AGV Routing and Timing via Time-Space Network Dijkstra
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing routing methods for automated guided vehicles in semiconductor manufacturing plants face inefficiencies, leading to high interference between vehicles and a low likelihood of achieving Just-In-Time (JIT) performance, as they struggle to generate schedules that balance vehicle routes and timing effectively.
Innovation Solution
A method that determines vehicle routing and timing using a graph-based approach, where a first graph represents possible routes and a second graph represents vehicle states and transitions, utilizing the Dijkstra algorithm to find a minimum cost path that considers earliness, tardiness, and staying costs, with parameters to weight JIT performance and transport efficiency, and iteratively adjusts for interference to avoid collisions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a schedule is generated to prevent interference between a large number of vehicles, then vehicle interference is reduced, but schedule generation efficiency decreases and JIT transport is less likely to be realized
Solution Approach 1:
The patent segments the scheduling problem into two distinct phases: (1) route determination for each vehicle using Dijkstra's algorithm to find minimum cost paths, and (2) schedule generation to coordinate vehicles and prevent interference. This segmentation allows each phase to be optimized independently, improving overall efficiency while maintaining reliability.
Solution Approach 2:
The patent introduces dynamic rerouting capabilities that allow schedules to be adjusted in real-time when deviations occur. The system can dynamically recalculate routes and timings to maintain JIT performance, making the schedule adaptive rather than static, thus improving both efficiency and reliability.
2Ease of operation
If a predefined schedule is used for vehicle transport, then schedule management is simplified, but JIT performance is compromised due to inability to adapt to deviations
Solution Approach 1:
The patent implements feedback mechanisms that continuously monitor vehicle positions and schedule adherence. When deviations are detected, the system provides feedback to the routing algorithm, which then adjusts routes and timings to restore JIT performance. This closed-loop control maintains simplicity while ensuring adaptability.
Solution Approach 2:
The patent performs preliminary route determination using Dijkstra's algorithm to establish optimal paths before vehicles depart. This preliminary action creates a solid foundation for schedule management while leaving room for dynamic adjustments during execution, balancing simplicity with adaptability.
3Quantity of substance
If the number of vehicles is increased to handle large-scale plant operations, then transport capacity is improved, but interference between vehicles increases and schedule efficiency decreases
Solution Approach 1:
The patent segments the scheduling problem into individual route determination tasks for each vehicle, allowing parallel processing of multiple vehicle routes. This segmentation enables the system to efficiently manage large numbers of vehicles without proportionally increasing computational complexity, thus maintaining schedule efficiency while scaling transport capacity.
Data Source
AI summary
A time-space network including a set of state nodes and a set of state transition edges each of which connects between the state nodes, is generated. Each of the state nodes represents a state in which a vehicle is present at a certain point node at a certain time. State transition costs increasing according to a difference of time corresponding to before or after transition and arrival time are defined. In the time-space network, by using the Dijkstra algorithm, a minimum cost transition path in which a sum of the state transition costs is minimum among paths from a starting point state node to any of the state nodes indicating that the vehicle is present at an arrival point is obtained. The starting point state node represents a state that the vehicle is present at a departure point at a reference time. The route and the timing of the vehicle are determined based on the minimum cost transition path.