Partial enumeration model predictive controller

1. A method of predictive control for a multiple input, multiple output(MIMO) system, the method employing at least one processor, comprising:

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 with the at least one processor 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;

wherein the providing comprises:

modeling the MIMO system with model factors;

generating the model predictive control problem from the model factor;

determining the partial parameter for an expected operating state;

solving the model predictive control problem offline for the problem solutions in the partial parameter region; and

population the solution table with the problem solutions.

2. The method of claim 1 wherein the generating comprises formulating a target calculation selected from the group consisting of linear objective functions, quadratic objective functions, hard constraint functions with precedence rank ordering, soft constraint functions, and multi-stage quadratic objective functions.

3. The method of claim 1 wherein the solving comprises:

adding random disturbances and setpoint changes to force the model predictive control problem into the partial parameter region; and

solving the model predictive control problem over a number of sampling times.

4. The method of claim 1 wherein the alternative solution comprises an alternative solution selected from the group consisting of a best suboptimal solution, a predetermined time current solution, and a simplified model current solution.

5. The method of claim 1 wherein the alternative solution comprises a simplified model current solution selected from the group consisting of a shortened time horizon solution and an additionally constrained solution.

6. The method of claim 1 further comprising:

calculating a current solution to the model predictive control problem for the MIMO system for the current parameters when the optimal solution is not in the solution table; and

adding the current solution to the solution table.

7. The method of claim 6 further comprising deleting one of the problem solutions.

8. The method of claim 6 further comprising adding a solution family related to the current solution to the solution table.

9. A computer readable medium storing a computer program for predictive control for a multiple input, multiple output (MIMO) system, comprising:

computer readable code for providing a solution table with problem solutions to a model predictive control problem for the MIMO system over a partial parameter region;

computer readable code for scanning the solution table for an optimal solution for current parameters;

computer readable code for using the optimal solution to control the MIMO system when the optimal solution is in the solution table; and

computer readable code for using an alternative solution to control the MIMO system when the optimal solution is not in the solution table;

wherein the computer readable code for providing comprises:

computer readable code for modeling the MIMO system with model factors;

computer readable code for generating the model predictive control problem from the model factors;

computer readable code for determining the partial parameter region for an expected operating state;

computer readable code for solving the model predictive control problem offline for the problem solutions in the partial parameter region; and

computer readable code for populating the solution table with the problem solutions.

10. The computer readable medium of claim 9 wherein the computer readable code for generating comprises computer readable code for formulating a target calculation selected from the group consisting of linear objective functions, quadratic objective functions, hard constraint functions with precedence rank ordering, soft constraint functions, and multi-stage quadratic objective functions.

11. The computer readable medium of claim 9 wherein the computer readable code for solving comprises:

computer readable code for adding random disturbances and setpoint changes to force the model predictive control problem into the partial parameter region; and

computer readable code for solving the model predictive control problem over a number of sampling times.

12. The computer readable medium of claim 9 wherein the alternative solution comprises an alternative solution selected from the group consisting of a best suboptimal solution, a predetermined time current solution, and a simplified model current solution.

13. The computer readable medium of claim 9 wherein the alternative solution comprises a simplified model current solution selected from the group consisting of a shortened time horizon solution and an additionally constrained solution.

14. The computer readable medium of claim 9 further comprising:

computer readable code for calculating a current solution to the model predictive control problem for the MIMO system for the current parameters when the optimal solution is not in the solution table; and

computer readable code for adding the current solution to the solution table.

15. The computer readable medium of claim 14 further comprising computer readable code for deleting one of the problem solutions.

16. The computer readable medium of claim 14 further comprising computer readable code for adding a solution family related to the current solution to the solution table.

17. A method of predictive control for a multiple input, multiple output (MIMO) system, the method employing at least one processor, comprising:

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 with the at least one processor 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;

using an alternative solution to control the MIMO system when the optimal solution is not in the solution table;

calculating a current solution to the model predictive control problem for the MIMO system for the current parameters when the optimal solution is not in the solution table;

adding the current solution to the solution table; and

deleting one of the problem solutions, the deleting comprising deleting a problem solution selected from group consisting of a least used solution and a least recently used solution.

18. A method of predictive control for a multiple input, multiple output (MIMO) system, the method employing at least one processor, comprising:

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 with the at least one processor 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;

using an alternative solution to control the MIMO system when the optimal solution is not in the solution table; and

determining a use score for each of the problem solutions over a predetermined number of sampling times;

wherein the scanning comprises scanning the problem solutions in order of the use score from most used to least used.

19. A method of predictive control for a multiple input, multiple output (MIMO) system, the method employing at least one processor comprising:

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 with the at least one processor 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;

using an alternative solution to control the MIMO system when the optimal solution is not in the solution table; and

determining an age score for each of the problem solutions;

wherein the scanning comprises scanning the problem solutions in order of the age score from newest to oldest.

20. A computer readable medium storing a computer program for predictive control for a multiple input, multiple output (MIMO) system, comprising:

computer readable code for providing a solution table with problem solutions to a model predictive control problem for the MIMO system over a partial parameter region;

computer readable code for scanning the solution table for an optimal solution for current parameters;

computer readable code for using the optimal solution to control the MIMO system when the optimal solution is in the solution table;

computer readable code for using an alternative solution to control the MIMO system when the optimal solution is not in the solution table;

computer readable code for calculating a current solution to the model predictive control problem for the MIMO system for the current parameters when the optimal solution is not in the solution table;

computer readable code for adding the current solution to the solution table; and

computer readable code for deleting one of the problem solutions;

wherein the computer readable code for deleting comprises computer readable code for deleting a problem solution selected from group consisting of a least used solution and a least recently used solution.

21. A computer readable medium storing a computer program for predictive control for a multiple input, multiple output (MIMO) system, comprising:

computer readable code for providing a solution table with problem solutions to a model predictive control problem for the MIMO system over a partial parameter region;

computer readable code for scanning the solution table for an optimal solution for current parameters;

computer readable code for using the optimal solution to control the MIMO system when the optimal solution is in the solution table;

computer readable code for using an alternative solution to control the MIMO system when the optimal solution is not in the solution table; and

computer readable code for determining a use score for each of the problem solutions over a predetermined number of sampling times;

wherein the computer readable code for scanning comprises computer readable code for scanning the problem solutions in order of the use score from most used to least used.

22. A computer readable medium storing a computer program for predictive control for a multiple input, multiple output (MIMO) system, comprising:

computer readable code for providing a solution table with problem solutions to a model predictive control problem for the MIMO system over a partial parameter region;

computer readable code for scanning the solution table for an optimal solution for current parameters;

computer readable code for using the optimal solution to control the MIMO system when the optimal solution is in the solution table;

computer readable code for using an alternative solution to control the MIMO system when the optimal solution is not in the solution table; and

computer readable code for determining an age score for each of the problem solutions;

wherein the computer readable code for scanning comprises computer readable code for scanning the problem solutions in order of the age score from newest to oldest.