Partial Enumeration MPC for MIMO Systems
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing model predictive control (MPC) systems for multiple input, multiple output (MIMO) processes are inefficient due to slow computation times, especially in large-scale processes with long time delays or fast sampling rates, and often rely on complete enumeration strategies that become impractical due to rapidly growing computational demands.
Innovation Solution
A partial enumeration model predictive controller that provides a solution table for a partial parameter region, scans for an optimal solution, and uses either the optimal or an alternative solution to control the MIMO system, with the ability to update and manage the solution table to maintain efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If complete enumeration strategies are used to solve MPC systems over the whole parameter region, then the optimal solution can be found, but the computation time increases rapidly and becomes unsuitable for real-time control
Solution Approach 1:
The parameter region is divided into multiple sub-regions, and solution tables are created for each sub-region separately. This segmentation allows the controller to search within smaller, manageable regions rather than the entire parameter space, significantly reducing computation time while maintaining solution quality for each local region.
Solution Approach 2:
Instead of enumerating all possible solutions over the complete parameter region, the method uses partial enumeration by creating solution tables that cover only selected sub-regions. This partial action approach provides a practical trade-off between computational feasibility and control performance.
2Adaptability or versatility
If solution tables are enlarged to cover the whole parameter region, then more optimal solutions are available, but the table size becomes unmanageable and searching becomes impractical
Solution Approach 1:
The large parameter region is segmented into multiple smaller sub-regions, each with its own solution table. This division makes the solution tables manageable in size while collectively covering the entire parameter region, allowing the controller to adapt to different operating conditions by selecting the appropriate sub-region table.
Solution Approach 2:
The approach introduces a new dimension of organization by dividing the parameter space into discrete sub-regions. This dimensional change allows the system to manage complexity through structured partitioning rather than attempting to handle the full parameter space as a single entity.
3Adaptability or versatility
If MPC systems are designed for large-scale processes with long time delays or fast sampling times, then they can handle complex processes, but the computation speed becomes too slow for real-time control
Solution Approach 1:
By segmenting the parameter region into sub-regions and creating dedicated solution tables for each, the system can quickly locate and apply pre-computed solutions without performing full-scale optimization for every control decision. This significantly improves computation speed while maintaining the ability to handle large-scale processes with complex dynamics.
Data Source
AI summary
A partial enumeration model predictive controller and method of predictive control for a multiple input, multiple output (MIMO) system, including providing a solution table with problem solutions to a model predictive control problem for the MIMO system over a partial parameter region; scanning the solution table for an optimal solution for current parameters; using the optimal solution to control the MIMO system when the optimal solution is in the solution table; and using an alternative solution to control the MIMO system when the optimal solution is not in the solution table.