Route Search Acceleration via Node Clustering and Path Approximation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing routing algorithms, such as Dijkstra's, are computationally intensive for large graphs like regional or continental maps, requiring significant resources and time to compute optimal vehicle routes, especially when direct access to underlying graph attributes is limited.
Innovation Solution
A method that approximates all-pairs distance matrices by clustering nodes into subgroups, calculating exact intra-group distances, and approximating inter-group distances, allowing for parallel queries to reduce computational effort and data requirements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If Dijkstra's algorithm is used to compute shortest paths for all pairs of customers on large regional or continental maps, then exact shortest path solutions are obtained, but the computational time and resource consumption become excessively high
Solution Approach 1:
The patent segments the complete set of customers into multiple subsets, each served by a dedicated routing algorithm instance. This segmentation divides the large computational problem into smaller, manageable sub-problems, reducing the computational burden on each algorithm instance while maintaining overall solution accuracy through coordinated routing across subsets.
Solution Approach 2:
The patent applies partial action by computing exact shortest paths only for critical routes within each subset using Dijkstra's algorithm, while using approximation methods for less critical routes. This selective computation approach reduces overall computational time while maintaining high accuracy for the most important routing decisions.
2Productivity
If additional data such as speed profiles, highway hierarchies, and traffic volume is used to accelerate the search algorithm, then route search speed increases, but the data buffering and processing requirements increase
Solution Approach 1:
The patent extracts only the essential routing information needed for computation, separating critical data elements from the broader dataset. By taking out only the necessary attributes (such as basic geographic coordinates and connectivity information) and processing these selectively, the system achieves efficient routing without the heavy data processing burden of comprehensive datasets.
Solution Approach 2:
The patent applies local quality by loading and processing only the specific data segments relevant to each customer subset and its geographic region. Rather than buffering and processing entire national or continental datasets, the system loads only the local data needed for each subset's routing computations, significantly reducing memory requirements and data processing overhead.
3Measurement precision
If raw map data is made available to the routing service, then accurate distance computation is enabled, but commercial and technological restrictions often prevent such data access
Solution Approach 1:
The patent introduces an intermediary layer between the routing service and the raw map data. This intermediary computes and stores pre-processed distance matrices and routing information, allowing the routing service to operate with summarized data rather than direct access to raw map data. The intermediary handles the data access restrictions while maintaining computation accuracy through pre-computed metrics.
Solution Approach 2:
The patent creates a simplified copy of the essential routing information in the form of pre-computed distance matrices and graph representations. Instead of requiring direct access to raw map data, the system uses these copied, pre-processed data structures that capture the necessary routing semantics, enabling accurate distance computation without the data access restrictions that would apply to raw map data.
Data Source
AI summary
A method for finding an approximation to the all-pairs shortest travel path between a number of predetermined nodes, comprising clustering nodes of an original road network into a plurality of subgroups so that the number of nodes in each subgroup is no greater than a first predetermined threshold and the maximum distance between any two nodes of a subgroup is no greater than a second predetermined threshold; adding information of intra group shortest paths for all pairs in a same subgroup to a newly created higher level road network; adding information of inter group shortest paths for the plurality of subgroups to the same road network; and searching the same road network for the shortest travel path. In those cases in which the path returned is not exact, the path represents one in the original map, even if not necessarily the best one.