Method for characterizing a complex system

• US 7,117,131 B2
• Filed: 05/07/2001
• Issued: 10/03/2006
• Est. Priority Date: 05/26/1998
• Status: Expired due to Term

• Current Assignee: International Business Machines Corporation
• Original Assignee: Definiens AG (Astrazeneca PLC)
• Inventors: Binnig, Gerd
• Primary Examiner(s): Jones, Hugh
• Application Number:
  US09/850,928
• Publication Number:
  US 20020010710A1
• Time in Patent Office: 1,975 Days
• Field of Search: 703/2, 703/6
• US Class Current: 703/2
• CPC Class Codes: G06F 18/21 Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation

Abstract

Disclosed is a computer-implemented method for characterizing a complex system, wherein said complex system consists of a plurality of objects and wherein relations are present between said objects, said method comprising the steps of (a) forming a hierarchical structure having a plurality of hierarchical levels, wherein super ordinate objects of a respective higher hierarchical level of the hierarchical structure are combinations of objects of a respective lower hierarchical level of the hierarchical structure, wherein a characteristic and a boundary condition is assigned to each of said objects within the hierarchical structure; (b) forming groups of objects which belong to the same superordinate object of a respective higher hierarchical level; determining relations between objects in each respective group by taking into account the respective boundary conditions of the objects within the respective group; (d) determining at least one characteristic for each group based on the determined relations between objects within a respective group while taking into account the respective boundary conditions of the objects within the respective group in order to describe a superordinate object; (e) determining boundary conditions for each object of a respective group based on the characteristic of an superordinate object which represents the respective group; (f) modifying the objects in accordance with the determined characteristics and boundary conditions; and (g) iterative repeating steps (b) to (f) until a predetermined condition is met.

Claims

1. Computer-implemented method for characterizing a complex system, wherein said complex system consists of a plurality of objects and wherein relations are present between said objects, said method comprising:

(a) forming a hierarchical structure having a plurality of hierarchical levels, wherein an initial complex system includes input data to be evaluated and wherein superordinate objects of a respective higher hierarchical level of the hierarchical structure are combinations of objects of a respective lower hierarchical level of the hierarchical structure, wherein a characteristic and a boundary condition is assigned to each of said objects within the hierarchical structure;

(b) forming groups of objects which belong to the same superordinate object of a respective higher hierarchical level;

(c) determining relations between objects in each respective group by taking into account the respective boundary conditions of the objects within the respective group;

(d) determining at least one characteristic for each group based on the determined relations between objects within a respective group while taking into account the respective boundary conditions of the objects within the respective group in order to describe a superordinate object;

(e) determining boundary conditions for each object of a respective group based on the characteristic of an superordinate object which represents the respective group;

(f) modifying the objects in accordance with the determined characteristics and boundary conditions; and

(g) iterative repeating steps (b) to (f) until a predetermined condition is met to form a final complex system.

2. The method of claim 1, wherein the highest hierarchical level of the hierarchical structure is represented by a single object.

3. The method of claim 1, wherein in forming the hierarchical structure an algorithm is used.

4. The method of claim 3, wherein the algorithm is a fractal algorithm and is employed in a similar manner for all objects within the hierarchical structure.

5. The method of claim 1, wherein determining at least one characteristic for each group is based on a rising order of hierarchy level while proceeding on a fractal manner.

6. The method of claim 5, wherein determining at least one characteristic for each group includes calculating according to a fractal algorithm.

7. The method of claim 1, wherein determining boundary conditions for each object of a respective group includes determining in a descending order of hierarchy level while proceeding in a fractal manner.

8. The method of claim 7, wherein proceeding in a fractal manner includes splitting in accordance with fractal algorithms, wherein splitting in accordance with fractal algorithms includes at least one of classifying into sub-topics, classifying into more specific boundary conditions, splitting resources and distributing resources.

9. The method of claim 1, wherein determining at least one characteristic for each group is performed before determining boundary conditions for each object of a respective group.

10. The method of claim 1, wherein determining at least one characteristic of each group is performed after determining boundary conditions for each object of a respective group.

11. The method of claim 1, wherein determining relations between objects in a respective group, determining at least one characteristic for each group and calculating boundary conditions for each object of a respective group are each performed according to an algorithm, wherein each algorithm includes factors of neighboring objects, subordinate objects and boundary relations.

12. The method of claim 11, wherein each algorithm modifies respective objects based on at least one of the current status, the neighboring objects, the context and the subordinate objects.

13. The method of claim 1, further comprising the step of performing object recognition having the form of a self-optimizing simulation.

14. The method of claim 13, wherein performing object recognition includes employing a fractal-hierarchical library.

15. The method of claim 1, wherein each characteristic is classified into one of an internal characteristic and an external characteristic and wherein each boundary condition is classified into one of a internal boundary condition and an external boundary condition.

16. The method of claim 15, wherein the internal characteristics are determined in an internal context from internal relations of the subordinate objects with their characteristics.

17. The method of claim 16, wherein the internal context constitutes the boundary condition for the internal relations of the subordinate object and is determined from an external context of the objects and their external characteristics.

18. The method of claim 17, wherein the external characteristics are calculated by bunching in the external context.

19. The method of claim 18, wherein the external context constitutes the boundary condition for determining the external characteristic and is calculated from Internal relations of the superordinate object and its internal context.

20. The method of claim 1, wherein the complex system is a satellite picture.