Project Overview
REINDEERS will officially open its global platform in December 2025. After the initial launch and stabilization phase, development of the new DVRP + WMS integrated logistics platform will begin, with a planned two-month development window from January to February 2026.
The project unifies WMS (Warehouse Management System) and DVRP (Dynamic Vehicle Routing & Planning) into a single AI-centric environment. It enables 3PL companies to manage inbound, storage, dispatching, and delivery through real-time AI decision-making. All features operate under a multi-tenant SaaS architecture where each logistics company is isolated as an independent tenant.
Development Scope
This project is part of the REINDEERS SaaS-based 3PL logistics platform, centered on AI-driven DVRP and WMS capabilities. The system will be developed as four integrated applications under a unified environment.
- ① System Administrator Portal — A global management web console for the entire SaaS platform. It enables the creation and control of logistics companies, pricing plans, system monitoring, and data analytics.
- ② SaaS Company Web System — A web application for logistics and warehouse companies using the platform. It covers employee, truck, warehouse, customer, dispatching, inbound/outbound, and settlement operations.
- ③ Client Company Web Portal — A dedicated web portal for 3PL client companies contracted with logistics providers. It allows inbound and outbound requests, inventory visibility, shelf-life alerts, and billing management.
- ④ Mobile Native Application (for Staff) — A mobile app for logistics company employees such as drivers and warehouse operators. It supports attendance tracking, daily work scheduling, inbound/outbound operations, GPS-based delivery, and digital signatures.
All components are interconnected through a unified API layer, maintaining data consistency and operational reliability within a single SaaS infrastructure.
What is DVRP?
DVRP (Dynamic Vehicle Routing & Planning) is a system that determines the optimal delivery routes and truck assignments in real time. Unlike traditional static route planners, DVRP continuously reacts to live operational conditions—such as GPS-based truck positions, traffic congestion, driver working hours, and delivery time constraints. The AI automatically recalculates routes and reallocates vehicles to minimize total cost, travel time, and idle distance.
AI-Centric Design Philosophy
The core concept of this platform is that AI becomes the proactive decision-maker across all operational processes. It is designed to automate logistics decisions from warehouse slotting to vehicle dispatching while learning from real-world data. AI integration spans five major domains:
- Vehicle dispatching & real-time route optimization
- Warehouse slotting & reallocation recommendations
- Picking route optimization
- Human resource and equipment workload balancing
- Continuous learning through feedback loops
AI in DVRP
The DVRP module is engineered to manage 1,000–2,000 trucks simultaneously through real-time data streaming. The system leverages AI to handle dynamic dispatching, considering both operational constraints and live environmental factors.
Key AI Capabilities
- AI Auto-Dispatch — Determines the best available truck within 2–4 seconds of DO creation, balancing cost, distance, and capacity.
- GPS-Driven Tracking — Uses live GPS data from each truck to monitor current position, estimated arrival, and idle status.
- Traffic-Aware Optimization — Integrates real-time traffic data and congestion patterns to re-route deliveries dynamically.
- Real-Time Re-Dispatch — Instantly reallocates shipments in case of breakdowns, delays, or road closures.
- Multi-Objective Planning — Optimizes for cost, time, or hybrid objectives based on company policy.
- Feedback-Based Learning — Incorporates manager evaluations (“appropriate / inappropriate”) into continuous AI retraining.
The dispatching logic is based on a hybrid AI model that combines statistical pattern learning from historical trips with heuristic algorithms that account for live constraints. This enables a continuously self-improving dispatch network that adapts to new delivery conditions automatically.
AI in WMS
The WMS (Warehouse Management System) uses AI to optimize storage efficiency and operational flow. Every inbound, outbound, and picking event feeds learning models that detect inefficiencies and recommend actions.
AI Slotting Optimization
- Respects hard constraints: temperature zones, hazardous goods, load limits, and dedicated areas
- Applies priority learning based on demand frequency, product size, expiration date, and client profile
- Suggests top 3–5 storage locations at inbound time; operators can accept or override
- Manual overrides are recorded and used to train future recommendation models
AI Reallocation Recommendation
Each night, the AI reviews full warehouse data to identify potential relocations that improve space utilization or picking efficiency. Items nearing expiration, high-turnover goods, or under-utilized racks are flagged for reassignment.
AI Picking Route Optimization
During outbound processing, the AI computes optimal picking paths by combining FEFO (First Expired, First Out) with minimal travel distance. Tasks are distributed among zone staff using collaborative algorithms to prevent overlapping movement paths.
AI Lifecycle
- Data Collection — Aggregates data from GPS, DOs, inventory, sensors, and warehouse activity
- Feedback Capture — Human confirmations and corrections are converted into training features
- Retraining Cycle — AI models are recalibrated nightly or weekly for improved prediction accuracy
- Performance Validation — Compares AI-suggested actions against real outcomes
- Reporting — Quantifies savings in time, cost, and distance achieved by AI optimization
Through this iterative process, the platform evolves into a self-learning logistics ecosystem that becomes increasingly autonomous over time.
Unified AI Architecture
Although DVRP and WMS operate with independent AI engines, they share a unified data and learning layer. Dispatching, slotting, and picking models interact through a centralized AI Decision Network that harmonizes logistics intelligence across fleet and warehouse domains.
- Inventory Intelligence — Tracks turnover, capacity, and spatial efficiency
- Fleet Intelligence — Learns vehicle health, route patterns, and utilization
- Operational Intelligence — Analyzes workforce efficiency and workload balance
- AI Decision Hub — Integrates all insights for unified decision-making
Human-in-the-Loop Philosophy
AI acts as the proactive recommender, while humans remain the verifier and decision calibrator. The system reduces repetitive decision load while allowing managers to focus on strategic supervision. Ultimately, fleets and warehouses operate together as a single intelligent network.