Real-Time Fleet Routing System Using Segmented Filtering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current systems for real-time order dispatching in same-day delivery services face challenges in efficiently routing orders to geographically distributed drivers due to the complexity of real-time traffic conditions and the inability to scale with increasing volumes of orders, leading to slow computation times and inefficiencies.
Innovation Solution
A routing system employing a three-layer architecture of filtering, routing, and ranking, utilizing machine learning algorithms to reduce the set of potential drivers, account for real-time traffic, and calculate optimal or pseudo-optimal routes, thereby speeding up the order assignment process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If traditional routing systems are used to assign orders to drivers, then routing decisions can be made, but computation time becomes prohibitively slow as order volume increases
Solution Approach 1:
The routing system is divided into multiple independent modules including a filtering module that reduces candidate drivers, a routing module that calculates optimal paths, and a ranking module that prioritizes drivers. This segmentation allows each module to process data independently and in parallel, dramatically reducing overall computation time while maintaining assignment accuracy
Solution Approach 2:
The filtering module performs preliminary action by pre-processing driver data and reducing the candidate set before the routing module executes complex path calculations. By eliminating unsuitable drivers early based on basic criteria (availability, location, capacity), the system avoids performing expensive routing computations on all drivers, thus reducing computation time significantly
2Measurement precision
If comprehensive routing calculations are performed for all drivers, then optimal routes can be determined, but the system cannot scale with increasing order volumes
Solution Approach 1:
The filtering module performs preliminary action by pre-processing driver data and reducing the candidate set before the routing module executes complex path calculations. By eliminating unsuitable drivers early based on basic criteria (availability, location, capacity), the system avoids performing expensive routing computations on all drivers, thus reducing computation time significantly
Solution Approach 2:
The routing system is divided into multiple independent modules including a filtering module that reduces candidate drivers, a routing module that calculates optimal paths, and a ranking module that prioritizes drivers. This segmentation allows each module to process data independently and in parallel, dramatically reducing overall computation time while maintaining assignment accuracy
3Reliability
If real-time traffic conditions are incorporated into routing decisions, then more accurate routes can be provided, but computation complexity increases
Solution Approach 1:
The system introduces an intermediary data structure (the reduced candidate driver set) between the raw traffic data and the final routing calculations. By first filtering drivers based on basic criteria and then applying complex routing algorithms only to this reduced set, the system incorporates real-time traffic conditions accurately while managing computational complexity through the intermediate filtering step
Data Source
AI summary
Embodiments of systems and methods for automated real-time routing within a fleet of geographically distributed drivers are disclosed. Embodiments may operate to dispatch orders and determine routing in real-time in a geographic area through application of rule-based filtering of drivers and selective application of optimal or non-optimal routing solutions utilizing the real-time locations of drivers, real-time conditions within the geographic area and the locations for the set of orders being routed by the system.