Operation Schedule Selection With Surrogate Model Filtering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing social simulation methods for optimizing measures, such as operation schedules, face challenges with long calculation times and low accuracy due to the limitations of both traditional simulations and surrogate models, making it difficult to find optimal measures effectively.
Innovation Solution
A measure specifying program and apparatus that utilizes a trained neural network to create a surrogate model for pre-filtering social measures, distinguishing between interpolation and extrapolation areas, and combining this with simulation for high accuracy and efficiency in finding optimal measures.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If social simulation is used to evaluate measures, then measurement precision is improved, but loss of time increases
Solution Approach 1:
The patent segments the evaluation process into two distinct phases: a first evaluation using a surrogate model for rapid screening of multiple measures, and a second evaluation using full social simulation only for selected measures. This segmentation allows the system to leverage the speed of surrogate models for initial filtering while reserving the accuracy of simulations for final validation, thereby resolving the contradiction between evaluation accuracy and calculation time.
Solution Approach 2:
The patent applies preliminary action by performing surrogate model evaluation before full social simulation. The surrogate model, trained on historical data, provides preliminary assessment results that guide the selection of measures warranting detailed simulation analysis. This preliminary filtering step eliminates the need to perform computationally expensive simulations on all candidate measures, thus reducing overall calculation time while maintaining evaluation precision for selected measures.
2Loss of time
If surrogate model is used to reduce calculation time, then loss of time is reduced, but measurement precision deteriorates
Solution Approach 1:
The patent merges two different evaluation approaches - surrogate model evaluation and full social simulation evaluation - into a unified two-stage process. The surrogate model provides rapid preliminary assessment, while the social simulation provides accurate final evaluation for selected measures. This combination allows the system to achieve both reduced calculation time (through surrogate model filtering) and maintained measurement precision (through simulation validation of selected measures).
Solution Approach 2:
The patent applies local quality by applying different evaluation methods to different subsets of measures based on their characteristics. High-priority measures or those requiring detailed analysis undergo full social simulation, while other measures receive surrogate model evaluation. This localized application of evaluation methods optimizes the balance between calculation time and measurement precision for different measures according to their specific requirements.
3Reliability
If comprehensive evaluation of multiple measures is performed, then reliability of measure selection is improved, but productivity decreases
Solution Approach 1:
The patent segments the comprehensive evaluation process into two productivity stages: rapid surrogate model screening that evaluates multiple measures simultaneously, followed by focused social simulation on selected measures. This segmentation maintains reliability by ensuring thorough evaluation of promising measures while improving productivity by avoiding exhaustive simulation of all measures, thus resolving the contradiction between selection reliability and optimization efficiency.
Solution Approach 2:
The patent applies partial action by performing full social simulation only on a subset of measures identified as promising by the surrogate model, rather than conducting exhaustive simulation on all candidate measures. This partial evaluation approach maintains sufficient reliability for the final selected measure while significantly improving productivity by reducing the total number of computationally expensive simulation runs required.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
An information processing apparatus receives data for a plurality of social measures that is a consideration target, detects a prediction model constructed from a trained neural network for substituting a simulation related to social measures by using machine learning, performs, based on characteristics of the prediction model, a filtering processing on the data for the plurality of social measures, performs the simulation on data for one or more of the social measures extracted by the filtering processing, specifies, based on a result of the simulation performed on the data for the one or more social measures, a first social measure that is an implementation target from among the plurality of social measures that is the consideration target, and outputs the first social measure.