Hybrid Evolutionary Algorithm for Triple-Patterning Layout Decomposition
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current semiconductor manufacturing technologies, such as double-patterning, struggle to handle design intents with complex and high pattern densities, as they are not capable of validating designs that require triple or higher multiple-patterning technologies due to difficulties in design intent validation.
Innovation Solution
A computer-implemented method using a hybrid evolutionary algorithm to decompose a graph representative of the design into at least three sets, forming a colored graph, which allows for the validation of design intents by assigning shapes to different masks, resolving conflicts through reworking the design and iteratively applying the hybrid evolutionary algorithm to minimize conflicts.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If double-patterning technology is used to increase pattern density, then integration density is improved, but the ability to handle complex design intents with triple or higher pattern density deteriorates due to validation difficulty
Solution Approach 1:
The patent segments the complex graph decomposition problem into smaller sub-problems by dividing the set of shapes into multiple subsets, where each subset is processed independently through iterative applications of the hybrid evolutionary algorithm. This segmentation approach makes the validation of triple-patterning and higher multiple-patterning designs computationally manageable while maintaining the ability to handle complex design intents.
2Quantity of substance
If triple-patterning technology is used to achieve higher pattern density, then integration density is improved, but design intent validation becomes significantly more difficult
Solution Approach 1:
The patent applies preliminary action by performing graph decomposition and conflict detection before actual mask fabrication. The hybrid evolutionary algorithm pre-validates the design intent by attempting to decompose the graph into valid mask assignments, identifying conflicts early in the design phase rather than during manufacturing, thereby reducing validation difficulty.
Solution Approach 2:
The patent implements feedback mechanisms through iterative applications of the hybrid evolutionary algorithm, where each iteration refines the graph decomposition based on conflict detection results. The algorithm uses feedback from previous iterations to adjust decomposition strategies, improving validation efficiency for complex triple-patterning designs while maintaining high pattern density capability.
Data Source
AI summary
A method is presented for layout decomposition including creating a first graph representative of an integrated circuit layout to be multiple-patterned, when a computer is invoked to decompose the layout, and decomposing each of a first subset of a multitude of sub-graphs into at least three sets when a valid coloring solution is returned for the layout. The multitude of sub-graphs is created from the first graph by dividing the first graph. The method further includes approximately decomposing each of the first subset into at least three sets using a hybrid evolutionary algorithm when the hybrid evolutionary algorithm does not return a valid coloring solution for the layout, and forming a colored graph representative of the layout by merging the at least three sets to generate one of at least three colors for each one of a multitude of vertices of the first graph.