Self-optimizing method and machine

• US 6,625,500 B1
• Filed: 03/16/1999
• Issued: 09/23/2003
• Est. Priority Date: 03/16/1999
• Status: Expired due to Fees

• Current Assignee: LI Family Holding Limited
• Original Assignee: LI Family Holding Limited
• Inventors: Li, Chou H.
• Primary Examiner(s): Gordon, Paul P.
• Application Number:
  US09/268,296
• Time in Patent Office: 1,652 Days
• Field of Search: 700/28, 700/29, 700/34, 700/44, 700/47, 700/48, 700/49, 700/73, 706/46, 706/47, 706/48, 706/62, 706/13, 706/25, 706/61, 706/45
• US Class Current: 700/29
• CPC Class Codes: G05B 13/021 in which a variable is automatically adjusted to optimise the performance
• 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

Abstract

A method for computer-generating interaction-specific knowledge base for rapidly improving or optimizing a performance of an object comprises performing, according to computer-designed test matrices, at least several automatic experimental cycles on selected control variables. In at least one of the automatic experimental cycles after the first the computer plans a new test matrix designed to minimize or remove at least one expected two-variable interaction from a main effect of a designated control variable. A machine operating according to the method is also available.

Claims

1. A method for self-optimizing a performance of an object comprising:

performing a plurality of computerized automatic experimental cycles on a plurality of selected control variables;

each experimental cycle comprising computer-planning a designed test matrix, executing the experimental cycle according to the designed test matrix to collect test results, and computer-analyzing, in real time, the test results to determine at least one way to combine optimally the selected control variables for optimizing or improving the performance of the object; and

in at least one of the experimental cycles after the first cycle computer-planning a new designed test matrix to minimize or remove at least one expected two-variable interaction effect from main effect of at least one designated control variable;

a main effect being a computer-analyzed individual effect of a single control variable on the performance of the object, while an interaction effect being a computer-analyzed combined effect of a plurality of control variables on the performance.

2. The method of claim 1 wherein the computer-analyzing step comprises performing a function selected from the group consisting of determining main effects of the control variables and finding an optimal combination of these control variables.

3. The method of claim 1 including computer-planning the new designed test matrix to isolate the at least one expected two-variable interaction effect from the main effect of the at least one designated control variable found to be important in a previous experimental cycle.

4. The method of claim 1 wherein the object is selected from the group consisting of a living object, a machine, a system, an industry, an office, a factory, and a business;

and

including performing, with real-time feed-back control, the plurality of the experimental cycles on the object.

5. The method of claim 1 wherein the designed test matrix is computer-planned according to a less than one-sixteenth fractionally replicated, factorial design;

and

including sensing the performance of the object with a computer-controlled sensor; and

connecting the object to both the sensor and a computer.

6. The method of claim 1 including electrooptically measuring the performance of the object with a measuring device;

and

including electrooptically connecting the object to both a computer and the measuring device.

7. The method of claim 1 including computer-planning the new designed test matrix to be similar to the designed test matrix but with a fewer number of control variables tested therein and designed to insure that the main effect of the at least one designated control variable is not contaminated by the at least one expected two-variable interaction effect.

8. The method of claim 1 including providing an interaction-variable identification knowledge base for use by a computer to computer-plan the new designed test matrix to specifically reserve a variable space for an expected two-variable interaction effect to appear therein so as not to contaminate, or appear as at least a part of, the main effect of the at least one designated control variable.

9. The method of claim 1 including providing computer-usable expert rules for use as an interaction-variable identification knowledge base to insure that the at least one expected two-variable interaction effect will not contaminate the main effect of the at least one designated control variable.

10. The method of claim 1 wherein the computer-planned, new designed test matrix consists essentially of the same designed test matrix but with a sequence of the control variables arranged differently from that in a previous experimental cycle, such that the main effect of the at least one designated control variable is isolated from the at least one expected two-variable interaction effect.

11. The method of claim 10 wherein a main effect of the designated control variable can be contaminated by the at least one expected two-variable interaction effect between the designated control variable and another control variable;

and

including computer-planning the new designed test matrix to separate the main effect of the at least one designated control variable from the at least one expected two-variable interaction effect.

12. The method of claim 1 wherein the main effects of two designated control variables can be contaminated by the at least one expected two-variable interaction effect, and

including computer-planning the new designed test matrix to separate the computer-analyzed main effects of the two designated control variables from the at least one expected two-variable interaction effect whereby the computer-analyzed main effects represent substantially true main effects of the two designated control variables substantially uncontaminated by the at least one expected two-variable interaction effect.

13. The method of claim 12 including using at least the substantially true main effects of the two designated control variables to develop a steepest ascent path on a multi-dimensional performance response surface for the performance to optimally climb the steepest ascent path thereby optimizing or improving the performance of the object.

14. The method as in claim 1 wherein the experimental cycles are performed on an object of interest and in actual operation, with same materials, parts, procedures, equipment, and environment as in said actual operation;

and

including quickly collecting and computer-analyzing the test results to optimize or improve the performance of the object during the actual operation thereof, whereby various ever-present and often critical errors of sampling, extrapolation, time delay, and scaling-up are minimized or eliminated.

15. The method of claim 1 wherein the performing step for each of the experimental cycles comprises computing for no more than a time selected from the group consisting of one minute, one second, one millisecond, and one microsecond.

16. The method of claim 1 wherein the performing step comprises generating new knowledge bases in a specific domain of interest;

and

including coding the generated new knowledge bases into a computer-usable form.

17. The method of claim 16 wherein the coding step comprises computer-coding, in real time, the generated new knowledge bases into a computer-usable form;

and

including immediately applying the computer-coded new knowledge bases onto the object for improving or optimizing the performance thereof.

57. A method for self-optimizing or self-improving performance of an object comprising providing a memory device which contains the knowledge base computer-generated by the method of claim 1, 34, 47, or 54.

18. A machine for self-optimizing a performance of an object comprising:

means for performing a plurality of computerized automatic experimental cycles on a plurality of selected control variables;

the performing means in each of the plurality of the experimental cycles comprising means for computer-planning a designed test matrix, means for executing the experimental cycle according to the designed test matrix to collect test results, and means for computer-analyzing, in real time, the test results to determine a way to combine optimally the selected control variables for achieving an optimized or improved performance of the object; and

in at least one of the experimental cycles after the first cycle the computer-planning means comprising means for computer-planning a new designed test matrix to minimize or remove at least one expected two-variable interaction effect from a main effect of at least one designated control variable;

a main effect being a computer-analyzed individual effect of a single control variable on the performance of the object, while an interaction effect being a computer-analyzed effect of a plurality of control variables working in combination on the performance.

19. The machine of claim 18 in which the computer-analyzing means comprises means selected from the group consisting of means for determining main effects of the selected control variables and means for finding an optimal combination of these control variables.

20. The machine of claim 18 in which the computer-planning means comprises means for computer-planning the new designed test matrix to isolate an expected two-variable interaction effect from the main effect of the at least one designated control variable which is found to be important in a previous experimental cycle.

21. The machine of claim 18 in which the object is selected from the group consisting of a living object, a machine, a system, an industry, an office, a factory, and a business;

and

the performing means comprises means for performing, with real-time feed-back control, the plurality of the experimental cycles on the object.

22. The machine of claim 18 including means for sensing the performance of the object by a computer-controlled sensor;

and

means for connecting the object to both the sensor and a computer.

23. The machine of claim 18 including a device for measuring the performance electrooptically;

and

means for electrooptically connecting the object to both the measuring device and a computer.

24. The machine of claim 18 in which the computer-planned, new designed test matrix is the same as the designed test matrix but with a fewer number of control variables tested therewith and designed to insure that the main effect of the at least one designated control variable is not contaminated by the at least one expected two-variable interaction effect.

25. The machine of claim 18 including means for providing an interaction-variable identification knowledge base for the computer-planning means to computer-plan the new designed test matrix in which a space is specifically reserved for an expected two-variable interaction effect to appear therein to avoid its contamination of the at least one designated control variable.

26. The machine of claim 18 including means for providing computer-usable expert rules for use as an interaction-variable identification knowledge base to insure that the at least one expected two-variable interaction effect will not contaminate the main effect of the at least one designated control variable.

27. The machine of claim 18 in which the computer-planned, new designed test matrix consists essentially of the same designed test matrix but with a sequence of the control variables arranged differently from that in a previous experimental cycle, such that the main effect of the at least one designated control variable is isolated from an expected two-variable interaction effect.

28. The machine of claim 18 in which the main effect of the at least one designated control variable can be contaminated by the at least one expected two-variable interaction effect between the at least one designated control variable and another control variable;

and

the computer-planning means comprises means for computer-planning the new designed test matrix to separate the main effect of the at least one designated control variable from the one expected two-variable interaction.

29. The machine of claim 18 in which main effects of at least two designated control variables are possibly contaminated by one expected two-variable interaction effect, and

including means for supplying an interaction-variable identification knowledge base to a computer for computer-planning the new designed test matrix for separating the computer-analyzed main effects of the at least two designated control variables from the expected two-variable interaction effect whereby the computer-analyzed main effects represent substantially true main effects of the at least two designated control variables which are substantially uncontaminated by the one expected two-variable interaction effect.

30. The machine of claim 29 including means for using the substantially true main effects of the at least two designated control variables to develop a steepest ascent path on a multi-dimensional performance response surface for the performance to optimally climb thereon thereby optimizing or improving the performance of the object.

31. The machine of claim 18 in which the performing means for each of the experimental cycles comprises means for computing for no more than a time selected from the group consisting of one minute, one second, one millisecond, and one microsecond.

32. The machine of claim 18 in which the performing means comprises means for generating new knowledge bases in a specific domain of interest;

and

including means for coding the generated new knowledge bases in a computer-usable form.

33. The machine of claim 32 wherein the coding means comprises means for computer-coding, in real time, the generated new knowledge bases into a computer-usable form for immediate application onto the object for improving or optimizing the performance thereof.

34. A method of self-optimizing or self-improving a performance of an object comprising:

creating a computer-generated knowledge base for use on the object including;

performing at least one automatic computerized, fractional replicated experiment involving a plurality of selected control variables on the object to obtain test results; and

computer-analyzing the test results in real time to generate the computer-generated knowledge base including information on at least two designated control variables for use in optimizing or improving said performance of the object;

said information being more than mere main linear effect of at least one of said designated control variables on the performance, but including at least one two-variable interaction effect between said at least one designated control variables and another control variable;

a main effect being a machine-analyzed individual effect of a single control variable on said performance of the object, and a two-variable interaction effect being a machine-analyzed combined effect of the two selected control variables on said specified performance of the object.

35. The method of claim 34 wherein said computer-generated knowledge base is interaction-specific.

36. The method of claim 34 including automatically computer-generating, in real time, the knowledge base using said at least one automatic computerized experiment involving a minimum number of control variables;

the minimum number being less than a number selected from the group consisting of 8, 16, 32, 64, 128, 256, 512, and 1,024.

37. The method of claim 34 including instantly computer-coding said computer-generated knowledge base for immediate, direct use on the object.

38. The method of claim 34 wherein said computer-generated knowledge base is in a specific domain of technology and is at least three times more variables and 10 times less costly than human-generated knowledge base in the same specific domain of technology.

39. The method of claim 34 including telecommunicating said computer-generated knowledge base to a computer.

40. The method of claim 34 including telecommunicating said computer-generated knowledge base to a computer for use as finalized optimizing process or product without any additional new automatic computerized experiments.

41. Apparatus for self-optimizing or self-improving a performance of an object comprising;

a computer usable medium having computer readable program code means embodied therein for self-optimizing or self-improving said performance of the object;

computer readable program code means for performing a plurality of computerized automatic experimental cycles on said object relative to a plurality of selected control variables;

means for computer-planning at least one designed experiment for each of the plurality of computerized automatic experimental cycles, said at least one designed experiment being computer-designed to insure that a main effect of at least one control variable is substantially uncontaminated by at least one interaction effect between the control variables;

means for computer-executing the design experiment to acquire experimental results; and

means for computer-analyzing in real time the experimental results to generate an interaction-specific knowledge base for use by the object;

a main effect being a machine-analyzed individual effect of a single control variable on said performance of the object, and a two-variable interaction effect being a machine-analyzed combined effect of the two selected control variables on said specified performance of the object.

42. The apparatus of claim 41 wherein at least one of said experimental cycles comprises an experiment designed specifically to isolate a most important main effect from at least one two-variable interaction effect.

43. The apparatus of claim 41 wherein said means for computer planning is capable of planning an experimental cycle after the first cycle and without human control comprising a new computer-designed experiment to estimate main effects of at least selected two control variables in addition to an expected two-variable interaction effect of these two control variables.

44. The apparatus of claim 43 wherein said means for computer-planning is capable of planning said new computer-designed experiment in a new experimental region not centered around a point representing the optimal variable combination in the immediately previous experimental cycle in order to broaden the experiment region.

45. The apparatus of claim 41 wherein said means for computer planning is capable of computer-planning the at least one designed experiment for determining a presence of at least one interaction effect of two designated control variables.

46. The apparatus of claim 41 wherein said means for computer-planning is capable of computer-planning, for at least one of the experimental cycles after the first cycle, a new computer-designed experiment to determine a presence of at least one suspected two-variable interaction effect of two selected control variables in addition to main effects of these two selected control variables to thereby generate an interaction-specific knowledge base in a specific domain of technology.

47. A program storage device containing computer readable program code embodied therein for self-optimizing or self-improving a performance of an object comprising:

a memory device readable by a computer, tangibly embodying a program of instructions executable by the computer to perform method steps for at least one automatic computerized experimental cycle each involving a plurality of selected control variables capable of use for self-optimizing or self-improving a performance of an object, said method steps comprising;

computer-planning a designed experiment for at least one automatic computerized experimental cycle;

said designed experiment being computer-designed to determine a presence of at least one interaction effect of two selected control variables;

a main effect being a computer-analyzed individual effect of a single control variable on the performance of the object, while an interaction effect being a computer-analyzed combined effect of a plurality of control variables working in combination on the performance.

48. The program storage device of claim 47 wherein said method steps further include:

using an interaction-variable identification knowledge base; and

causing the computer to automatically plan the designed experiments specifically to isolate a most important main effect from at least one likely two variable interaction effect.

49. The program storage device of claim 47 wherein said method steps further include:

computer-planning the designed experiment to determine presence or absence of at least one interaction effect of two of designated control variables.

50. A method for self-optimizing or self-improving a performance of an object comprising:

creating a memory for storing a data structure for access by an application program being executed on a data processing system;

said data structure being related to a designated field of technology, being a real-time computer-generated knowledge base and containing at least a determination of a presence of an expected interaction effect of at least two designated control variables in addition to main effects of the two designated control variables;

a main effect being a computer-analyzed individual effect of a single control variable on the performance of the object, and a two-variable interaction effect being a computer-analyzed combined effect of two of the selected control variables on the performance of the object.

51. A method for self-optimizing or self-improving a performance of an object comprising:

creating a processing system for executing an application program and containing a computer-generated knowledge base used by said application program;

said processing system comprising;

a computer device for processing said application program; and

a memory component for holding said computer-generated knowledge base for access by said application program, said knowledge base being composed of information resident for use by said application program and including;

at least a determination of presence of an expected interaction effect of at least two designated control variables in addition to main effects of these two designated control variables;

a main effect being a computer-analyzed individual effect of a single control variable on the performance of the object, and a two-variable interaction effect being a computer-analyzed combined effect of two of the selected control variables on the performance of the object.

52. A method for self-optimizing or self-improving a performance of an object comprising:

creating a computer-readable medium containing instructions for self-optimizing or self-improving a performance of an object relative to a plurality of control variables, by;

computer-planning at least one computerized automatic experimental cycle containing a computer-designed test matrix;

said computer-planning step including computer-planning for determining a presence of at least one interaction effect of at least two designated variables in addition to substantially uncontaminated main effects of these at least two designated control variables; and

computer-executing the computerized automatic experimental cycle according to the at least one computer-designed test matrix to collect test results;

computer-analyzing, in real time, the test results to determine a way to determine substantially uncontaminated main effects, of said at least two designated control variables;

computer-coding the computer-analyzed test results for storage in a selected computer-readable program storage means; and

when it is determined that the performance of the object is substantially optimized, keeping the object in the thus-optimized or improved condition and recording into the selected computer readable program memory means the optimum combination of the control variables;

a main effect being a computer-analyzed individual effect of a single control variable on the performance of the object, while an interaction effect being a computer-analyzed combined effect of a plurality of control variables working in combination on the performance.

53. A method for self-optimizing or self-improving a performance of an object comprising:

creating a computer-readable data transmission medium containing a real-time computer-generated knowledge base including an optimum combination of selected control variables in a specific field of technology, and a determination of presence of an expected interaction effect between two designated control variables in addition to main effects of these two designated control variables;

a main effect being a computer-analyzed individual effect of a single control variable on the performance of the object, and a two-variable interaction effect being a computer-analyzed combined effect of two of the selected control variables on the performance of the object.

54. A method for self-optimizing or self-improving a performance of an object wherein said method comprises:

creating a computer system for communicating with an automation system;

real-time automatic computerized experimenting on a plurality of selected control variables in at least one computer-designed statistical experiment involving at least 7 or 15 control variables with a test number no more than one-sixteenth of a test number normally required for complete factorial experimentation;

collecting the experimental results;

computer-analyzing the experimental results to generate a new interaction-specific knowledge base including a substantially pure main effect of a most important control variable and at least one two-variable interaction of the most important variable;

computer-coding in real time the new knowledge base into a machine-usable form; and

in response to the automation system, telecommunicating the computer-coded, new interaction-specific knowledge base to the automation system for its immediate implementation to achieve a substantially optimal condition on the automation system;

a main effect being a computer-analyzed individual effect of a single control variable on the performance of the object, and a two-variable interaction effect being a computer-analyzed combined effect of two of the selected control variables on the performance of the object.

55. The method of claim 54 wherein the new interaction specific knowledge base is real-time computer-generated according to the at least one designed experiment and computer-analyzed and coded ready for telecommunicating for immediate application to instantly achieve optimal or improved results on the automation system without additional automatic computerized experimentation on the automation.

56. The method of claim 54 further including providing a memory for storing said new interaction-specific knowledge base for access by an application program being executed on a data processing system;

said interaction-specific knowledge base including information resident in a database used by said application.

58. A method for self-optimizing or self-improving a performance of an object including generating a knowledge base for access by an application program executable by a data processor, said method comprising the method steps, executed by said data processor, of:

a) computer-planning a designed experiment for at least one computerized automatic experimental cycles on a plurality of selected control variables to obtain main effects of all the control variables including two designated control variables; and

one interaction effect of these two designated variables;

b) computer-executing the at least one designed experiment to obtain test results;

c) computer-analyzing, in real time, the test results to determine at least a way to combine the selected control variables for improving or optimizing said performance of the object;

said computer-analyzing step determining a presence of at least one suspected two-variable interaction effect in addition main effects of at least these two control variables of the suspected two-variable interaction effect;

a main effect being a computer-analyzed individual effect of a single control variable on the performance of the object, and a two-variable interaction effect being a computer-analyzed combined effect of two of the selected control variables on the performance of the object.

59. A method for self-optimizing or self improving a performance of a computer optimizable object for access by an application program being executable by a data processor, said method comprising the method steps, executed by said data processor, of:

a) planning with a computer a designed experiment for each of a plurality of computerized automatic experimental cycles on a plurality of selected control variables;

said designed experiment being computer-planned for at at least two different experimental regions;

b) sequentially computer-executing the experimental cycles according to the respective computer-planned designed experiments to obtain test results;

c) computer-analyzing, in real time, the test results to computer-generate an interaction specific knowledge base relating to presence or absence of the at least one combined effect of at least two selected control variables and also to a way for optimizing or improving the performance of said object; and

d) computer-coding the computer-generated knowledge base for storage in a selected computer memory device; and

e) computer-telecommunicating, in real time, said computer-generated interaction specific knowledge base to a computing means for use not in starting its own new computerized automatic experimentation, but as an embedded expert system that performs at least a plurality of the following;

aa) understanding the computer-generated and computer-coded knowledge base;

bb) logically comparing different parts of the knowledge base;

cc) selecting relevant compared parts of the knowledge base;

dd) synthesizing new method steps, based on logical decisions, for further improving the performance of the computer optimizable object; and

ee) providing a new synthesized recipe for achieving an improved performance of the computer-optimizable object with the use of said computing means.

60. A method of self-optimizing or self-improving a performance of an object comprising:

computer-planning to computer-design in real time at least one experiment on a plurality of selected control variables;

insuring that main effect of at least one designated control variable is substantially uncontaminated by at least one interaction effect between the selected control variables;

executing the designed experiment to acquire experimental performance results; and

computer-analyzing in real time the experimental results to generate an interaction-specific knowledge base for use by the object;

a main effect being a computer-analyzed individual effect of a single control variable on the performance of the object, while an interaction effect being a computer-analyzed combined effect of a plurality of control variables working in combination on the performance.

61. The method as in claim 60 including computer generating a interaction-variable identification knowledge base:

computer-coding the computer-generated interaction-specific knowledge base in a computer-usable form; and

applying the computer-generated and computer-coded, interaction-specific knowledge base onto the object to improve or optimize a performance of the object.

62. A method as in claim 60 including coupling the computer-planning, executing, and computer-analyzing steps to achieve real-time closed-loop automation.

63. A method as in claim 60 wherein the computer-planning step comprises computer-planning the designed experiment to insure that at least one designated control variable is not significantly contaminated by an expected two-variable interaction effect.

64. A method as in claim 60 wherein the computer-analyzing step comprises computer-generating an interaction-specific knowledge base.

65. A method as in claim 64 wherein the computer-analyzing step comprises providing computer-coded rules containing probabilistic statements derived from an experimental error obtained in the computer-analyzing step;

and

the computer-coding step comprises computer-coding, in real time, at least a part of the computer-generated interaction-specific knowledge base into an inference engine.

66. A method as in claim 64 wherein the computer-coding step comprises computer-coding the computer-generated interaction-specific knowledge base in real-time and without any human involvement in a form selected from the group consisting of guidance, reviews, analyses, interactions, and supervision.

67. A method as in claim 60 for optimizing a specific growth performance of a selected growing object;

and

wherein the control variables are growth-controlling variables selected from the group consisting of equipment, materials, parts, procedures, and environment.

68. A method as in claim 60 including:

instantly feeding back information on status of improvement or optimization of the performance of the object;

reperforming at least one subsequent computer-planned designed experiment on the object, with emphasis on isolating at least one specific two-variable interaction effect;

computer-reanalyzing in real time the new experimental results to generate a more recent computer-codeable and interaction-specific knowledge base;

computer-recoding in real time the computer-generated more recent computer-codeable, interaction-specific knowledge base in computer-usable form; and

reapplying the computer-recoded, more recent interaction-specific knowledge base for improving or optimizing the performance of the object.

69. A method as in claim 68 wherein:

the computer-analyzing or computer-reanalyzing step comprises determining a best combination of the control variables for achieving optimization or most improvement at the time of the computer-planned designed experiment; and

the reperforming step after the first automatic, computer-planned designed experiment comprises reperforming the at least one subsequent computer-planned designed experiment around, but not at, the best control variables combination determined in an immediately previous computer-planned designed experiment.

70. A method as in claim 69 wherein the computer-analyzing and computer-reanalyzing steps comprise determining main effect, at different times of the computer-planned designed experiments, of at least one of the control variables to achieve a time effect thereof.

71. A method as in claim 60 wherein the computer-planning step comprises computer-planning a plurality of close-loop designed experiments on the plurality of control variables;

each of the computer-planned designed experiments being performed on an object whose performance is to be optimized without relying on invalid scaling-up law or extrapolations form sampled test results obtained on other similar objects.

72. A method as in claim 60 including the additional step of sensing the performance of the object;

and

compensating, in the sensed performance, at least a portion of sensing errors due to a cause selected from the group consisting of zero-point miscorrection, proportionality constant deviation, and use of wrong sensors.

73. A method as in claim 60 including minimizing or eliminating in the computer-planned designed experiment at least one error selected from the group consisting of those due to sampling, deficient knowledge bases, misused or miscalibrated sensors, actuator backlash, and drifting equipment.

74. A method as in claim 60 including computer-sensing by a computer during the executing step at least part of the experimental results on a performance of the object.

75. A method as in claim 60 for improving or optimizing a performance of object which is selected from the group consisting of human, animal, plants, and virus, and bacteria;

and

wherein the performance relates to a physical, mental, medical, psychological, or physiological condition or activity of the object; and

including selecting the control variables from the group consisting of noise or music type, intensity, and tempo of the sound or noise;

smell;

lighting;

smell;

lighting;

temperature;

humidity;

pressure;

visual display and recorded tapes or disks; and

other factors that stimulate the object physically, mentally, or psychologically.

76. A method as in claim 60 for computer-generating the interaction-specific knowledge base for an improved or optimized effect on a physical, mental, medical, psychological, or physiological condition or activity of a living object;

and

wherein the control variables are selected from the group consisting of type, tempo, frequency, and intensity of sound, noise, or vibration;

smell;

lighting;

temperature;

humidity;

pressure on the object;

video or audio signal; and

other factors that stimulate the living object physically, mentally, medically, psychologically, or physiologically.

77. A method as in claim 60 wherein the computer-planning step comprises computer-planning a plurality of close-loop computer-planned designed experiments on the plurality of the control variables;

each of the computer-planned designed experiments being performed on the object to be improved or optimized without relying on invalid scaling-up laws or extrapolations from sampled test results obtained on other similar objects.

78. A method of self-improving or self-optimizing a performance of an object including:

performing at least one automatic computerized experiment on an object whose performance is to be improved or optimized, comprising;

computer-planning at least one designed experiment on a plurality of selected control variables;

the designed experiment being of a fractionally factorial type such that main effect of at least one designated control variable is not appreciably contaminated by an expected interaction effect of selected two of the control variables;

executing the at least one automatic computerized experiment to acquire experimental results; and

computer-analyzing in real time the experimental results to computer generate new interaction-specific knowledge bases in a computer-usable form;

a main effect being a computer-analyzed individual effect of a single control variable on the performance of the object, while an interaction effect being a computer-analyzed combined effect of a plurality of control variables on the performance.

79. A method as in claim 78 including coupling the computer-planning, executing, and computer-analyzing steps to achieve real-time closed loop automation;

and including automatically coding at least some of the computer-generated new knowledge bases into computer-usable form.

80. A method as in claim 78 in which the new computer generated interaction-specific knowledge bases are in a specific domain of technology where human knowledge bases are costly, unreliable, and difficult to be translated even by humans for uses by computers, and wherein human coding of the knowledge base is expensive if at all possible;

and

including instantly applying the computer-generated, computer-usable new interaction-specific knowledge bases on the object to immediately improve or optimize the performance of the object.

81. A method as in claim 78 for optimizing a specific growth performance of a selected growing object;

and wherein the control variables are growth-controlling variables selected from the group consisting of equipment, materials, parts, procedures, and environment.

82. A method as in claim 81 wherein the object is a living object selected form the group consisting of human, animal, plants, virus, and bacteria, and

the specific growth performance relates to a physical, mental, medical, psychological, or physiological activity or condition of the object;

and

including selecting the control variables from the group consisting of type, tempo, frequency, and intensity of sound, noise, or vibration;

smell;

lighting;

temperature;

humidity;

pressure on the object;

video or audio signal from tapes or disks; and

other factors that stimulate the living object physically, mentally, medically, psychologically, or physiologically.

83. A method as in claim 78 for computer-generating new knowledge bases for optimizing or improving a physical or mental activity or condition of a living object selected from the group consisting of human, animal, bacteria, virus, germs, and plant;

the activity or condition being selected from the group consisting of learning, relaxing, resting, sleeping, dreaming, enjoying, and being stimulated or entertained; and

the control variables being selected for the group consisting of time;

equipment;

procedures;

materials; and

parts related to said activity or condition;

type, tempo, frequency, and intensity of sound, noise, or vibration;

smell;

lighting;

temperature;

humidity;

pressure on the object;

video or audio signal from tapes or disks; and

other factors that stimulate the living object physically, mentally, medically, psychologically, or physiologically.

84. A method as in claim 78 including minimizing or eliminating at least one of error selected from the group consisting of those due to sampling, deficient knowledge bases, misused or miscalibrated sensors, actuator backlash, drifting equipment, and statistical data fluctuations associated with the automatic computerized experimentation.”