Rule Set Minimization via Application and Violation Graphs
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing business rule management systems struggle to effectively eliminate redundancy in rule sets, especially when dealing with rules involving attributes with infinite domains and complex decision policies, leading to inefficiencies in decision automation and rule management.
Innovation Solution
A method is introduced to build a rule set application graph and a rule violation graph, allowing for the identification of a reduced subset of rules that reproduces the original rule set's behavior without redundancy, using a processor to explore candidate subsets and prioritize rule elimination based on preferences.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If rules are automatically generated by rule learning and data mining tools or assembled from different sources, then the decision automation capability is enhanced, but redundant rules are introduced into the rule set
Solution Approach 1:
The patent extracts and removes redundant rules from the rule set through a dedicated rule analysis module. The system identifies rules that are subsumed by other rules (where one rule's condition and action are fully covered by another rule) and eliminates them, thereby reducing the quantity of redundant rules while preserving the decision automation capability provided by the remaining rules.
Solution Approach 2:
The patent discards redundant rules that are automatically generated or assembled from multiple sources. By analyzing the logical relationships between rules, the system identifies and discards rules that do not add unique decision-making value, thereby reducing rule set size while maintaining the essential decision automation functionality.
2Quantity of substance
If the number of rules is reduced by making rules more general, then the rule set size is minimized, but the precision of decision-making for specific cases may be compromised
Solution Approach 1:
The patent applies universality by making rules as general as possible while ensuring they remain precise through logical subsumption analysis. General rules that cover multiple cases are preferred, and the system identifies when a general rule subsumes specific rules, allowing the specific rules to be removed. This maintains decision precision for specific cases while minimizing rule set size through universal, multi-functional rules.
Solution Approach 2:
The patent merges specific rules into more general rules when the general rule's condition and action fully cover the specific rule's condition and action. By combining multiple specific rules into a single general rule that subsumes them, the system reduces rule set size while maintaining the same decision-making precision for all covered cases.
3Adaptability or versatility
If rules are overlapping to handle diverse cases, then the decision coverage is improved, but the complexity of rule management increases
Solution Approach 1:
The patent segments the rule set into distinct, non-redundant rules through subsumption analysis. By identifying and removing rules that are fully covered by other rules, the system creates a segmented structure where each remaining rule provides unique decision-making value. This reduces management complexity while maintaining comprehensive decision coverage through the segmented, non-overlapping rule set.
Solution Approach 2:
The patent inverts the traditional approach by not trying to manage overlapping rules but instead systematically removing them through subsumption analysis. Rather than adding rules to handle diverse cases and managing the resulting complexity, the system starts with a comprehensive rule set and eliminates redundant rules, thereby reducing management complexity while preserving decision coverage through the remaining essential rules.
Data Source
AI summary
A method, system and computer program product for minimizing a rule set. A rule set application graph that describes a set of rules including actions and cases is built, whereby each rule defines an action for one or more cases. A rule violation graph for each rule describing the actions and cases not defined for that rule is built. A subset of rules having a reduced number of rules with the same set of actions and cases as the rule set application graph is determined by exploring candidate subsets of the rule set that are sufficient for reproducing the behavior of the original rule set.