Systems and methods for efficiently updating solutions to multi-objective hierarchical linear programming problems

1. A system of efficiently solving subsequent runs of a supply chain planning problem to reduce computer solve time, wherein the supply chain planning problem is modeled as a multi-objective hierarchical linear programming problem, comprising:

the computer, comprising a processor and memory, the computer solves the supply chain planning problem by:

receiving supply chain input data for the supply chain planning problem;

modeling the supply chain planning problem as a multi-objective hierarchal linear programming problem having a first objective and at least one additional objective, wherein the first objective and the at least one additional objective are ordered and defined by an objective hierarchy, and based, at least in part, on the supply chain input data;

solving a base run of the multi-objective hierarchical linear programming problem for the first objective and the at least one additional objective;

generating a base plan by converting a solution of the base run of the multi-objective hierarchical linear programming problem;

receiving one or more changes to the supply chain input data;

identifying primal feasibility changes of the one or more changes that affect only primal feasibility;

identifying dual feasibility changes of the one or more changes that affect only dual feasibility;

loading the multi-objective hierarchical linear programming problem and a first optimal basis of the first objective of the base run;

applying the dual feasibility changes;

modifying the first optimal basis of the first objective of the base run to generate a first starting basis for a new run;

solving the multi-objective hierarchical linear programming problem for the first objective of the new run using the first starting basis and a primal simplex method;

modifying a second optimal basis of the first objective of the new run to generate a second starting basis;

applying the primal feasibility changes;

solving the multi-objective hierarchical linear programming problem for the first objective using a dual simplex method and the second starting basis;

iteratively solving, according to the order specified by the objective hierarchy, each of the at least one additional objective of the multi-objective hierarchical linear programming problem for the new run, wherein decision variables for each subsequent objective are fixed at upper and lower bounds corresponding to objectives higher in the objective hierarchy that have already been solved; and

generating a new supply chain plan based, at least in part, on a solution of the new run.

2. The system of claim 1 , wherein the computer:

monitors for one or more supply chain changes that impact the base plan; and

when the one or more supply chain changes that impact the base plan are detected, calculate the one or more changes to the supply chain input data based, at least in part, on the one or more supply chain changes that impact the base plan.

3. The system of claim 2 , wherein the computer:

identifies current inventory levels of items at one or more supply chain entities; and

causes one or more items to be transported by a transportation network to fulfill the supply chain plan of the new run.

4. The system of claim 3 , wherein the dual feasibility changes comprise one or more of an addition of a new variable, a change in an objective coefficient, and an addition of a constraint comprising a new variable.

5. The system of claim 4 , wherein the primal feasibility changes comprise one or more of an addition of a new constraint, a change in a lower bounds, a change in an upper bounds, a change in a right hand side value, a change in a lower bounds to undo variable fixing based on the base run and to apply variable fixing based on previous objectives solved in the new run; and a change in an upper bounds to undo variable fixing based on the base run and to apply variable fixing based on previous objectives solved in the new run.

6. The system of claim 5 , wherein modifying a need date of an existing demand comprises addition of a new demand, an adjustment of a priority of a demand by changing an objective coefficient of an existing demand to zero, and an adjustment of a need quantity of an existing demand to zero.

7. The system of claim 6 , wherein addition of a new lateable demand comprises modifying a starting basis to retain primal feasibility by:

setting a newly added constraint at the lower bound in the starting basis; and

setting at least one variable in the newly added constraint as a basic variable in the starting basis.

8. A method of efficiently solving subsequent runs of a supply chain planning problem to reduce computer solve time, wherein the supply chain planning problem is modeled as a multi-objective hierarchical linear programming problem, comprising:

receiving, by the computer comprising a processor and memory, supply chain input data for the supply chain planning problem;

modeling the supply chain planning problem as a multi-objective hierarchal linear programming problem having a first objective and at least one additional objective, wherein the first objective and the at least one additional objective are ordered and defined by an objective hierarchy, and based, at least in part, on the supply chain input data;

solving a base run of the multi-objective hierarchical linear programming problem for the first objective and the at least one additional objective;

generating a base plan by converting a solution of the base run of the multi-objective hierarchical linear programming problem;

receiving one or more changes to the supply chain input data;

identifying primal feasibility changes of the one or more changes that affect only primal feasibility;

identifying dual feasibility changes of the one or more changes that affect only dual feasibility;

loading the multi-objective hierarchical linear programming problem and a first optimal basis of the first objective of the base run;

applying the dual feasibility changes;

modifying the first optimal basis of the first objective of the base run to generate a first starting basis for a new run;

solving the multi-objective hierarchical linear programming problem for the first objective of the new run using the first starting basis and a primal simplex method;

modifying a second optimal basis of the first objective of the new run to generate a second starting basis;

applying the primal feasibility changes;

solving the multi-objective hierarchical linear programming problem for the first objective using a dual simplex method and the second starting basis;

iteratively solving, according to the order specified by the objective hierarchy, each of the at least one additional objective of the multi-objective hierarchical linear programming problem for the new run, wherein decision variables for each subsequent objective are fixed at upper and lower bounds corresponding to objectives higher in the objective hierarchy that have already been solved; and

generating a new supply chain plan based, at least in part, on a solution of the new run.

9. The method of claim 8 , further comprising:

monitoring for one or more supply chain changes that impact the base plan; and

when the one or more supply chain changes that impact the base plan are detected, calculating the one or more changes to the supply chain input data based, at least in part, on the one or more supply chain changes that impact the base plan.

10. The method of claim 9 , further comprising:

identifying current inventory levels of items at one or more supply chain entities; and

causing one or more items to be transported by a transportation network to fulfill the supply chain plan of the new run.

11. The method of claim 10 , wherein the dual feasibility changes comprise one or more of an addition of a new variable, a change in an objective coefficient, and an addition of a constraint comprising a new variable.

12. The method of claim 11 , wherein the primal feasibility changes comprise one or more of an addition of a new constraint, a change in a lower bounds, a change in an upper bounds, a change in a right hand side value, a change in a lower bounds to undo variable fixing based on the base run and to apply variable fixing based on previous objectives solved in the new run; and a change in an upper bounds to undo variable fixing based on the base run and to apply variable fixing based on previous objectives solved in the new run.

13. The method of claim 12 , wherein modifying a need date of an existing demand comprises addition of a new demand, an adjustment of a priority of a demand by changing an objective coefficient of an existing demand to zero, and an adjustment of a need quantity of an existing demand to zero.

14. The method of claim 13 , wherein addition of a new lateable demand comprises modifying a starting basis to retain primal feasibility by:

setting a newly added constraint at the lower bound in the starting basis; and

setting at least one variable in the newly added constraint as a basic variable in the starting basis.

15. A non-transitory computer-readable medium embodied with software for efficiently solving subsequent runs of a supply chain planning problem to reduce computer solve time, wherein the supply chain planning problem is modeled as a multi-objective hierarchical linear programming problem, wherein the software when executed by the computer:

receives supply chain input data for the supply chain planning problem;

models the supply chain planning problem as a multi-objective hierarchal linear programming problem having a first objective and at least one additional objective, wherein the first objective and the at least one additional objective are ordered and defined by an objective hierarchy, and based, at least in part, on the supply chain input data;

solves a base run of the multi-objective hierarchical linear programming problem for the first objective and the at least one additional objective;

generates a base plan by converting a solution of the base run of the multi-objective hierarchical linear programming problem;

receives one or more changes to the supply chain input data;

identifies primal feasibility changes of the one or more changes that affect only primal feasibility;

identifies dual feasibility changes of the one or more changes that affect only dual feasibility;

loads the multi-objective hierarchical linear programming problem and a first optimal basis of the first objective of the base run;

applies the dual feasibility changes;

modifies the first optimal basis of the first objective of the base run to generate a first starting basis for a new run;

solves the multi-objective hierarchical linear programming problem for the first objective of the new run using the first starting basis and a primal simplex method;

modifies a second optimal basis of the first objective of the new run to generate a second starting basis;

applies the primal feasibility changes;

solves the multi-objective hierarchical linear programming problem for the first objective using a dual simplex method and the second starting basis;

iteratively solves, according to the order specified by the objective hierarchy, each of the at least one additional objective of the multi-objective hierarchical linear programming problem for the new run, wherein decision variables for each subsequent objective are fixed at upper and lower bounds corresponding to objectives higher in the objective hierarchy that have already been solved; and

generates a new supply chain plan based, at least in part, on a solution of the new run.

16. The non-transitory computer-readable medium of claim 15 , wherein the software when executed further:

monitors for one or more supply chain changes that impact the base plan; and

when the one or more supply chain changes that impact the base plan are detected, calculates the one or more changes to the supply chain input data based, at least in part, on the one or more supply chain changes that impact the base plan.

17. The non-transitory computer-readable medium of claim 16 , wherein the software when executed further:

identifies current inventory levels of items at one or more supply chain entities; and

causes one or more items to be transported by a transportation network to fulfill the supply chain plan of the new run.

18. The non-transitory computer-readable medium of claim 17 , wherein the dual feasibility changes comprise one or more of an addition of a new variable, a change in an objective coefficient, and an addition of a constraint comprising a new variable.

19. The non-transitory computer-readable medium of claim 18 , wherein the primal feasibility changes comprise one or more of an addition of a new constraint, a change in a lower bounds, a change in an upper bounds, a change in a right hand side value, a change in a lower bounds to undo variable fixing based on the base run and to apply variable fixing based on previous objectives solved in the new run; and a change in an upper bounds to undo variable fixing based on the base run and to apply variable fixing based on previous objectives solved in the new run.

20. The non-transitory computer-readable medium of claim 19 , wherein modifying a need date of an existing demand comprises addition of a new demand, an adjustment of a priority of a demand by changing an objective coefficient of an existing demand to zero, and an adjustment of a need quantity of an existing demand to zero.