Location-Based Services (LBS) in On‑Site Fulfillment: Geofencing, Order Allocation, Routing and Autonomous Scheduling

1 What is LBS

Location‑Based Services (LBS) use various positioning technologies to obtain the current location of a device and provide information and basic services via mobile internet. Users first determine their spatial position, then retrieve location‑related resources and information. LBS integrates mobile communications, internet, spatial positioning, location data, and big data, using a mobile‑internet platform for data updates and interaction.

2 Terminology

Engineer : On‑site fulfillment courier

Geofence : A closed polygon formed by points (latitude‑longitude vertices)

3 Business Overview

The "Zhuanzhuan" on‑site fulfillment service relies on a C2B model for 3C digital product recycling, offering fast and precise on‑site service. The workflow includes checking city support, address eligibility, valuation, order placement, automatic order assignment, and on‑site collection.

4 Geofence‑Based Exposure Ordering and Order Allocation

4.1 Exposure Ordering

Because on‑site collection cannot cover every city, accurately judging whether a user's address is within a supported area is crucial. The process converts the address to latitude‑longitude coordinates and checks if the point lies inside a geofence.

4.1.1 Preliminary Filter: Minimum Bounding Rectangle

Any irregular polygon can be enclosed by a rectangle defined by its top‑right and bottom‑left coordinates, allowing fast exclusion of most geofences.

4.1.2 Precise Filter: Ray‑Casting Method

The ray‑casting algorithm shoots a ray from the point in a direction and counts intersections with the polygon; an odd count means the point is inside, even means outside.

Common point‑in‑polygon methods include ray‑casting, corner‑angle, cross‑product, area, binary search, etc. The ray‑casting method was chosen for its simplicity and applicability to any polygon.

4.1.3 Simple Retrieval Process

Two stages: (1) Service loads geofence data from the database, builds an in‑memory index; (2) User query checks the point against the index using the chosen algorithm.

4.1.4 Retrieval Index

To avoid costly linear scans, an R‑tree index is built on the minimum bounding rectangles, reducing lookup complexity to O(log N).

4.2 Order Allocation

Unlike food delivery, the order density is lower, so a low‑cost, fast allocation method is needed. Engineers are attached to geofences representing their service area; when an order arrives, its coordinates are matched to a geofence, and the corresponding engineer is selected based on business rules.

5 Location‑Based Route Planning and Autonomous Order Scheduling

5.1 Route Planning

To improve fulfillment efficiency, the system calculates distances between orders and engineers. Simple straight‑line (great‑circle) distance is insufficient for complex terrain, so third‑party navigation services are used. Because synchronous calls would increase latency, the solution asynchronously invokes the navigation service, caches results, and combines them with business rules.

5.2 Autonomous Order Scheduling

When automated rules cannot handle complex scenarios, on‑site engineers can manually reassign orders. Two typical cases are described where engineers exchange orders based on proximity.

To quickly find nearby orders or engineers, Elasticsearch GEO queries are employed. Engineers periodically report their coordinates, orders store their coordinates, and GEO distance queries retrieve entities within a specified radius.

{
  "filter": [{
    "geo_distance": {
      "location": {
        "lat": 20.12345,
        "lon": 100.223344
      },
      "distance": "3km",
      "distance_type": "arc"
    }
  }]
}

6 Summary

The article outlines several LBS‑driven scenarios for on‑site fulfillment, including geofence checking, order allocation, routing, and autonomous scheduling, and notes that as data volume grows, continuous optimization of retrieval and service mechanisms is required.

7 References

https://www.cnblogs.com/lbser/p/4471742.html

https://my.oschina.net/1024bits/blog/782820

https://www.cnblogs.com/yym2013/p/3673616.html

https://blog.csdn.net/WilliamSun0122/article/details/77994526

https://toutiao.io/posts/4as8i9/preview