Multi-Modal Transport Scheduling with Stochastic Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional scheduling methods for multi-modal transportation networks fail to efficiently integrate private and public transport routes, particularly when dealing with uncertain passenger information, leading to overcapacity, increased travel times, and environmental impact.
Innovation Solution
A system utilizing two-stage stochastic programming and decision diagrams to jointly schedule early riders and late riders across multiple transportation modes, optimizing routes and vehicle assignments to minimize travel time and environmental impact.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If conventional routing methods are used separately for each transportation mode, then the routing process is simple to implement, but the integration of multi-modal routes is difficult and travel time increases
Solution Approach 1:
The patent combines multiple transportation mode networks into a unified multi-modal transportation network, integrating private transport, public transport, and on-demand services into a single coordinated system that optimizes routes across all modes simultaneously rather than sequentially
Solution Approach 2:
The routing system is designed to handle multiple transportation modes universally through a single platform that can process and optimize routes for different transport types (private vehicles, public transit, on-demand services) using common algorithms and data structures
2Productivity
If all passenger information is assumed to be available ahead of time for optimization, then scheduling efficiency is improved, but the system fails to handle uncertain passenger information
Solution Approach 1:
The scheduling system dynamically adapts to changing passenger information by continuously updating schedules as new requests arrive or existing requests are modified, rather than relying on static pre-collected data, allowing the system to handle both known and uncertain passenger information effectively
Solution Approach 2:
The system incorporates feedback mechanisms where passenger requests and schedule changes are continuously monitored and fed back into the optimization algorithm, enabling real-time adjustments to schedules based on actual passenger information rather than assumptions
3Manufacturing precision
If separate optimization is performed for private and public transportation networks, then each network can be optimized independently, but the integration to provide true multi-modal routing is not achieved
Solution Approach 1:
The patent merges private and public transportation networks into a unified multi-modal system where routing optimization considers all transport modes simultaneously, enabling seamless transitions between different transportation types while maintaining optimization precision
Data Source
Figure 1
Figure 2A
Figure 2B
AI summary
Systems and methods for scheduling early riders (ERs) and late riders (LRs) to vehicles in a multi-modal transportation network (MTN). Stored instructions, when executed, cause a processor to perform acts of forecasting a finite set of scenarios, each scenario having a possible set of forecast LRs (FLRs) itinerary requests. Iteratively, generate ER groups and FLR groups for each scenario, based on a desired time of arrival at a destination. Assign a commuter vehicle (CV) for each ER and FLR group in each scenario. Iteratively, determine for each ER and FLR group a start time and an arrival time at the destination in the corresponding CV, for which, the ER and FLR group are assigned. The iterations continue until a joint schedule for the ERs and the FLRs form each scenario that minimizes an objective function. Formulate assignment information, and transmit the assignment information to the ERs and the assigned CVs.