John J Koza

US Patent:

6360191, Mar 19, 2002

Filed:

Jan 5, 1999

Appl. No.:

09/225643

Inventors:

John R. Koza - Los Altos Hills CA 94022
David Andre - Menlo Park CA
Martin A. Keane - Chicago IL

Assignee:

John R. Koza - Los Altos Hills CA

International Classification:

G06G 748

US Classification:

703 6, 703 2, 703 7, 703 14, 703 22, 716 1, 716 2

Abstract:

An automated design process and apparatus for use in designing complex structures, such as circuits, to satisfy prespecified design goals, using genetic operations. The present invention uses a population of entities which may be evolved to generate structures that may potentially satisfy the design goals. The behavior of such generated structures is evaluated in view of the design goals, and those structures more closely meeting the design goals are evolved further until a structure is generated that either meets the prespecified design goal or some other process completion criteria. In this manner, a design complex structure may be obtained.

US Patent:

6424959, Jul 23, 2002

Filed:

Jun 17, 1999

Appl. No.:

09/336373

Inventors:

John R. Koza - Los Altos Hills CA 94022

Assignee:

John R. Koza - Los Altos Hills CA

International Classification:

G06F 9445

US Classification:

706 13, 706919, 706921, 714 8, 714724, 716 8

Abstract:

The present invention consists of a method and apparatus for the automatic creation of the topology, component sizing, placement, and routing of complex structures, such as electronic circuits or mechanical systems, to satisfy prespecified high-level design goals. The present invention uses a population of entities which are evolved over a series of generations by an iterative process involving the application of operations, such as mutation, crossover, reproduction, and architecture-altering operations. The individuals in the population are each developed, in a developmental process, into a structure that may potentially satisfy the design goals. The present invention also determines the placement of components within the developing structure and determining the routing of the connecting means (wires for electrical circuits) between the components. The behavior of the structures is evaluated in relation to the high-level design goals, and those structures more closely satisfying the design goals are evolved further until a structure is generated that either meets the prespecified design goal. In this manner, the topology, component sizing, placement, and routing of a complex structure may be obtained.

US Patent:

6532453, Mar 11, 2003

Filed:

Apr 12, 1999

Appl. No.:

09/290521

Inventors:

John R. Koza - Los Altos Hills CA 94022
David Andre - Menlo Park CA
Martin A. Keane - Chicago IL

Assignee:

John R. Koza - Los Alto Hills CA

International Classification:

G06F 1518

US Classification:

706 13

Abstract:

The present invention is a genetic programming problem solver that automatically generates computer programs to solve problems. The genetic programming problem solver incorporates architecture-altering operations. In one embodiment, the genetic programming problem solver uses architecture-altering operations for automatically defined functions and loops, together with indexed memory, to generate the resulting computer programs. In a second embodiment, the genetic programming problem solver uses architecture-altering operations of automatically defined function, loops, recursions, and stores to generate the resulting computer programs.

US Patent:

6564194, May 13, 2003

Filed:

Sep 10, 1999

Appl. No.:

09/393863

Inventors:

John R. Koza - Los Altos Hills CA 94022
Martin A. Keane - Chicago IL
Jessen Yu - Sunnyvale CA
William Mydlowec - Sunnyvale CA

Assignee:

John R. Koza - Los Altos Hills CA

International Classification:

G06N 312

US Classification:

706 13, 706 52, 364472, 364476

Abstract:

A general automated method for synthesizing the design of both the topology and parameter values for controllers is described. The automated method automatically makes decisions concerning the total number of signal processing blocks to be employed in the controller, the type of each signal processing block, the topological interconnections between the signal processing blocks, the values of all parameters for the signal processing blocks, and the existence, if any, of internal feedback between the signal processing blocks within the controller. The general automated method can simultaneously optimize prespecified performance metrics (such as minimizing the time required to bring the plant outputs to the desired values as measured by the integral of the time-weighted absolute error or the integral of the squared error), satisfy time-domain constraints (such as overshoot, disturbance rejection, limits on control variables, and limits on state variables), and satisfy frequency domain constraints (bandwidth).

US Patent:

6767284, Jul 27, 2004

Filed:

Mar 14, 2000

Appl. No.:

09/524857

Inventors:

John R. Koza - Los Altos Hills CA 94022

Assignee:

John R. Koza - Los Altos CA

International Classification:

A63F 924

US Classification:

463 29, 463 42

Abstract:

Skill games are described that are implemented using network communications. The subject matter of the present invention concerns games of skill that are legal, under current law, in most states of the United States and in many jurisdictions of other countries and the game includes a mechanism for determining if a player is eligible.

US Patent:

6964608, Nov 15, 2005

Filed:

Oct 17, 2000

Appl. No.:

09/690923

Inventors:

John R. Koza - Los Altos Hills CA, US

International Classification:

A63F009/24

US Classification:

463 9, 379 9313, 340323 R, 700 92

Abstract:

Skill games are described that are implemented using network communications. The subject matter of the present invention concerns games of skill that are legal, under current law, in most states of the United States and in many jurisdictions of other countries and the game includes a mechanism for determining if a player is eligible.

US Patent:

7117186, Oct 3, 2006

Filed:

Jan 30, 2003

Appl. No.:

10/355443

Inventors:

John R. Koza - Los Altos Hills CA, US
Martin A. Keane - Chicago IL, US
Jessen Yu - Sunnyvale CA, US
William Mydlowec - Sunnyvale CA, US

International Classification:

G08N 5/00

US Classification:

706 13, 706 45

Abstract:

A general automated method for synthesizing the design of both the topology and parameter values for controllers is described. The automated method automatically makes decisions concerning the total number of signal processing blocks to be employed in the controller, the type of each signal processing block, the topological interconnections between the signal processing blocks, the values of all parameters for the signal processing blocks, and the existence, if any, of internal feedback between the signal processing blocks within the controller. The general automated method can simultaneously optimize prespecified performance metrics (such as minimizing the time required to bring the plant outputs to the desired values as measured by the integral of the time-weighted absolute error or the integral of the squared error), satisfy time-domain constraints (such as overshoot, disturbance rejection, limits on control variables, and limits on state variables), and satisfy frequency domain constraints (bandwidth).

US Patent:

8356000, Jan 15, 2013

Filed:

Apr 13, 2000

Appl. No.:

09/548637

Inventors:

John R. Koza - Los Altos Hills CA, US
Oscar Stiffelman - Seattle WA, US

International Classification:

G06N 3/12

US Classification:

706 13

Abstract:

A method and apparatus for the automatic creation of novel designs, specifically electronic circuits, controllers, antennas, and mechanical systems to satisfy prespecified design goals, using search procedures, such as genetic programming, genetic algorithms, simulated annealing and hill climbing is described. Further, the techniques include automatically creates designs which do not posses key characteristics of preexisting technology. The present invention uses a population of entities which may be evolved to generate structures that may potentially satisfy the design goals. The behavior of the structures is evaluated in view of the design goals, and the structures are compared to a preexisting structure. Those structures more closely meeting the design goals and not similar to the preexisting structure are favored further until a structure is generated that either meets the prespecified design goal or some other process completion criteria. In this manner, a novel structure may be obtained.

Author:

John R. Koza

ISBN #:

0262111705

Author:

John R. Koza

ISBN #:

0262111896

Author:

John R. Koza

ISBN #:

0262611279