Nondeterministic Graph Query via Probability Calculation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current query algorithms and index mechanisms for large-scale graph data are ineffective for nondeterministic graphs due to their deterministic structure basis, leading to high overhead and cost in query processing.
Innovation Solution
A method and apparatus for querying nondeterministic graphs by determining vertices and calculating probabilities of path events, allowing for quick query results by sorting paths and calculating probabilities of first and second events based on edge existence, thereby simplifying query conditions and reducing complexity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If graph searching is performed on exponential graph instances derived from a nondeterministic graph to ensure probability guarantee, then query accuracy is improved, but query complexity and processing cost increase exponentially
Solution Approach 1:
The patent segments the exponential graph instances into smaller subgraphs by identifying and extracting connected components. Instead of processing all exponential instances globally, the system divides them into manageable subgraphs that can be processed independently, reducing the overall complexity while maintaining probability guarantees through localized processing.
Solution Approach 2:
The patent introduces dynamic thresholding mechanisms that adjust processing depth and probability calculation granularity based on query requirements and graph characteristics. The system dynamically adapts between exact probability calculation (when high accuracy is needed) and approximate methods (when efficiency is prioritized), creating a flexible query processing framework.
2Ease of manufacture
If deterministic graph query algorithms are applied to nondeterministic graphs, then implementation simplicity is improved, but query reliability deteriorates due to inability to handle probability-based edge existence
Solution Approach 1:
The patent introduces probability annotations as an intermediary layer between deterministic graph algorithms and nondeterministic graph data. The system wraps deterministic algorithms with probability calculation modules that inject probability information at critical points, allowing standard algorithms to work on nondeterministic graphs while maintaining reliability through probability-aware modifications.
Solution Approach 2:
The patent transforms nondeterministic graph problems into deterministic equivalents by changing the parameter representation from probability distributions to concrete instance realizations. The system samples or determinizes the probabilistic edges according to given thresholds, converting the problem into a form that deterministic algorithms can handle reliably.
3Measurement precision
If all possible paths are enumerated to calculate exact probabilities, then measurement precision is improved, but loss of time increases due to exponential path enumeration
Solution Approach 1:
The patent implements partial path enumeration by limiting the depth and breadth of path exploration based on probability thresholds and query constraints. Instead of enumerating all possible paths, the system selectively explores only those paths that meet minimum probability criteria or contribute significantly to the query result, achieving acceptable accuracy with reduced time cost.
Solution Approach 2:
The patent performs preliminary probability estimation and path filtering before exact probability calculation. The system pre-processes the graph to identify high-probability paths and prune low-probability branches, preparing a reduced search space that maintains accuracy for important paths while eliminating time-consuming exploration of negligible paths.
Data Source
AI summary
The present invention discloses a method and apparatus for querying a nondeterministic graph, which are used to implement quick query of a nondeterministic graph, reduce query complexity, and improve query efficiency. The method comprises receiving a query instruction, where the query instruction is used to query a nondeterministic graph for data that satisfies a query condition; determining two vertices in the nondeterministic graph according to the query instruction; determining all possible paths that use one vertex in the two vertices as a start point and the other vertex as an end point; calculate a probability of a first event or a second event corresponding to each of the paths; and obtaining, according to the probability of the first event or the probability of the second event, a query result corresponding to the query instruction.