CONFIDENTIAL - TRADE SECRET
This document contains proprietary algorithms and methodologies that constitute trade secrets of HaulR Inc. Any unauthorized use, disclosure, or reproduction of this information is strictly prohibited and may result in legal action. Access is restricted to authorized personnel only.
Document Classification: INTERNAL ONLY | Created: January 26, 2025
HaulR Proprietary Algorithms
Technical documentation of our competitive advantages
HaulR's fundamental innovation is eliminating empty drives through intelligent batching of 3-5 deliveries along optimized routes. This increases driver utilization from ~60% to 90%+, enabling 2-3x earnings compared to traditional single-stop platforms.
Patent Strategy
The specific methodology described below is the subject of a provisional patent application filed January 2025. Patent covers: route corridor calculation, earnings-per-mile optimization, driver home address integration, and real-time job batching logic.
The Problem We Solve
Traditional delivery platforms (Uber, DoorDash, Roadie) operate on a single-stop model:
- • Driver accepts Job A at Restaurant/Warehouse X
- • Drives to pickup location (often empty drive)
- • Delivers to Customer at Location Y
- • Job ends - driver now must find new job from Location Y
- • Typically drives empty 30-40% of time looking for next job
Result: 60% utilization, $15-20/hour earnings after expenses
When a driver sets a planned route (Home → Work, or Start → End destination), we calculate a geographic "corridor" using configurable width (default 10 miles).
Method:
- • Use Haversine formula to calculate great-circle distances
- • For each job: find perpendicular distance from route line
- • Filter jobs within corridor width threshold
- • Calculate additional detour miles required
Proprietary optimization: We pre-compute corridor segments for common routes in each market, reducing real-time calculation overhead by 80%.
Each job receives a score based on profitability relative to the driver's planned route.
Scoring Formula:
Where detour_miles = additional distance beyond the driver's planned route
Ranking Logic:
- • Higher earnings per mile = higher priority
- • Jobs on-route (0 detour) ranked highest
- • Filter out jobs below $2/mile threshold
- • Secondary sort by pickup time proximity
Advanced feature: Machine learning model adjusts scoring based on time-of-day traffic patterns, historical driver preferences, and seasonal demand fluctuations.
AI continuously analyzes available jobs to create optimal multi-stop sequences for each driver.
Batching Algorithm:
- Identify all jobs within driver's corridor
- Calculate pickup time windows and delivery deadlines
- Use modified traveling salesman algorithm to find optimal sequence
- Ensure pickups happen before deliveries (constraint satisfaction)
- Limit batch size to 3-5 stops (driver capacity + time constraints)
- Re-optimize every 60 seconds as new jobs appear
Proprietary optimization: We use a custom heuristic that reduces computation time from O(n!) to O(n²), enabling real-time optimization even in high-density markets.
Revolutionary feature: Drivers can specify a "home base" or starting point, and we filter jobs based on radius from that location OR their planned route.
Implementation:
- • Driver sets home address (geocoded to lat/lon)
- • Two filtering modes: "Radius" or "Route"
- • Radius mode: filter jobs within X miles of home (configurable)
- • Route mode: filter jobs along planned route corridor
- • Prevents drivers from seeing irrelevant jobs far from operational area
Competitive advantage: No other platform lets drivers pre-define operational zones based on their personal geography. This dramatically improves job relevance and driver satisfaction.
Our algorithms improve automatically as we process more jobs, creating a compounding data advantage:
Historical Route Data
We learn which routes are most popular, optimal times for each corridor, and typical traffic patterns. This data is unique to our platform and cannot be replicated by competitors.
Driver Preferences
Machine learning tracks which jobs drivers accept/reject, preferred times, vehicle capacity utilization, and customer ratings to improve matching accuracy over time.
Pricing Intelligence
Historical bid data, acceptance rates, and market pricing inform dynamic pricing recommendations, optimizing for both customer affordability and driver earnings.
Q2 2025: Predictive Job Availability
AI predicts when/where jobs will appear based on historical patterns, allowing drivers to position themselves proactively for maximum earnings.
Q3 2025: Multi-Driver Collaborative Routing
Coordinate multiple drivers in the same area to hand off packages mid-route, creating even more efficient multi-stop sequences across the entire driver network.
Q4 2025: Customer Delivery Time Prediction
Machine learning provides customers with accurate delivery time estimates based on real-time driver location, traffic, and historical completion times for similar routes.
Document Access & Security
Authorized Personnel: Founder/CEO, CTO (when hired), Lead Engineers (with NDA)
Prohibited Actions: Screenshots, copying, forwarding, discussing with third parties
Storage: This document must remain on secure HaulR servers only
Version Control: All edits tracked with timestamps and author identification
If you have received this document in error, immediately delete all copies and notify the sender.