Learning Management System for Complex Subsystem Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing techniques face challenges in simulating and predicting optimal actions in complex systems like supply chains and energy management systems, especially when dealing with vast types of goods and rapidly changing demands, and are not reliable when data is far from past records.
Innovation Solution
A learning management system that includes record data storage, multiple correlation models, and learning units to determine parameters by simulating behavior, generating measure data, and adjusting parameters based on evaluation logic, allowing for appropriate subsystem control in complex systems.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If simulation is used to predict optimal actions in complex systems, then prediction capability is improved, but it becomes difficult to cover every possible pattern when types of goods and demands are enormous and rapidly changing
Solution Approach 1:
The patent implements a dynamic hybrid prediction system that automatically switches between simulation-based prediction and record data-based prediction based on the characteristics of the input situation. The system evaluates whether the current situation resembles past records and adjusts the prediction method accordingly, making the system adaptable to both familiar and novel scenarios without requiring complete re-simulation of all possible patterns.
Solution Approach 2:
The patent segments the prediction problem into two distinct approaches: simulation-based prediction for novel situations far from past records, and record data-based prediction for familiar situations. This segmentation allows each method to be applied where it is most effective, avoiding the limitation of using simulation for all cases when it cannot cover every possible pattern.
2Productivity
If record data is used to perform prediction, then prediction speed is improved, but the prediction is not reliable at all in a situation which is far from the past record data
Solution Approach 1:
The patent implements a feedback mechanism that evaluates the similarity between the current situation and past record data. Based on this evaluation, the system provides feedback to select the appropriate prediction method: using record data-based prediction when similarity is high (ensuring efficiency), and switching to simulation-based prediction when similarity is low (ensuring reliability). This feedback loop resolves the contradiction by making the prediction method selection adaptive to the situation.
Solution Approach 2:
The patent changes the prediction approach parameter based on the characteristics of the input situation. When the situation is far from past records, the system transitions from record data-based prediction to simulation-based prediction, effectively changing the prediction parameter to maintain reliability while preserving efficiency where possible.
3Adaptability or versatility
If simulation covers every possible pattern in complex systems, then completeness of prediction is improved, but system complexity and computational requirements increase significantly
Solution Approach 1:
The patent applies partial action by using simulation only when necessary (i.e., when the situation is far from past records and record data-based prediction is unreliable). Instead of performing complete simulation for all possible patterns, the system uses simulation selectively for novel situations, reducing computational complexity while maintaining prediction completeness where needed.
Solution Approach 2:
The patent uses record data as a copy of past situations to make predictions for familiar scenarios, avoiding the need to perform simulation for every possible pattern. This copying approach reduces system complexity by leveraging historical data instead of requiring comprehensive simulation coverage for all patterns.
Data Source
AI summary
A retail agent determines a parameter of an activity proposal model by using data stored in a past record database, and determines parameters of an activity determination model and an activity value evaluation model by further using base activity simulation data. Consequently, it is possible to appropriately determine parameters of a subsystem control method in a complex system which cannot be embodied as a simulator and shows a significant change with respect to past record data.