Haul Truck Task Assignment Using Reinforcement Learning and Energy State
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current mining operations face challenges in efficiently managing haul truck dispatch and task assignment, particularly in large-scale mining sites, where manual or restricted automatic dispatch methods are inefficient and do not account for the unique constraints of greenhouse-gas-free machines, which require frequent recharging and have limited energy storage, leading to operational inefficiencies.
Innovation Solution
A computer-implemented method using a reinforcement-learning model to optimize task assignments for a fleet of haul trucks by inputting state data into a trained model that learns an assignment policy to optimize a reward function, considering haul truck performance variance due to changing haul weights and material weights, thereby automating task assignments and optimizing operations such as material movement and energy consumption.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If manual dispatch is used to direct haul trucks to tasks, then operators can manage tasks with flexibility, but the dispatch efficiency and productivity are insufficient compared to automated systems
Solution Approach 1:
The system enables haul trucks to autonomously request tasks from the assignment engine based on their current state (location, battery charge, task completion status). The trucks self-manage their task assignments without continuous human intervention, improving dispatch efficiency while maintaining operational flexibility through automated decision-making
Solution Approach 2:
The patent replaces manual human dispatch operations with an automated machine-learning-based assignment engine. The system uses reinforcement learning models to automatically optimize task assignments, substituting human operators' mechanical dispatch decisions with automated computational algorithms that process truck states and generate optimized assignments in real-time
2Extent of automation
If restricted automatic dispatch is used with rules and limitations on haul truck assignments, then automation is improved, but the system requires continuous controller intervention to optimize production
Solution Approach 1:
The assignment engine autonomously manages task assignments without requiring continuous controller intervention. The system self-adjusts assignments based on real-time truck states, battery levels, and task priorities, enabling fully automated operation while eliminating the need for controllers to continuously manage restrictions and optimize production
Solution Approach 2:
The system continuously receives feedback from haul trucks about their state (location, battery charge, task completion) and automatically adjusts task assignments in real-time. This closed-loop feedback mechanism enables the system to self-optimize production without human intervention, maintaining high automation while reducing operational complexity
3Object-affected harmful factors
If greenhouse-gas-free machines with limited energy storage are used, then environmental sustainability is improved, but the machines require frequent recharging which limits operational cycles
Solution Approach 1:
The assignment engine proactively assigns tasks to haul trucks based on their predicted battery discharge rates and remaining charge capacity. By calculating future energy requirements before tasks are executed, the system ensures trucks are assigned tasks they can complete without running out of charge, enabling sustained operational cycles while maintaining environmental sustainability
Solution Approach 2:
The system dynamically adjusts task assignment parameters based on battery state of charge and predicted discharge rates. By changing assignment criteria to account for energy constraints, the system optimizes task allocation to maximize operational cycles between recharging events while maintaining zero-emission operation
4Device complexity
If haul truck assignments do not account for performance variance due to changing haul weights, then assignment simplicity is maintained, but assignment accuracy and optimization are reduced
Solution Approach 1:
The assignment engine dynamically adjusts task assignments based on real-time haul weight data and predicted performance variance. By continuously updating assignments according to changing truck loads and their impact on discharge rates, the system maintains high assignment accuracy while adapting to varying operational conditions without requiring overly complex static models
Data Source
AI summary
Systems and methods are disclosed for managing task assignments for a fleet of haul trucks at a mine site. An assignment engine may: receive state data for a haul truck including haul weight data and second state data for the mine site that is indicative of a plurality of available tasks and associated task material weight data; assign a task to the haul truck by inputting the state data into a trained reinforcement-learning model, wherein: the model has been trained to learn an assignment policy that optimizes a reward function, such that the learned policy accounts for vehicle performance variance due to changing vehicle haul weight and/or road conditions; and cause the at least one haul truck to be operated according to the at least one task assignment.