Vehicle Request Clustering for Allocation Efficiency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current systems for managing vehicle requests, such as taxi services, often result in inefficient allocation of vehicles, leading to varying wait times for users and uneven distribution of requests among drivers, as they typically match vehicles to requests on a proximity-based, individual request basis without considering a broader overview of multiple requests.
Innovation Solution
A method and system that clusters vehicle requests based on geographical and temporal criteria before matching them with available vehicles, allowing for a more optimized allocation that balances wait times for users and travel distances for drivers by considering all requests within a cluster simultaneously.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If vehicles are matched to requests on an individual request-by-request basis using proximity, then the matching process is simple and fast, but the allocation efficiency is low and wait times vary significantly
Solution Approach 1:
The patent segments the vehicle request management process into two distinct phases: a clustering phase where requests are grouped based on spatial and temporal proximity, and a matching phase where vehicles are assigned to clusters. This segmentation allows the system to handle individual requests efficiently while optimizing overall allocation through cluster-level processing, thereby improving productivity without excessive complexity.
Solution Approach 2:
The patent applies preliminary action by performing clustering of vehicle requests before the actual vehicle matching process. By pre-grouping requests that are geographically and temporally close together, the system prepares optimized sets of requests that can be efficiently assigned to vehicles, reducing overall wait times and improving allocation efficiency before the matching step occurs.
2Loss of time
If vehicles are allocated based on closest proximity to individual requests, then the response time for individual requests is fast, but the overall balance of demands among drivers is uneven
Solution Approach 1:
The patent merges multiple individual vehicle requests into clusters based on their spatial and temporal characteristics. By combining nearby requests that occur close together in time, the system can assign a single vehicle to serve multiple requests in sequence, thereby reducing overall wait times while creating more balanced and efficient routes for drivers, improving both user experience and driver workload distribution.
Solution Approach 2:
The system performs preliminary clustering of requests before vehicle assignment, grouping requests that are geographically proximate and temporally close. This preliminary organization allows the matching algorithm to optimize vehicle routes to serve multiple clustered requests efficiently, reducing total wait time for users while creating more balanced and manageable demand patterns for drivers.
3Productivity
If the system considers all requests globally before matching, then the allocation optimization is maximized, but the computational complexity and processing time increase significantly
Solution Approach 1:
The patent divides the global request set into multiple smaller clusters based on spatial and temporal criteria. This segmentation transforms the computationally intensive problem of optimizing all requests globally into multiple smaller, more manageable sub-problems that can be solved independently and efficiently, maintaining high allocation optimization while reducing computational complexity and processing time.
Solution Approach 2:
The patent introduces spatial and temporal dimensions for organizing requests into clusters. By grouping requests based on their geographic proximity and temporal proximity, the system creates a multi-dimensional structure that naturally partitions the problem space, enabling efficient optimization within each dimension without requiring exhaustive global analysis, thus balancing optimization quality with computational feasibility.
Data Source
AI summary
A vehicle request management system having a server (10) arranged to communicate with a plurality of vehicle requesting devices (12) and a plurality of vehicles (14), each being equipped with a device (200) having route planning and navigation functionality. Upon receipt of a vehicle request, the server allocates the request to an existing cluster of vehicle requests if it is related to the requests. If it is not related to the requests, the request is used to generate a new cluster. The relationship between new vehicle requests and those of existing clusters is assessed based on a proximity of a pick-up location associated with the new request to a pick-up location of an existing request. The relationship is assessed using a parameter that is selected based on a geographical and temporal density of vehicle requests in a region associated with the cluster. When the cluster meets a threshold in relation to age of an oldest request associated with it, or alternatively a maximum size threshold, the requests in the cluster are subjected to a matching process to match a vehicle to each request for fulfilling the request.