Graph Neural Network Vehicle Assignment Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for assigning vehicles to transport tasks, such as e-hailing services, are inefficient and fail to optimize vehicle assignment due to unknown true costs, leading to suboptimal customer satisfaction and operational inefficiencies.
Innovation Solution
A method using graph neural networks to determine similarity weights between vehicle and transport task graphs, enabling a learning-based approach to optimize vehicle assignments by processing node features and applying differentiable assignment algorithms, trained through reinforcement learning.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If traditional assignment methods are used for vehicle allocation, then the system is simple to implement, but the assignment efficiency and customer satisfaction deteriorate due to suboptimal matching
Solution Approach 1:
The patent replaces traditional mechanical assignment algorithms with a graph neural network-based learning system. The GNN processes vehicle and transport task graphs to automatically learn optimal assignment strategies from data, substituting rule-based mechanical decision-making with an intelligent system that adapts to complex real-world patterns and dynamics.
Solution Approach 2:
The patent transforms the assignment problem from a static optimization task into a dynamic learning process. By using learnable parameters in the graph neural network that adapt through training on historical data, the system continuously improves assignment quality based on changing conditions, rather than relying on fixed assignment rules.
2Measurement precision
If learning-based assignment methods are implemented, then the assignment quality improves through capturing real-time influences, but the computational complexity and processing time increase
Solution Approach 1:
The patent performs preliminary training of the graph neural network offline using historical assignment data. This pre-learning phase captures patterns and relationships in advance, allowing the trained model to make rapid, high-quality assignments in real-time without undergoing complex computation during actual assignment operations.
Solution Approach 2:
The patent uses graph representations as simplified copies of complex real-world systems. By creating graph models of vehicles and transport tasks that capture essential features and relationships, the system can process assignments efficiently using these simplified representations rather than dealing with full system complexity.
3Adaptability or versatility
If graph neural networks are used to process vehicle and transport task data, then the ability to capture locality and real-time influences improves, but the computational resources and energy consumption increase
Solution Approach 1:
The patent segments the overall assignment problem into independent graph processing units. Each vehicle and transport task is represented as a separate graph node with its own features, allowing the graph neural network to process information in a modular fashion that captures local relationships while maintaining computational efficiency through structured graph operations.
Data Source
AI summary
Aspects concern a method for controlling vehicles to perform transport tasks comprising supplying information about vehicles and information about transport tasks to a graph neural network by associating each vehicle with a vehicle graph node and each transport task with a transport task graph node, processing the vehicle and the transport graph by the neural network, wherein the neural network determines a feature for each graph node, determining, for each pair of a transport graph node and vehicle graph node, a weight representing a similarity between the features determined for the transport graph node and the vehicle graph node, selecting an assignment between the transport graph nodes and the vehicle graph nodes from a set of possible assignments, wherein the selected assignment maximizes the sum of the weights of the pairs and controlling each vehicle according to the selected assignment.