211 (29.3%)

Integration Architecture: Start With One, Scale When Ready

OptiMile Pro is designed for zero-integration onboarding. You get more out of the system if you integrate with your telematics provider but you don't have to.

Day One (No Integration): Upload routes manually (CSV, Excel, or through the web UI), upload or provide contracted pricing, and receive optimized fuel plans immediately. The full optimization engine runs on manually uploaded data.

When You're Ready (Optional Integration): REST API (fully documented, versioned), TMS integration, and telematics hookup for real-time re-optimization mid-route. The integration path is incremental. Start manual, automate what makes sense, expand at your own pace.

Security and Data Handling

Data Architecture: Each fleet's data is logically isolated with no cross-customer data sharing. AES-256 encryption for all stored data, including pricing, routes, and generated plans. TLS 1.2+ for all API and web communications. Configurable retention policies — you own your data and can export or delete at any time.

Compliance: No third-party data sharing — your data is used exclusively to generate your fuel plans. Role-based access with audit logging. Infrastructure hosted on GCP with multi-AZ deployment, automated backups, and infrastructure as code.

What We Don't Need Access To: TMS (unless you want integration), driver PII, load or customer data, financial systems, or ELD/HOS data. Minimum data required: routes (origin/destination/waypoints) and contracted fuel pricing.

What a Proof of Concept Looks Like

A POC should answer one question: does this work with your data, on your routes, at your scale?

Week 1: You provide 5-10 representative routes and your contracted pricing. We run them through the optimization engine and deliver detailed results. Week 2: Your team reviews the results. We walk through the optimization logic for any route you want to examine. Week 3: If the numbers hold, we discuss integration path. If they don't, we part ways. No cost for the POC.

The entire evaluation requires zero changes to your existing infrastructure. No staging environment. No sandbox integration. No IT tickets. Just data in, results out.

OptiMile Pro is fuel stop optimization for trucking fleets with 10+ tractors. Built for fleets that want results, not marketing.