SPECULATIVE ASYNCHRONOUS SUB-POPULATION EVOLUTIONARY COMPUTING

• US 20130006902A1
• Filed: 08/23/2012
• Published: 01/03/2013
• Est. Priority Date: 06/30/2011
• Status: Abandoned Application

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Cantin, Jason F.
• Application Number:
  US13/592,851
• Publication Number:
  US 20130006902A1
• Time in Patent Office: Days
• US Class Current: 706/13
• CPC Class Codes: G06N 20/00 Machine learning
• CPC Class Codes: G06N 3/126 Evolutionary algorithms, e.g. genetic algorithms or genetic programming

Abstract

A tool computes fitness values for a first generation of a first sub-population of a plurality of sub-populations. A population of candidate solutions for an optimization problem was previously divided into the plurality of sub-populations. The population of candidate solutions was created for an iterative computing process in accordance with an evolutionary algorithm to identify a most fit candidate solution for the optimization problem. The tool determines a speculative ranking of the first generation of the first sub-population prior to the fitness values being computed for all candidate solutions in the first generation of the first sub-population. The tool generates a next generation of the first sub-population based, at least in part, on the speculative ranking prior to completion of computation of the fitness values for the first generation of the first sub-population.

Claims

1. A method comprising:

receiving fitness values for a first generation of a first sub-population of a plurality of sub-populations, wherein a population of candidate solutions for an optimization problem was previously divided into the plurality of sub-populations, wherein the population of candidate solutions was created for an iterative computing process in accordance with an evolutionary algorithm to identify a most fit candidate solution for the optimization problem;

determining that the first generation of the first sub-population does not satisfy a termination criterion for the iterative computing processing;

determining that the first generation corresponds to a later iteration of the iterative computing process than a second generation of a second sub-population of the plurality of sub-populations;

determining that a difference between the first generation and the second generation does not exceed a termination speculation threshold; and

generating a third generation of the first sub-population responsive to said determining that the difference between the first generation and the second generation does not exceed the termination speculation threshold, wherein said generating the third generation of the first sub-population is based, at least in part, on the fitness values.

2. The method of claim 1, wherein said generating the third generation of the first sub-population is in accordance with one or more evolutionary computing techniques implemented in accordance with the evolutionary algorithm.

3. The method of claim 2, wherein the evolutionary computing techniques comprise at least one of crossover, mutation, and reproduction.

4. The method of claim 1 further comprising dividing the population of candidate solutions for the optimization problem into the plurality of sub-populations of candidate solutions and assigning the plurality of sub-populations to computing resources.

5. The method of claim 1, wherein said determining that the first generation corresponds to a later iteration of the iterative computing process than the second generation of the second sub-population of the plurality of sub-populations comprises:

accessing generation tracking data based on a first computing resource identifier and a second computing resource identifier,

wherein the generation tracking data indicates an iteration of the iterative computing process for each of a plurality of computing resources,

wherein the first computing resource identifier identifies a first of the plurality of computing resources and a second computing resource identifier identifies a second of the plurality of computing resources,

wherein the first computing resource is associated with the first sub-population and the second computing resource is associated with the second sub-population; and

comparing a first iteration indicated for the first computing resource and a second iteration indicated for the second computing resource identifier.

6. The method of claim 1 further comprising:

determining that the third generation of the first sub-population should receive a migrant candidate solution from a neighboring one of the plurality of sub-populations;

determining that the migrant candidate solution from the neighboring one of the plurality of sub-populations is not available;

creating a speculative migrant candidate solution; and

inserting the speculative migrant candidate solution into the third generation of the first sub-population of candidate solutions.

7. The method of claim 6 further comprising determining that a condition for speculatively creating a migrant candidate solution is satisfied before creating the speculative migrant candidate solution.

8. The method of claim 6 further comprising tagging the speculative migrant candidate solution to indicate speculative.