Route Design System Using Inverse Reinforcement Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing route design systems for public transportation, such as bus routes, require significant manual effort and expertise to optimize routes and bus stop locations due to various constraints, making frequent updates inefficient and time-consuming.
Innovation Solution
A route design system that uses inverse reinforcement learning to calculate a cost function based on weighted features, selecting candidate relay points and designing routes efficiently, thereby automating the process of optimizing routes and bus stop locations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If manual review and coordination of routes and bus stop locations is performed, then route design quality can be maintained through expert know-how, but the process becomes time-consuming and difficult to update frequently
Solution Approach 1:
The patent replaces the manual mechanical process of expert review and coordination with an automated computer-based system. The system uses a cost function that quantitatively evaluates route designs based on multiple factors (passenger demand, operational costs, constraints), automatically selecting optimal routes and bus stop locations without requiring manual expert intervention for each evaluation.
Solution Approach 2:
The patent transforms the qualitative expert judgment process into a quantitative parameter-based evaluation system. By defining a cost function with specific parameters (passenger demand, operational costs, various constraints), the system enables automated calculation and comparison of different route designs, allowing frequent updates through systematic parameter optimization rather than manual review.
2Measurement precision
If individual evaluation of correlation between evaluation data values and analysis targets is performed, then specific factors can be analyzed, but optimizing the overall system requires significant know-how and time
Solution Approach 1:
The patent merges the individual evaluation of multiple correlation factors into a unified cost function. Instead of evaluating passenger demand, operational costs, and constraints separately and then manually integrating them, the system combines all these factors into a single comprehensive cost function that automatically calculates and compares overall route design quality, significantly improving system optimization productivity.
Solution Approach 2:
The cost function serves multiple functions simultaneously: it evaluates passenger demand, calculates operational costs, checks constraint satisfaction, and ranks different route design options. This multi-functional evaluation tool replaces the need for separate analysis processes and manual synthesis, enabling efficient overall system optimization.
3Adaptability or versatility
If frequent updates of routes and bus stop locations are attempted, then responsiveness to regional development and demographic changes improves, but manual coordination becomes impractical
Solution Approach 1:
The patent replaces manual coordination operations with automated computer-based calculations. The system can frequently re-evaluate and update route designs by simply inputting new demographic and regional data, automatically optimizing routes based on updated parameters without requiring manual coordination efforts.
Solution Approach 2:
The system performs self-service optimization by automatically adjusting route designs in response to input data changes. When regional development or demographic data is updated, the system autonomously re-evaluates and optimizes routes using the cost function, eliminating the need for manual intervention and enabling frequent adaptive updates.
Data Source
AI summary
The function input means 71 accepts input of a cost function that calculates a cost incurred in at least one of a selection of a candidate relay point and a design of a route, the cost function being represented as a linear sum of terms weighted by degree of importance attached to each of the features that an expert is assumed to intend when selecting the candidate relay point and design of the route. The learning means 72 learns the cost function by inverse reinforcement learning using training data that includes relay point information which is data that maps information indicating a relay point with surrounding information of the relay point and usage information of the relay point, and route result information which is result data of a route that pass through each relay point.