Production optimizer for supply chain management

• US 8,112,300 B2
• Filed: 04/12/2006
• Issued: 02/07/2012
• Est. Priority Date: 04/22/2005
• Status: Expired due to Fees

• Current Assignee: Air Liquide Large Industries U.S. LP (L'Air Liquide SA)
• Original Assignee: Air Liquide Large Industries U.S. LP (L'Air Liquide SA)
• Inventors: Harper, Charles N.
• Primary Examiner(s): Choi, Peter
• Application Number:
  US11/279,557
• Publication Number:
  US 20060241986A1
• Time in Patent Office: 2,127 Days
• Field of Search: 705/7.12, 705/7.26
• US Class Current: 705/7.26
• CPC Class Codes: G06N 3/126 Evolutionary algorithms, e.g. genetic algorithms or genetic programming
• CPC Class Codes: G06Q 10/04 Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem" market predictions or forecasting for commercial activities G06Q30/0202
• CPC Class Codes: G06Q 10/063 Operations research, analysis or management
• CPC Class Codes: G06Q 10/0631 Resource planning, allocation, distributing or scheduling for enterprises or organisations
• CPC Class Codes: G06Q 10/06316 Sequencing of tasks or work
• CPC Class Codes: G06Q 10/0639 Performance analysis of employees; Performance analysis of enterprise or organisation operations
• CPC Class Codes: Y02P 90/80 Management or planning

Abstract

One embodiment of the invention provides a method for optimizing a supply chain management (SCM) problem. A genetic algorithm optimization technique may be used to generate a production solution for the production side of an SCM problem and an ant colony optimization technique may be used to generate a solution for the distribution side of the SCM problem. Together, the genetic algorithm optimization technique and the ant colony optimization technique operate to quickly identify high-quality solutions to an SCM problem.

Claims

1. A computer-implemented method of optimizing a supply chain, comprising:

a) generating, by operation of one or more processors executing a first application, an optimized production solution for the supply chain by using a genetic algorithm optimization technique to optimize a population of seed solutions, wherein the optimized production solution specifies at least, for each day of an optimization period;

(i) a production level for each of one or more commodity products produced at each of one or more production plants and (ii) a production split between the commodity products for each of the one or more production plants; and

b) generating, by operation of one or more processors executing a second application, an optimized distribution solution for distributing the commodity products produced according to the optimized production solution using an ant colony optimization technique, wherein generating the optimized distribution solution for the optimized production solution using the ant colony optimization technique, comprises;

(i) initializing a plant weight vector for each of the production plants in the supply chain, wherein components of the plant weight vector indicate preferences for a delivery from the plant represented by the plant weight vector to an installation represented by a vector component,

(ii) initializing an installation vector for each delivery location in the supply chain, and wherein components of installation vector indicate a preference for a subsequent delivery from the installation represented by the installation vector to an installation represented by a vector component, and

(iii) generating the distribution solution based on a set of weights specified by components of the plant weight vectors and installation vectors, wherein the distribution solution describes one or more delivery routes from a given one of the production plants to at least one delivery location.

2. The computer-implemented method of claim 1, wherein generating an optimized production solution for the supply chain, comprises:

a) reading input data, wherein the input data describes a supply chain problem to be optimized;

b) generating the population of seed solutions, wherein each solution in the population specifies a feasible solution for a supply chain problem to be optimized;

c) evaluating each solution in the population and sorting the solutions in the population according to the evaluation; and

d) performing the genetic algorithm optimization technique on the population until a termination condition specified by the input data is satisfied.

3. The computer-implemented method of claim 2, wherein performing the genetic algorithm optimization technique comprises:

a) modifying at least one aspect of at least one of the solutions in the population;

b) evaluating the performance of the modified solution;

c) adding the modified solution to the population; and

d) removing a solution from the population with the poorest evaluation, relative to other solutions in the population.

4. The computer-implemented method of claim 3, wherein modifying at least one aspect of at least one of the solution in the population, comprises:

a) selecting a solution modification technique;

b) selecting the at least one solution from the population to modify; and

c) modifying the at least one solution, according to the selected solution modification technique.

5. The computer-implemented method of claim 3, wherein the solution modification technique randomly modifies at least one aspect of the at least one solution.

6. The computer-implemented method of claim 3, wherein the solution modification technique modifies the at least one solution according to a heuristic rule.

7. The computer-implemented method of claim 3, wherein the solution modification technique cross-breeds two or more solutions to generate the modified solution.

8. The computer-implemented method of claim 3, wherein modifying at least one aspect of at least one of the solutions in the population comprises probabilistically selecting one or more solutions, wherein a solution with a superior evaluation is more likely to be selected for modification than a solution with an inferior evaluation.

9. The computer-implemented method of claim 1, wherein the population of seed solutions includes a first group of solutions and a second group of solutions, wherein the first group of solutions are generated according to heuristic rules and wherein the second group of solutions are randomly generated.

10. A computer-readable storage medium containing a program which, when executed, performs operations for optimizing a supply chain, comprising:

a) generating an optimized production solution for the supply chain by using a genetic algorithm optimization technique to optimize a population of seed solutions, wherein the optimized production solution specifies at least, for each day of an optimization period;

(i) a production level for each of one or more commodity products produced at each of one or more production facilities and (ii) a production split between the commodity products for each of the one or more production faculties; and

b) generating an optimized distribution solution for distributing the commodity products produced according to the optimized production solution using an ant colony optimization technique, wherein generating the optimized distribution solution for the optimized production solution using the ant colony optimization technique, comprises;

(i) initializing a plant weight vector for each of the production plants in the supply chain, wherein components of the plant weight vector indicate preferences for a delivery from the plant represented by the plant weight vector to an installation represented by a vector component,

(ii) initializing an installation vector for each delivery location in the supply chain, and wherein components of installation vector indicate a preference for a subsequent delivery from the installation represented by the installation vector to an installation represented by a vector component, and

(iii) generating the distribution solution based on a set of weights specified by components of the plant weight vectors and installation vectors, wherein the distribution solution describes one or more delivery routes from a given one of the production plants to at least one delivery location.

11. The computer-readable storage medium of claim 10, wherein the operation for generating an optimized production solution for the supply chain comprise:

a) reading input data, wherein the input data describes a supply chain problem to be optimized;

b) generating the population of seed solutions, wherein each solution in the population specifies a feasible solution for a supply chain problem to be optimized;

c) evaluating each solution in the population and sorting the solutions in the population according to the evaluation; and

d) performing the genetic algorithm optimization technique on the population until a termination condition specified by the input data is satisfied.

12. The computer-readable storage medium of claim 11, wherein the operations for performing the genetic algorithm optimization technique, comprise:

a) modifying at least one aspect of at least one of the solutions in the population;

b) evaluating the performance of the modified solution;

c) adding the modified solution to the population; and

d) removing a solution from the population with the poorest evaluation, relative to other solutions in the population.

13. The computer-readable storage medium of claim 12, wherein the operations for modifying at least one aspect of at least one of the solution in the population, comprise:

a) selecting a solution modification technique;

b) selecting the at least one solution from the population to modify; and

c) modifying the at least one solution, according to the selected solution modification technique.

14. The computer-readable storage medium of claim 12, wherein the solution modification technique randomly modifies at least one aspect of the at least one solution.

15. The computer-readable storage medium of claim 12, wherein the solution modification technique modifies the at least one solution according to a heuristic rule.

16. The computer-readable storage medium of claim 12, wherein the solution modification technique cross-breeds two or more solutions to generate the modified solution.

17. The computer-readable storage medium of claim 12, wherein modifying at least one aspect of at least one of the solutions in the population comprises probabilistically selecting one or more solutions, wherein a solution with a superior evaluation is more likely to be selected for modification than a solution with an inferior evaluation.

18. The computer-readable storage medium of claim 10, wherein the population of seed solutions includes a first group of solutions and a second group of solutions, wherein the first group of solutions are generated according to heuristic rules and wherein the second group of solutions are randomly generated.

19. A computing device used to optimize operations of a supply chain, comprising:

a) a processor;

b) a production optimizer which, when executed by the processor, is configured to generate an optimized production solution for the supply chain by using a genetic algorithm optimization technique to optimize a population of seed solutions, wherein the optimized production solution specifies at least, for each day of an optimization period;

(i) a production level for each of one or more commodity products produced at each of one or more production facilities and (ii) a production split between the commodity products for each of the one or more production faculties; and

c) a distribution optimizer which, when executed by the processor, is configured to generate an optimized distribution solution for distributing the commodity products produced according to the optimized production solution using an ant colony optimization technique, wherein generating the optimized distribution solution for the optimized production solution using the ant colony optimization technique, comprises;

(i) initializing a plant weight vector for each of the production plants in the supply chain, wherein components of the plant weight vector indicate preferences for a delivery from the plant represented by the plant weight vector to an installation represented by a vector component,

(ii) initializing an installation vector for each delivery location in the supply chain, and wherein components of installation vector indicate a preference for a subsequent delivery from the installation represented by the installation vector to an installation represented by a vector component, and

(iii) generating the distribution solution based on a set of weights specified by components of the plant weight vectors and installation vectors, wherein the distribution solution describes one or more delivery routes from a given one of the production plants to at least one delivery location.

20. The computing device of claim 19, wherein the production optimizer is configured to generate an optimized production solution for the supply chain by performing the steps of:

a) reading input data, wherein the input data describes a supply chain problem to be optimized;

b) generating the population of seed solutions, wherein each solution in the population specifies a feasible solution for a supply chain problem to be optimized;

c) evaluating each solution in the population and sorting the solutions in the population according to the evaluation; and

d) performing the genetic algorithm optimization technique on the population until a termination condition specified by the input data is satisfied.

21. The computing device of claim 20, wherein the step of performing the genetic algorithm optimization technique comprises:

a) modifying at least one aspect of at least one of the solutions in the population;

b) evaluating the performance of the modified solution;

c) adding the modified solution to the population; and

d) removing a solution from the population with the poorest evaluation.

22. The computing device of claim 21, wherein the step of modifying at least one aspect of at least one of the solution in the population, comprises:

a) selecting a solution modification technique;

b) selecting the at least one solution from the population to modify; and

c) modifying the at least one solution, according to the selected solution modification technique.

23. The computing device of claim 21, wherein the solution modification technique randomly modifies at least one aspect of the at least one solution.

24. The computing device of claim 21, wherein the solution modification technique modifies the at least one solution according to a heuristic rule.

25. The computing device of claim 21, wherein the solution modification technique cross-breeds two or more solutions to generate the modified solution.

26. The computing device of claim 21, wherein the step of modifying at least one aspect of at least one of the solutions in the population comprises probabilistically selecting one or more solutions, wherein a solution with a superior evaluation is more likely to be selected for modification than a solution with an inferior evaluation.

27. The computing device of claim 19, wherein the population of seed solutions includes a first group of solutions and a second group of solutions, wherein the first group of solutions are generated according to heuristic rules and wherein the second group of solutions are randomly generated.