Hybrid Process Prediction Model for Physical Constraint Compliance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing data-based probabilistic models for predicting the time evolution of variables influenced by technical processes often lack the ability to effectively incorporate prior knowledge about the process, leading to predictions that may violate known laws or boundary conditions.
Innovation Solution
A method that utilizes an existing probabilistic model in conjunction with a process model that represents prior knowledge about the given process. This method assesses candidates for future values of a variable by considering a history that includes both previous values of the variable and context information, and assigns scores based on compatibility with known laws and boundary conditions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If data-based probabilistic models are used to predict future values based on historical data, then prediction capability is improved, but predictions may violate known laws or boundary conditions
Solution Approach 1:
The patent merges data-based probabilistic models with process models that encode prior knowledge about physical laws and boundary conditions. The hybrid system combines the pattern recognition strength of probabilistic models with the constraint enforcement capability of process models, ensuring predictions are both accurate and physically plausible.
Solution Approach 2:
The patent introduces an intermediary scoring mechanism that evaluates candidate predictions against known laws and boundary conditions. This scoring system acts as a mediator between the probabilistic model's predictions and the physical constraints, filtering or adjusting predictions that violate established principles.
2Measurement precision
If comprehensive history and context information are incorporated into predictions, then prediction accuracy is improved, but computational effort increases
Solution Approach 1:
The patent performs preliminary action by pre-processing and structuring historical and contextual information before the actual prediction process. By organizing data in advance and pre-evaluating candidate predictions against constraints, the system reduces the computational burden during the main prediction phase.
Solution Approach 2:
The patent segments the prediction process into distinct stages: generating candidate predictions from probabilistic models, evaluating them against process constraints, and selecting the best candidates. This segmentation allows parallel processing and optimization of each stage independently, improving overall computational efficiency.
Data Source
AI summary
A method for predicting the time evolution of a variable x that is influenced by a given process. The method includes: proceeding from the history Vt for the time step t, k candidates xt+11, . . . , xt+1k are ascertained for the value xt+1 of the variable x in the time step t+1; for candidates xt+11, . . . , xt+1k, scores st+11, . . . , st+1k are ascertained in cooperation between a probabilistic model and a process model that represents prior knowledge about the given process; from the set of candidates xt+11, . . . , xt+1k, a proper subset xt+1i, i∈I⊂{1, . . . , k}, is selected based on the associated scores st+11, . . . , st+1k; proceeding from new selected candidates xt+1i for the time step t+1, l candidates xt+21, . . . , xt+2l are ascertained for the value xt+2 of the variable x in the time step t+2.