Transit Trip Pattern Generation via Iterative Cost Model Revision
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current transit system trip planning services often fail to provide a diverse range of trip options due to their reliance on computationally expensive algorithms, either returning only a single result or overwhelming users with too many options, and may miss identifying certain trip patterns.
Innovation Solution
A computer-implemented method that iteratively determines optimal transit trips by revising a cost model based on previously found trips, penalizing arcs associated with those trips to find new, diverse patterns, and merging trip patterns using graph traversal algorithms to identify additional unique routes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If computationally expensive search algorithms are used to identify a large number of trips, then the diversity of trip results is improved, but the computational cost and time consumption increase significantly
Solution Approach 1:
The patent segments the trip identification process into multiple iterative phases. In each iteration, the algorithm identifies one optimal trip pattern, extracts it, and then modifies the cost model to prevent reselection of the same pattern. This segmentation transforms a single complex search into multiple simpler iterative searches, reducing overall computational burden while maintaining diversity.
Solution Approach 2:
The patent dynamically changes the cost model parameters between iterations. After each optimal trip is identified, the cost associated with arcs belonging to that trip is increased, effectively changing the parameter landscape for the next search. This parameter modification ensures diverse results without requiring completely new search algorithms for each trip.
2Device complexity
If a single optimal trip is returned based on a complex cost model, then the computational complexity is reduced, but the diversity and completeness of trip options are limited
Solution Approach 1:
The patent maintains continuous useful action by iteratively applying the same optimal trip identification algorithm multiple times. Rather than switching to different algorithms or stopping after one result, the system continuously generates new trips by modifying the cost model between iterations, ensuring both algorithmic simplicity and result diversity.
Solution Approach 2:
The patent implements feedback by using the identified optimal trip to modify the cost model for subsequent iterations. The cost model is updated based on the previously found trip pattern, creating a feedback loop that guides the algorithm toward discovering new, diverse trip patterns while maintaining computational efficiency.
3Adaptability or versatility
If arc costs are increased for arcs associated with previously found optimal trips, then the diversity of identified trip patterns is improved, but the cost model becomes more complex
Solution Approach 1:
The patent applies dynamics by making the cost model adaptive and changing over time. The cost model is not static but evolves during the iterative process, with arc costs dynamically adjusted based on previously identified trips. This dynamic modification enables diverse trip pattern discovery while keeping the underlying algorithm simple.
Data Source
AI summary
Systems and methods for generating a plurality of trip patterns are provided. One exemplary method includes receiving transit graph data describing a plurality of nodes respectively corresponding to a plurality of transit stations and a plurality of arcs respectively connecting the plurality of nodes. The method also includes performing a plurality of identification iterations. Each identification iteration includes determining an optimal transit trip connecting an origin node to a destination node based on a cost model. Each identification iteration also includes revising the cost model based on the determined optimal transit trip, such that the arc costs associated with one or more arcs associated with the optimal transit trip are increased. Each optimal transit trip can have an associated trip pattern describing a sequence of nodes traversed by such optimal transit trip. One exemplary system can include a transit planning platform that includes a trip pattern identification module.