Deterministic Optimization Control for Real-Time Trajectory Selection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Optimization-based control (OBC) systems for complex dynamic control systems are computationally demanding, making them inefficient for real-time control due to the complexity of solving constrained optimization problems, particularly quadratic programming issues, which can take seconds or minutes to solve, and are challenging to predict within a predetermined time frame.
Innovation Solution
A deterministic optimization-based control method that involves determining a linear approximation of a non-linear process model, convex approximation of constraints, and using a stabilization function to produce a stable feasible solution within a predetermined time window, employing a feasible search strategy like the primal active set method to efficiently handle simple bounds, reducing computational complexity and enabling real-time control.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If general optimization solvers are used to solve constrained optimization problems, then optimal control solutions can be obtained, but the computation time becomes unpredictable and may take seconds or minutes
Solution Approach 1:
The patent segments the control horizon into multiple prediction steps and divides the optimization problem into smaller sub-problems that can be solved more quickly. By breaking down the complex constrained optimization into manageable segments, the system achieves faster computation while maintaining solution quality.
Solution Approach 2:
The patent performs preliminary actions by pre-calculating feasible control trajectories and storing them for future use. When real-time control is needed, the system selects from pre-computed trajectories rather than solving the full optimization problem from scratch, dramatically reducing computation time while ensuring feasibility.
2Productivity
If deterministic optimization-based control is implemented to provide feasible control actions within predetermined time, then real-time control capability is achieved, but the system complexity increases
Solution Approach 1:
The patent implements dynamic adjustment of control strategies based on system state and time constraints. The controller adapts its approach by selecting from multiple pre-computed trajectories or adjusting optimization parameters in real-time, achieving deterministic performance without requiring overly complex fixed-structure systems.
Solution Approach 2:
The patent introduces an intermediary layer that mediates between the optimization algorithm and the control system. This intermediary manages the complexity by handling trajectory selection, validation, and adjustment, allowing the core control system to remain relatively simple while achieving sophisticated real-time optimization capabilities.
3Productivity
If optimization-based control is used to operate closer to constraints, then system performance is improved, but the computational demand increases significantly
Solution Approach 1:
The patent applies partial optimization by focusing computational efforts only on the most critical control variables and time steps. Rather than optimizing all variables to full precision, the system achieves sufficient performance by optimizing key parameters partially, reducing computational energy demand while maintaining effective control near constraints.
Data Source
Figure 1~2
Figure 3~5B
Figure 4
AI summary
The embodiments described herein include one embodiment that provides a control method, including determining a first stabilizing feasible control trajectory of a plurality of variables of a controlled process, determining a second stabilizing feasible control trajectory for the plurality of variables for a second time step subsequent to the first time step, determining a first cost of applying the first feasible control trajectory at the second time step, determining a second cost of applying the second feasible control trajectory at the second time step, comparing the first and second costs, selecting the first feasible control trajectory or the second feasible control trajectory based upon the comparison in a predetermined time frame, and controlling the controlled process by cost of the selected control trajectory.