Substrate Dispatching With Deep Reinforcement Learning Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current dispatching systems in substrate fabrication facilities rely on manual setting of dispatching parameters and ranking orders, leading to uncertainty and varying performance due to changing manufacturing conditions.
Innovation Solution
The implementation of deep reinforcement learning to automatically generate and modify dispatching parameters and ranking orders based on real-time data from the fabrication facility, using a software agent trained in a simulation environment.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If manual setting of dispatching parameters and ranking orders is used, then system complexity is reduced, but adaptability to changing manufacturing conditions deteriorates
Solution Approach 1:
The dispatching system automatically adjusts dispatching parameters and ranking orders based on real-time manufacturing conditions without manual intervention. The system serves itself by using machine learning models to autonomously optimize dispatching decisions, eliminating the need for manual parameter setting while adapting to changing conditions.
Solution Approach 2:
The system dynamically changes dispatching parameters and ranking orders based on real-time manufacturing conditions. Machine learning models continuously update parameter values and ranking sequences according to current facility state, enabling adaptability while the underlying system structure remains relatively simple.
2Productivity
If manual setting of dispatching parameters is used, then ease of operation is improved, but productivity deteriorates
Solution Approach 1:
The dispatching system automatically optimizes parameters and rankings without requiring operator intervention. This self-service capability maintains ease of operation while significantly improving productivity through data-driven, real-time optimization of substrate processing throughput.
Solution Approach 2:
Manual mechanical parameter setting is replaced with automated machine learning-based optimization. The system substitutes human operators with intelligent algorithms that continuously adjust dispatching parameters to maximize throughput, improving productivity while maintaining operational simplicity.
3Adaptability or versatility
If automated deep reinforcement learning is implemented, then adaptability to changing conditions is improved, but device complexity worsens
Solution Approach 1:
Machine learning models serve as intermediaries between manufacturing conditions and dispatching decisions. These models absorb the complexity of adaptive optimization, allowing the rest of the dispatching system to remain relatively simple while gaining advanced adaptability through the intermediary layer.
Solution Approach 2:
The system uses simulation environments to create virtual copies of the fabrication facility for training machine learning models. This copying approach allows complex adaptive behavior to be developed and tested in silico before deployment, reducing the complexity burden on the actual production system.
Data Source
AI summary
A method for substrate dispatching management at a substrate fabrication facility is provided. The method includes obtaining data about a state of a fabrication facility and providing the data as input to an agent of a predictive subsystem associated with the fabrication facility to obtain one or more outputs indicative of one or more settings of one or more dispatching factors. The one or more dispatching factors comprise a dispatching parameter or ranking order. A dispatching decision is generated using the one or more settings of the one or more dispatching factor and a set of operations on a candidate set of substrates, based on the dispatching decision, is initiated.