Method for optimization of element placement in a thinned array

• US 5,774,690 A
• Filed: 09/14/1995
• Issued: 06/30/1998
• Est. Priority Date: 09/14/1995
• Status: Expired due to Fees

• Current Assignee: the united states of america as represented by the secretary of the navy
• Original Assignee: The United States of America As Represented By The Secretary of Agriculture
• Inventors: O'Neill, Daniel J.
• Primary Examiner(s): Zanelli, Michael
• Assistant Examiner(s): FREJD, RUSSELL WARREN
• Application Number:
  US08/536,310
• Time in Patent Office: 1,020 Days
• Field of Search: 395/500, 395/13, 364/726, 364/491, 342/379, 382/159, 367/13, 367/19, 367/22, 367/56, 367/62, 367/905
• US Class Current: 716/123
• CPC Class Codes: G01V 1/20 Arrangements of receiving elements, e.g. geophone pattern
• CPC Class Codes: G06N 3/126 Evolutionary algorithms, e.g. genetic algorithms or genetic programming

Abstract

A method for generating a thinned array having an optimal directional response wherein an initial population of possible element occupancy patterns is generated and a relative fitness value for each of the element occupancy patterns in the initial population is determined. Successive populations are generated by manipulating the element occupancy patterns using reproduction, crossover, mutation, elitism, and equality operators. The array response of each occupancy pattern in the successive population is evaluated to determine a fitness value for each. The process of generating and evaluating successive populations is repeated for a specified number of generations, until the maximum population fitness does not significantly increase between generations, or until the population diversity decreases.

Claims

1. A method for determining placement of elements within a discrete lattice array comprising the steps of:

building a plurality of arrays, wherein each array is defined by an element placement pattern indicating the placement and weighting of each element within the array, said plurality of arrays defining an initial population;

determining a fitness value for each array within said plurality of arrays in the initial population using a fitness function which provides a measure of a relationship between peak sidelobe level and mainlobe width of an angular response of an array;

generating a reproduction population by selecting arrays from the initial population with probability proportional to their fitness values;

generating a successor population from said reproduction population, said successor population comprising a plurality of arrays wherein each array is defined by an element placement pattern indicating the placement and weighting of each element within the array;

generating a fitness value for each array in the successor population using said fitness function; and

repeating the steps of generating a reproduction set, generating a successor population, and generating a fitness value for each array in the successor population until the successor population meets a threshold.

2. The method of claim 1 wherein said step of generating a successor population from said reproduction population comprises the steps of:

generating a crossover population from said reproduction population by applying a crossover operator; and

applying a mutation operator to said crossover population.

3. The method of claim 2 wherein said step of generating a successor population from said reproduction population further comprises applying an elitism operator.

4. The method of claim 3 wherein said step of generating a successor population from said reproduction population further comprises applying an equality operator.

5. The method of claim 4 wherein said fitness function is given by ##EQU2## where ml is the 6 dB mainlobe width in degrees and sl is the maximum sidelobe level in dB down from the mainlobe response.

6. A method for determining placement of elements within a discrete lattice array comprising the steps of:

building a plurality of arrays, said plurality of arrays defining an initial population;

determining a fitness value for each of the arrays in the initial population;

generating a reproduction population by selecting arrays from the initial population with probability proportional to their fitness values;

generating a successor population from said reproduction population;

generating a fitness value for each array in the successor population by measuring the angular response of each of the arrays in the successor population and using a fitness function given by ##EQU3## where ml is the 6 dB mainlobe width in degrees and sl is the maximum sidelobe level in dB down from the mainlobe response, to generate the fitness value for each array based on the angular response of the array; and

repeating the steps of generating a reproduction set, generating a successor population, and generating a fitness value for each array in the successor population until the successor population meets a threshold.

7. The method of claim 6 wherein said step of generating a successor population from said reproduction population comprises the steps of:

generating a crossover population from said reproduction population by applying a crossover operator; and

applying a mutation operator to said crossover population.

8. The method of claim 7 wherein said step of generating a successor population from said reproduction population further comprises applying an elitism operator.

9. The method of claim 8 wherein said step of generating a successor population from said reproduction population further comprises applying an equality operator.

10. A method for determining placement of elements within a discrete lattice array comprising the steps of:

building a plurality of arrays, said plurality of arrays defining an initial population;

determining a fitness value for each of the arrays in the initial population;

generating a reproduction population by selecting arrays from the initial population with probability proportional to their fitness values;

generating a successor population from said reproduction population;

generating a fitness value for each array in the successor population; and

repeating the steps of generating a reproduction set, generating a successor population, and generating a fitness value for each array in the successor population until the successor population meets a threshold;

wherein said step of generating a successor population from said reproduction population comprises generating a crossover population from said reproduction population by applying a crossover operator, applying a mutation operator to said crossover population, and applying an elitism operator.

11. The method of claim 10 wherein said step of generating a successor population from said reproduction population further comprises applying an equality operator.

12. The method of claim 11 wherein said step of generating a fitness value for each array in the successor population comprises the steps of:

measuring the angular response of each of the arrays; and

using a fitness function to generate a fitness value for each array based on the angular response of the array.

13. The method of claim 12 wherein said fitness function is given by ##EQU4## where ml is the 6 dB mainlobe width in degrees and sl is the maximum sidelobe level in dB down from the mainlobe response.

14. A method for determining placement of elements within a discrete lattice array comprising the steps of:

building a plurality of arrays, wherein each array is defined by an element placement pattern indicating the placement and weighting of each element within the array it defines, said plurality of arrays defining an initial population;

determining a fitness value for each of the arrays in the initial population;

generating a reproduction population by selecting arrays from the initial population with probability proportional to their fitness values;

generating a successor population from said reproduction population;

generating a fitness value for each array in the successor population by translating each element placement pattern into a thinned array, wherein the element placement and element weighting for the thinned array is described by the element placement pattern, measuring the angular response of each of the thinned arrays in the successor population, and using a fitness function to generate a fitness value for each element placement pattern based on the angular response of the thinned array defined by the element placement pattern; and

repeating the steps of generating a reproduction set, generating a successor population, and generating a fitness value for each array in the successor population until the successor population meets a threshold.

15. The method of claim 14 wherein said step of generating a successor population from said reproduction population comprises the steps of:

generating a crossover population from said reproduction population by applying a crossover operator; and

applying a mutation operator to said crossover population.

16. The method of claim 15 wherein said step of generating a successor population from said reproduction population further comprises applying an elitism operator.

17. The method of claim 16 wherein said step of generating a successor population from said reproduction population further comprises applying an equality operator.