Semiconductor Material Scheduling for Real-Time Global Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing material scheduling methods for semiconductor processing equipment are inefficient, as they either consume significant time or only provide local optimal scheduling results, failing to achieve real-time optimal solutions due to exhaustive search methods and lack of global optimization.
Innovation Solution
A material scheduling method that divides materials into groups and subgroups based on process recipes, establishing a scheduling task list and using a solver to calculate an optimal movement sequence, which improves calculation speed and ensures real-time attainment of global optimal scheduling results by selecting and calculating material scheduling task subgroups in batches.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If an enumeration method based on search tree is used to exhaustively list all possible movement orders, then the scheduling optimality is improved, but the calculation time increases significantly making real-time optimization impossible
Solution Approach 1:
The patent segments the exhaustive search space by introducing a search depth parameter that divides the material scheduling problem into multiple levels. At each level, only a subset of possible movements is explored based on the current search depth, thereby reducing the total number of nodes in the search tree while still finding optimal or near-optimal solutions within acceptable time limits.
Solution Approach 2:
The patent applies partial action by performing enumeration only up to a predetermined search depth rather than exhaustively exploring all possible movement orders. This partial exploration is sufficient to achieve optimal scheduling in practice while avoiding the exponential time complexity of complete enumeration, thus enabling real-time optimization.
2Productivity
If an N-step segment scheduling strategy is used to define a limited search range, then the calculation speed is improved, but only local optimal scheduling results can be obtained without global optimization guarantee
Solution Approach 1:
The patent introduces a dynamic search depth parameter that can be adjusted based on problem characteristics and computational resources. Rather than using a fixed N-step limit, the search depth adapts to balance between calculation speed and solution quality, allowing the system to explore deeper when needed while maintaining fast performance when sufficient solutions are found earlier in the search.
Data Source
AI summary
Embodiments of the present disclosure provide a material scheduling method and a material scheduling device for semiconductor processing equipment. The method includes establishing a material list, establishing a first scheduling task list according to process recipes and the material list, and inputting the first scheduling task list into a solver to calculate and output a scheduling result with shortest time for performing all material scheduling tasks in the first scheduling task list and parsing the scheduling result to obtain a movement sequence of all materials. In the technical solutions of the material scheduling method and the material scheduling device for the semiconductor processing equipment of embodiments of the present disclosure, the overall scheduling result can be improved, and the calculation speed can be improved. Thus, the scheduling result can be obtained in real-time.