Radix Tree Compression for Manufacturing Process Chains
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing methods for storing and accessing execution graphs in manufacturing systems are inefficient due to high memory usage and slow access times, particularly when dealing with large numbers of redundant subgraphs, and current compression techniques do not adequately address these issues.
Innovation Solution
The method involves mapping process chains into a shared memory data structure using radix trees, which allows for efficient compression and quick access by identifying and merging shared nodes and edges across multiple process chains, thereby reducing memory footprint and improving access speeds.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If execution graphs are stored in a list format, then access simplicity is maintained, but memory usage increases significantly due to redundant subgraphs
Solution Approach 1:
The patent merges identical subgraphs across multiple execution graphs by introducing a canonical representation system. When subgraphs are detected to be identical through isomorphism checking, they are merged into a single stored instance with references from multiple parent graphs, eliminating redundant storage while maintaining access simplicity through the unified data structure.
2Quantity of substance
If compression techniques like BDDs or MDDs are applied to process chains, then memory efficiency improves, but compression effectiveness is limited for complex diagram structures
Solution Approach 1:
The patent changes the representation parameters of process chains by introducing a hierarchical canonical form with standardized node labeling and ordering. This transformation enables more effective compression by creating a consistent parameter space where identical subgraphs have identical representations, allowing for superior compression ratios compared to traditional BDD/MDD approaches.
3Quantity of substance
If file-level binary compression (e.g., ZIP) is used, then memory footprint is reduced, but access speed decreases due to slow lookup and query operations
Solution Approach 1:
The patent performs preliminary action by pre-processing execution graphs into a canonical compressed format during the storage phase, organizing data in a way that enables efficient access. The canonical representation with standardized node ordering and hierarchical structure allows for rapid lookup and query operations without requiring decompression, thus maintaining both small memory footprint and fast access speed.
4Adaptability or versatility
If the number of execution graphs increases to improve flexibility and machine park utilization, then production flexibility improves, but storage capacity is exceeded
Solution Approach 1:
The patent uses copying references instead of duplicating actual graph data. Multiple execution graphs can reference the same canonical subgraph instances, allowing the system to support a large number of execution graphs for improved production flexibility while storing only unique subgraphs once in memory, thus avoiding storage capacity exhaustion.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Method for producing a product selected from a set of products, which set of products comprises a first and at least one second product, each defined by a process chain (G1-G4), wherein a process data structure in the form of a directed, acyclic graph describes the respective process chain (G1-G4), and the process chains (G1-G4) are stored in a common memory, the selected process chain is retrieved from the memory by a control device, and a production means with elements (AF, 1-9) of the selected process chain is controlled by the control device and the selected product is produced, wherein the process chains (G1-G4) are mapped into a single, common memory data structure (R) if the process data structure for the first and the at least one second product is identical in at least one respective subset and the memory data structure (R) is stored in memory.