Method for optimizing configuration of pick-and-place machine

• US 20040249492A1
• Filed: 06/06/2003
• Published: 12/09/2004
• Est. Priority Date: 06/06/2003
• Status: Active Grant

• Current Assignee: Visteon Global Technologies Incorporated (Visteon Corporation)
• Original Assignee: Visteon Global Technologies Incorporated (Visteon Corporation)
• Inventors: Welch, Kurt D.
• Granted Patent: US 7,076,313 B2
• Time in Patent Office: Days
• US Class Current: 700/97
• CPC Class Codes: B25J 9/1664 characterised by motion, path, trajectory planning
• CPC Class Codes: G05B 2219/40384 Optimize teached path by data acquisition followed by genetic algorithm
• CPC Class Codes: G05B 2219/40463 Shortest distance in time, or metric, time optimal
• CPC Class Codes: G05B 2219/45063 Pick and place manipulator
• CPC Class Codes: Y10T 29/5313 Means to assemble electrical device

Abstract

A method for optimizing the configuration of a pick-and-place machine utilizes a genetic algorithm that creates an initial population of possible configurations and selects an optimum configuration based upon lowest cycle time. The method then creates a next generation by selecting possible configurations from the prior generation and randomly mutating instructions. The method compares the mutated configuration having the lowest cycle time from the next generation and the configuration having the lowest cycle time from the prior generation, and selects the optimum configuration. The steps are repeated to evaluate additional generations of mutated configurations. After several mutated generations have been evaluated, the method creates a new initial population and evaluates the new initial population and mutated generations derived therefrom, compares the optimum configuration from the new initial populations with the optimum configuration from the prior initial population, and selects the optimum configuration. In this manner, the method may evaluate several initial populations and mutated generations therefrom in selecting the optimum configuration for use in operating the machine.

Claims

1. A method of optimizing a configuration of a pick-and-place machine comprising a plurality of placement devices, said configuration comprising a series of instructions that direct each placement device to pick up one of a plurality of electronic components at one of a plurality of feeders and place the component at a designated location on a substrate, said method comprising:

(a) creating an initial population of possible configurations to be a first generation,

(b) estimating cycle time for each possible configuration and selecting the possible configuration having a lowest estimated cycle time as an optimum configuration;

(c) creating a next generation of possible configurations by selecting possible configurations from the prior generation;

(d) randomly mutating instructions of the possible configurations of the next generation;

(e) estimating cycle time for each possible configuration of the next population,

(f) comparing the cycle time of the possible configuration of the next generation having the lowest cycle time and the cycle time of the optimum configuration, and selecting the possible configuration having the lower cycle time as the optimum configuration;

(g) repeating steps (c) through (g); and

(h) selecting the optimum configuration for use in operating the pick-and-place machine.

2. A method in accordance with claim 1 wherein the step of creating a next generation comprises selecting possible configurations from the prior generation based upon a fitness probability.

3. A method in accordance with claim 2 wherein the fitness probability is based upon a fitness factor determined by a standard deviation in individual cycle times of the placement devices.

4. A method in accordance with claim 1 wherein the step of mutating at least one instruction of the possible configuration comprises randomly selecting an instruction that directs a particular placement device to pick up a particular component, and directing a different placement device to pick up the particular component.

5. A method in accordance with claim 1 wherein the step of mutating at least one instruction of the possible configuration comprises randomly selecting an instruction that directs a particular placement device to pick up a particular component at a particular time, and directing the particular placement device to pick up the particular component at a different time.

6. A method in accordance with claim 1 wherein the possible configuration comprises a feeder set up, and wherein the step of mutating the possible configuration comprises changing the feeder set up.

7. A method in accordance with claim 1 wherein the step of creating the next generation includes selecting the optimum configuration.

8. A method in accordance with claim 1 further comprising, following a predetermined stop event, creating a new initial population;

repeating steps (b) through (h); and

selecting the optimum configuration in step (i) from the optimum configurations derived from the prior initial population and the new initial population.

9. A method of optimizing a configuration of a pick-and-place machine comprising a plurality of placement devices, said configuration comprising a series of instructions that direct each placement device to pick up one of a plurality of electronic components at one of a plurality of feeders and place the component at a designated location on a substrate, said method comprising:

(a) creating a first population of possible configurations to be a first generation,

(b) estimating a cycle time for each possible configuration in the generation;

(c) selecting the possible configuration having the lowest estimated cycle time as an optimum configuration,

(d) determining a fitness factor for each possible configuration;

(e) creating a next generation of possible configurations by selecting possible configurations from the prior generation based upon a fitness probability using the fitness factors;

(f) randomly mutating instructions of the possible configuration of the next generation;

(g) estimating a cycle time of each possible configuration in the next population;

(h) comparing the cycle time of the possible configuration of the next population having the lowest cycle time and the cycle time of the optimum configuration, and selecting the possible configuration having the lowest cycle time as the optimum configuration;

(i) repeating steps (d) through (i) until a stop event; and

(j) creating a next population of possible configurations and repeating steps (b) through (h); and

(k) selecting the optimum configuration based upon the lowest cycle time between the optimum configurations derived from the first population and the optimum configuration derived from the next population.

10. A method in accordance with claim 9 wherein the fitness probability is determined by, for each possible configuration, calculating individual cycle times for the possible devices, determining a fitness factor based upon a standard deviation for the cycle times, squaring the difference between the fitness factor of each possible configuration and the highest fitness factor for the prior generation;

and randomly selecting possible configurations for the next generation based upon the squared difference.

11. A method in accordance with claim 9 wherein the step of mutating at least one instruction of the possible configuration comprises at least one of

(a) randomly selecting an instruction that directs a particular placement device to pick up a particular component, and selecting a different placement device to pick up the particular component;

(b) randomly selecting an instruction that directs a particular placement device to pick up a particular component at a particular time, and selecting a different time, or

(c) randomly changing the feeder set up.