Sequence Simulator Using Variable-Length Substring Partitioning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional random number generators fail to reproduce the statistical characteristics of original data sequences, making it difficult to generate simulated sequences for testing and graphics applications that require accurate replication of real data characteristics.
Innovation Solution
A process that partitions an input sequence into substrings of varying lengths and rearranges them randomly to produce a simulated sequence with the same statistical properties, using techniques like the Lempel-Ziv incremental parsing rule and tree structures to ensure similarity and randomness.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional random number generators are used to generate sequences, then the generation process is simple and fast, but the statistical characteristics of the generated sequences do not match the original data
Solution Approach 1:
The patent divides the original sequence into multiple substrings of varying lengths using parsing rules (such as Lempel-Ziv parsing). These substrings are then randomly permuted to create the simulated sequence. This segmentation approach preserves statistical characteristics by maintaining the distribution of substring lengths and patterns while introducing randomness through reordering.
Solution Approach 2:
The patent performs preliminary parsing of the original sequence to extract substrings and their length distributions before generating the simulated sequence. This preliminary analysis of the original sequence's statistical properties enables the generation of fake sequences that accurately replicate those properties without requiring complex real-time computation.
2Manufacturing precision
If substrings of equal length are used in the sequence, then the generation process is simple, but visual characteristics and statistical properties are compromised
Solution Approach 1:
The patent applies different parsing strategies to create substrings of varying lengths at different positions in the sequence. By allowing substrings to have different lengths (e.g., through Lempel-Ziv parsing which creates phrases of variable length), the method preserves local statistical properties and visual characteristics of the original sequence while maintaining overall randomness through random permutation.
3Reliability
If tiling processes are used to create textures, then the generation process is efficient, but visual characteristics at boundaries and repeated patterns become noticeable
Solution Approach 1:
The patent segments the texture into multiple substrings of varying lengths and randomly permutes them to create a seamless pattern. This eliminates the repetitive tiling effect by ensuring that no two regions have identical patterns, while the random permutation maintains computational efficiency compared to generating entirely new textures.
Solution Approach 2:
The patent creates a simulated texture by copying and rearranging substrings from the original texture rather than generating entirely new pixel values. This copying approach preserves the statistical and visual characteristics of the original texture while avoiding the repetitive patterns associated with tiling, as the random permutation ensures unique arrangements.
Data Source
AI summary
A process for simulating an input sequence partitions the input sequence into a set of substrings and a tail. The substrings preferably have lengths that are not all equal. An output sequence then includes the substrings in a random order.