CLASS-BASED DISTRIBUTED EVOLUTIONARY ALGORITHM FOR ASSET MANAGEMENT AND TRADING

• US 20100274736A1
• Filed: 04/28/2010
• Published: 10/28/2010
• Est. Priority Date: 04/28/2009
• Status: Active Grant

• Current Assignee: Cognizant Technology Solutions U.S. Corporation (Cognizant Technology Solutions Corp.)
• Original Assignee: Genetic Finance Holdings Limited AMS Trustees Limited
• Inventors: Hodjat, Babak, Shahrzad, Hormoz
• Granted Patent: US 8,768,811 B2
• Time in Patent Office: Days
• US Class Current: 705/36.R
• CPC Class Codes: G06N 3/126 Evolutionary algorithms, e.g. genetic algorithms or genetic programming
• CPC Class Codes: G06Q 40/04 Trading; Exchange, e.g. stocks, commodities, derivatives or currency exchange
• CPC Class Codes: G06Q 40/06 Asset management; Financial planning or analysis

Abstract

A server computer and a multitude of client computers form a network computing system that is scalable and adapted to continue to evaluate the performance characteristics of a number of genes generated using a software application. Each client computer continues to periodically receive data associated with the stored genes stored in its memory. Using this data, the client computers evaluate the performance characteristic of their genes by comparing a solution provided by the gene with the periodically received data associated with that gene. Accordingly, the performance characteristic of each gene may be updated and varied with each periodically received data. The performance characteristic of a gene defines its fitness. The genes may be virtual asset traders that recommend trading options. The genes may be assigned initially to different classes to improve convergence but may later be decided to merge with genes of other classes to improve diversity.

Claims

1. A networked computer system comprising M client computers each assigned to a different one of N classes, a first one of the M client computers comprising:

a memory operative to store a plurality of genes characterized by a plurality of conditions and a subset of indicators associated with the class to which the client computer is assigned;

a communication port adapted to continue to periodically receive a first set of data associated with the plurality of genes; and

a processor operative to evaluate performance characteristic of each of the plurality of genes by comparing a solution provided by each gene with the periodically received first set of data associated with that gene, the performance characteristic of each gene defining a fitness of that gene and being adjusted with every periodically received data.

2. The networked computer system of claim 1 wherein at least two of the subsets of plurality of indicators comprise overlapping indicators.

3. The networked computer system of claim 1 wherein the first set of data associated with each gene is historical trading data covering a first plurality of days and wherein the solution provided by each gene is a trading recommendation made by the gene representative of the fitness of the gene.

4. The networked computer system of claim 3 wherein each client computer is configured to discard one or more of the genes it receives after evaluating the fitness of the genes for the first plurality of days, each of the discarded genes having a fitness that falls below a first predefined threshold value, each of the remaining genes being a surviving gene.

5. The networked computer system of claim 4 wherein said first client computer is configured to:

evaluate the fitness of its surviving genes using a second set of historical data covering a second plurality days;

discard the genes whose fitness as evaluated during a sum of the first and second plurality of days is below a second predefined threshold value;

repeat the evaluation and the discard operations S times where S is an integer equal to or greater than one, wherein genes surviving S evaluation and discard operations define a first elitist gene pool; and

transmit one or more of the genes of the first elitist pool to at least one server computer.

6. The networked computer system of claim 5 wherein the first and second predefined threshold values are equal.

7. The networked computer system of claim 5 wherein said at least server computer is adapted to:

receive the one or more genes of the elitist gene pool of the client computers;

store the received one or more genes in a server gene pool; and

transmit, for fitness evaluation using a third set of historical trading data covering a third plurality of days, only to at least a first subset of the plurality of client computers of class j, one or more genes previously evaluated by and received from the client computers of class j, wherein j is an integer varying from 1 to N.

8. The networked computer system of claim 4 wherein said at least one server computer is adapted to:

receive one or more genes from the client computers in class k;

store the received class k genes in a server gene pool; and

transmit, for fitness evaluation using a third set of historical trading data covering a third plurality of days, to at least a first subset of the plurality of class i client computers, one or more class k genes, wherein k and i are different integers representing different classes.

9. The networked computer system of claim 8 wherein said server computer causes class k genes and class i genes to be merged thereby generating a new class of genes represented by a union of the subset of indicators of class k and class i genes, wherein said class k and class i genes are merged only if a measure of convergence of class k and class i genes meets a predefined condition.

10. The networked computer system of claim 8 wherein said server computer causes class k genes and class i genes to be merged thereby generating a new class of genes represented by a new subset of indicators randomly generated, wherein said class k and class i genes are merged only if a measure of convergence of class k and class i genes meets a predefined condition.

11. The networked computer system of claim 8 wherein said third plurality of days do not overlap with said first plurality of days.

12. The networked computer system of claim 5 wherein a gene is transmitted to said at least one or more server computer only if the gene'"'"'s fitness after S evaluation and discard operations is computed as being higher than the fitness of the gene when previously stored previously by the one or more servers.

13. The networked computer system of claim 4 wherein the genes stored in the client computers are generated in accordance with computer instructions stored in the client computers and executed by the processors of the first client computer.

14. The networked computer system of claim 8 wherein said server pool is configured to store a fixed number of genes.

15. The networked computer system of claim 8 wherein said sever is further adapted to combine a fitness value associated with each gene that the server receives from a client computer with a corresponding fitness value previously stored by the server for that gene.

16. The networked computer system of claim 4 wherein said plurality of indicators are evaluated as a logical expression conjoined by logical AND and/or modified by logical NOT operations.

17. The networked computer system of claim 16 wherein the at least one action recommended by each gene is selected from a group consisting of buy, sell, hold, long exit and short exit recommendations.

18. A networked computer system comprising a server computer adapted to:

receive one or more genes from a plurality of client computers assigned to N different classes;

each gene characterized by a plurality of conditions and a subset of indicators associated with the class to which the client computer is assigned;

store the received one or more genes in a server gene pool; and

transmit to each of at least a first subset of the plurality of client computers one or more genes of the server gene pool for fitness evaluation spanning W trading days, the genes received from the plurality of client computers generated by evaluating performance characteristic of each of the N genes by comparing a solution provided by each gene with the periodically received data associated with the gene, the performance characteristic of each gene being adjusted with each periodically received data and defining a fitness of the gene, each gene further characterized by at least one action.

19. The networked computer system of claim 18 wherein the data associated with each gene is historical trading data and wherein the solution provided by each gene is a trading recommendation made by the gene.

20. The networked computer system of claim 19 wherein said first one of the plurality of client computers is configured to discard M of the N genes after evaluating the fitness of the N genes for P days, each of the discarded genes having a fitness that falls below a first predefined threshold value, each of the remaining N−

M genes being a surviving gene, wherein M and P are integers and wherein M is smaller than N.

21. The networked computer system of claim 20 wherein said first one of the plurality of client computers is further configured to:

evaluate, for a plurality of Q days in addition to the P days, the fitness of the remaining genes;

discard the genes whose fitness as evaluated during a sum of the P+Q days is below a second predefined threshold value; and

repeat the evaluation and the discard operations at least S times, wherein Q and S are integers, and wherein S is equal to or greater than 1, wherein genes surviving S evaluation and discard operations define the first elitist gene pool, each gene being V trading days olds, wherein V is defined by the P+S*Q, wherein * represents a multiplication operation.

22. The networked computer system of claim 19 wherein said server computer receives only genes whose fitness is higher than the fitness of genes previously stored in the server gene pool.

23. The networked computer system of claim 22 wherein said server gene pool is configured to store a fixed number of genes.

24. The networked computer system of claim 22 wherein said sever is further adapted to combine a plurality of fitness values associated with each gene the server receives with a corresponding fitness value stored in the server for that gene.

25. The networked computer system of claim 19 wherein said plurality of indicators are evaluated as a logical expression conjoined by logical AND and/or modified by logical NOT operations.

26. The networked computer system of claim 25 wherein the at least one action recommended by each gene is selected from a group consisting of buy, sell, hold, long exit and short exit recommendations.

27. The networked computer system of claim 21 wherein said server computer transmits for fitness evaluation using a third set of historical trading data covering a third plurality of days, only to at least a first subset of the plurality of client computers of class j, one or more genes previously evaluated by and received from the client computers of class j, wherein j is an integer.

28. The networked computer system of claim 20 wherein said computer server transmit, for fitness evaluation using a third set of historical trading data covering a third plurality of days, to at least a first subset of the plurality of class i client computers, one or more class k genes, wherein k and i are different integers representing different classes.

29. The networked computer system of claim 28 wherein said server computer causes class k genes and class i genes to be merged thereby generating a new class of genes represented by a union of the subset of indicators of class k and class i genes, wherein said class k and class i genes are merged only if a measure of convergence of class k and class i genes meets a predefined condition.

30. The networked computer system of claim 28 wherein said server computer causes class k genes and class i genes to be merged thereby generating a new class of genes represented by a new subset of indicators randomly generated, wherein said class k and class i genes are merged only if a measure of convergence of class k and class i genes meets a predefined condition.

31. A method for solving a computational problem, the method comprising:

storing N genes, each gene characterized by a plurality of conditions and a subset of indicators associated with a class to which the gene belongs;

wherein N is an integer greater than one;

continuing to periodically receive data associated with the N genes; and

evaluating performance characteristic of each of the N genes by comparing a solution provided by each gene with the periodically received data associated with that gene, the performance characteristic of each gene being adjusted with each periodically received data and defining a fitness of the gene, each gene being further characterized by at least one action.

32. The method of claim 31 wherein the data associated with each gene is historical trading data and wherein the solution provided by each gene is a trading recommendation made by the gene.

33. The method of claim 32 further comprising:

discarding M of the N genes after evaluating the fitness of the N genes for a first plurality of days, each of the discarded genes having a fitness that falls below a first predefined threshold value, each of the remaining N−

M genes being a surviving gene, wherein M and P are integers and wherein M is smaller than N.

34. The method of claim 33 further comprising:

evaluating the fitness of the surviving genes using a second set of historical data covering a second plurality of days;

discarding genes whose fitness as evaluated during a sum of the first and second plurality of days is below a second predefined threshold value;

repeating the evaluation and the discarding operations S times where S is an integer greater than one, wherein genes surviving S evaluation and discarding operations define a first elitist gene pool; and

transmitting one or more of the genes of the first elitist pool to a server.

35. The method of claim 34 wherein the first and second predefined threshold values are equal.

36. The method of claim 34 further comprising:

receiving the one or more genes of the elitist gene pool;

storing the received one or more genes; and

transmitting, for fitness evaluation using a third set of historical trading data covering a third plurality of days, only to at least a first subset of the plurality of client computers of a first class, one or more genes previously evaluated by and received from client computers of the first class

37. The method of claim 34 further comprising:

receiving the one or more genes of the elitist gene pool;

storing the received one or more genes; and

transmitting, for fitness evaluation using a third set of historical trading data covering a third plurality of days, only to at least a first subset of the plurality of class i client computers, one or more class k genes previously evaluated by and received from the client computers of a second class wherein k and i are different integers representing different classes.

38. The method of claim 37 further comprising:

merging class k and class i genes to generate a new class of genes represented by a union of the subset of indicators of class k and class i genes.

39. The method of claim 37 further comprising:

merging class k and class i genes to generate a new class of genes represented by a new subset of indicators randomly generated.

40. The method of claim 37 wherein said third plurality of days do not overlap with said first plurality of days.

41. The method of claim 34 further comprising:

transmitting a gene for storage only if the gene'"'"'s fitness after S evaluation and discard operations is computed as being higher than the fitness of the gene at a time of previous storage.

42. The method of claim 33 further comprising:

generating the genes stored in the client computers in accordance with computer instructions stored in and executed by the client computers.

43. The method of claim 37 further comprising:

storing a fixed number of genes in a server gene pool.

44. The method of claim 37 further comprising:

combining a plurality of fitness values associated with each gene received by the server with a corresponding fitness value previously stored in the server for that gene.

45. A stand alone computer comprising:

at least one processing core configured as a sever;

a plurality of processing cores configured as clients each assigned to a different one of N classes, each of the plurality of clients comprising;

a memory operative to store N genes, each gene characterized by a plurality of conditions, a subset of indicators and at least one action, wherein N is an integer greater than one;

a port adapted to continue to periodically receive data associated with the N genes;

each client processing core configured to evaluate performance characteristic of each of the N genes by comparing a solution provided by each gene with the periodically received data associated with that gene, the performance characteristic of each gene being adjusted with each periodically received data and defining a fitness of the gene.

46. The stand alone computer of claim 45 wherein the data associated with each gene is historical trading data and wherein the solution provided by each gene is a trading recommendation made by the gene.

47. The stand alone computer of claim 46 wherein a first one of the plurality of client processing cores is configured to discard M of the N genes after evaluating the fitness of the N genes for P days, each of the discarded genes having a fitness that falls below a first predefined threshold value, each of the remaining N−

M genes being a surviving gene, wherein M and P are integers and wherein M is smaller than N.

48. The stand alone computer of claim 47 wherein said first one of the client processing cores is further configured to:

evaluate, for a plurality of Q days in addition to the P days, the fitness of the remaining genes;

discard the genes whose fitness as evaluated during a sum of the P+Q days is below a second predefined threshold value; and

repeat the evaluation and the discard operations at least S times, wherein Q and S are integers, and wherein S is equal to or greater than 1, wherein genes surviving S evaluation and discard operations define a first elitist gene pool each gene being V trading days olds, wherein V is defined by the P+S*Q, wherein * represents a multiplication operation; and

transmit the first elitist pool of genes to the server processing core.

49. The stand alone computer of claim 48 wherein the first and second predefined threshold values are equal.

50. The stand alone computer of claim 48 wherein said sever processing core is adapted to:

selectively receive the first elitist gene pool from the first client processing core;

store the received genes in a server gene pool; and

transmit for fitness evaluation only to at least a first subset of the plurality of client processing cores of class j, one or more genes previously evaluated by and received from the client processing cores of class j, wherein j is an integer varying from 1 to N.

51. The stand alone computer of claim 48 wherein said sever processing core is adapted to:

selectively receive the first elitist gene pool from the first client processing core;

store the received genes in a server gene pool; and

transmit for fitness evaluation to at least a first subset of a plurality of client processing cores of class j, one or more genes previously evaluated by and received from the client processing cores of class k, wherein k and j are unequal integers.

52. The stand alone computer of claim 51 wherein said server processing core causes class k genes and class i genes to be merged to generate a new class of genes represented by a union of the subset of indicators of class k and class i genes.

53. The stand alone computer system of claim 51 wherein said server processing core causes class k genes and class i genes to be merged to generate a new class of genes represented by a new subset of indicators randomly generated.

54. The stand alone computer of claim 48 wherein each of the at least a first subset of the plurality of client processing cores is adapted to:

receive one or more genes from the server processing core;

evaluate a fitness of the received genes by comparing trading recommendations made by each of the received genes with an associated historical trading data spanning the W trading days;

discard genes whose fitness is less than a third threshold value;

report discarded genes to the server processing core; and

transmit genes that are not discarded to the server processing core for storage in the server gene pool.

55. The stand alone computer of claim 54 wherein said W trading days and V trading days do not overlap.

56. The stand alone computer of claim 54 wherein said third threshold value is equal to the first and second threshold values.

57. The stand alone computer of claim 54 wherein said genes are transmitted to the server processing core only if their fitness is higher than the fitness of genes previously stored in the server pool.

58. The stand alone computer of claim 46 wherein the genes stored in each client processing core are generated in accordance with computer instructions stored in and executed by the stand alone computer.

59. The stand alone computer of claim 54 wherein said server processing core is configured to store a fixed number of genes in its pool.

60. The stand alone computer of claim 57 wherein said sever processing core is further adapted to combine a plurality of fitness values associated with each gene the server processing core receives from the at least first subset of the plurality of client processing cores with a corresponding fitness value stored in the server processing core for that gene.

61. The stand alone computer of claim 47 wherein said plurality of indicators are evaluated as a logical expression conjoined by logical AND and/or modified by logical NOT operations.

62. The stand alone computer of claim 48 wherein the at least one action recommended by each gene is selected from a group consisting of buy, sell, hold, long exit and short exit operations.

63. The stand alone computer of claim 51 wherein class k genes and class i genes are merged only if a measure of convergence of class k and class i genes meets a predefined condition.

64. The stand alone computer of claim 53 wherein class k genes and class i genes are merged only if a measure of convergence of class k and class i genes meets a predefined condition.

65. The stand alone computer of claim 60 wherein said stand alone computer is a mainframe computer.