Monte Carlo Simulation Scenario Clustering and Transition Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing simulation systems, such as those described in WO2016/194051, fail to provide an effective method for users to easily analyze and display many scenarios generated by Monte Carlo simulations, particularly in identifying important junction points or index vectors at specific time steps.
Innovation Solution
A simulation system comprising a processor, memory, and display device that classifies index vectors into clusters, calculates transition probabilities, and displays an analysis map showing transition probabilities and transmissible clusters, enabling users to easily analyze and visualize scenario transitions across time steps.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If many scenarios are generated by Monte Carlo simulation, then the simulation accuracy and comprehensiveness are improved, but the difficulty of analyzing and displaying the scenarios increases
Solution Approach 1:
The patent segments the scenarios into clusters based on their characteristics at each time step. By dividing the large set of scenarios into manageable clusters, the system maintains simulation accuracy while making analysis more feasible. The clustering process organizes complex scenario data into structured groups that can be easily displayed and interpreted.
Solution Approach 2:
The patent introduces a temporal dimension to the analysis by displaying scenarios across multiple time steps in a structured format. The display mechanism organizes scenarios by time step and cluster, adding a dimensional structure that transforms the complexity of many scenarios into an organized, navigable presentation that maintains accuracy while improving analyzability.
2Loss of information
If many scenarios are displayed in detail, then the user can see all possible outcomes, but the user cannot easily identify important junction points or key information
Solution Approach 1:
The patent segments scenarios into clusters and further organizes them by time steps, creating a hierarchical structure that preserves complete information while making it accessible. Users can navigate through time steps and clusters to find important junction points without being overwhelmed by the full scenario set.
Solution Approach 2:
The system performs preliminary clustering and organization of scenarios before display, pre-structuring the information to highlight important patterns and junction points. This preliminary processing maintains information completeness while preparing the data for easy user analysis and identification of key moments.
3Ease of operation
If scenarios are organized by time steps and clusters, then the display becomes more manageable and analytical, but the complexity of processing and calculating transition probabilities increases
Solution Approach 1:
The patent segments the processing into discrete steps: first clustering scenarios at each time step, then calculating transition probabilities between clusters across time steps. This segmentation of the complex processing task into manageable stages reduces overall computational complexity while maintaining the structured, manageable display format.
Solution Approach 2:
The system performs clustering as a preliminary action before calculating transition probabilities. By pre-organizing scenarios into clusters at each time step, the subsequent calculation of transition probabilities becomes more efficient and manageable, reducing the overall processing complexity while maintaining display organization.
Data Source
AI summary
A simulation system stores information representing a plurality of scenarios of index vectors generated by simulation and a plurality of clusters in which the index vectors in each of a plurality of time steps are classified, classifies the clusters of a final time step of the plurality of scenarios into a plurality of groups, calculates a transition probability at which the index vectors belonging to the cluster in the time step transition to each group in the final time step for each of combinations of the time steps and the clusters and a cluster in which the index vectors belonging to the cluster in the time step are transmissible in a subsequent step, and displays the transition probability and the transmissible cluster.