Substrate Processing Scheduling for Utility-Constrained Operations

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing substrate processing systems face challenges in efficiently managing utility usage, particularly during periods of limited resource availability or chemical shortages, which can degrade processing quality and require significant time and effort to adjust settings.

Innovation Solution

A substrate processing system utilizing reinforcement learning to create schedules that optimize utility usage by positioning operations along a time sequence, incorporating trained models to generate multiple schedules that minimize utility consumption while maintaining processing efficiency.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Loss of energy

If manual adjustment of setting values is performed to suppress utility usage, then utility consumption is reduced, but time and effort are significantly increased

Engineering Contradiction:

Improveutility usageVSAvoidadjustment time

Core Design Contradiction:

Loss of energyVSLoss of time

Solution Approach 1:

The schedule creation apparatus automatically creates optimized schedules that minimize utility usage without requiring manual intervention. The system uses trained models to autonomously determine optimal operation timings and settings, eliminating the need for operators to manually adjust parameters while achieving utility suppression.

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

The patent replaces manual mechanical adjustment of setting values with an automated information processing system. The schedule creation apparatus uses computer-based algorithms and trained models to calculate optimal schedules, substituting human operators with an automated computational system that rapidly generates utility-optimized schedules.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Loss of energy

If setting values are changed to suppress utility usage, then utility consumption is reduced, but processing quality may be degraded

Engineering Contradiction:

Improveutility usageVSAvoidprocessing quality

Core Design Contradiction:

Loss of energyVSManufacturing precision

Solution Approach 1:

The system changes operational parameters dynamically based on the created schedule. Instead of arbitrarily adjusting setting values, the schedule creation apparatus determines optimal parameter combinations that achieve both utility suppression and quality maintenance. The trained models have learned the relationships between parameters, utility consumption, and processing quality outcomes.

Inventive Principle:

Principle #35Parameter changes

Solution Approach 2:

The system incorporates feedback mechanisms where the schedule creation apparatus evaluates multiple candidate schedules and selects those that optimize both utility usage and processing quality. The trained models are built using learning input data that includes information about processing outcomes, allowing the system to learn from past performance and make informed decisions about parameter settings.

Inventive Principle:

Principle #23Feedback

3Loss of energy

If multiple schedules are generated using trained models, then utility usage is suppressed, but system complexity increases

Engineering Contradiction:

Improveutility usageVSAvoidsystem complexity

Core Design Contradiction:

Loss of energyVSDevice complexity

Solution Approach 1:

The system performs preliminary training of models using learning input data before actual schedule creation. This preliminary action prepares the system in advance, allowing it to quickly generate optimized schedules during operation. The trained models encapsulate complex relationships learned during the training phase, simplifying the inference phase where schedules are created based on new input data.

Inventive Principle:

Principle #10Preliminary action

Data Source

PatentUS20260094079A1Substrate processing system, schedule creating method, storage medium, and schedule creating program

Publication Date: 2026.04.02 SCREEN HOLDINGS CO LTD

AI summary

A substrate processing includes a substrate processing apparatus and a controller. The controller is configured or programmed to create a schedule of operations of the substrate processing apparatus. The controller includes a storage portion and a controlling portion. The storage portion stores a plurality of trained models for creating a plurality of schedules that differ from each other in the usage amount of the utility. The controlling portion is capable of creating the plurality of schedules based on the plurality of trained models. The plurality of trained models are each constructed by executing reinforcement learning based on learning input data. The learning input data includes substrate count information, recipe information, and utility information. The utility information includes the usage amount of the utility used in each action included in a procedure of the operations of the substrate processing apparatus.