Distributed Interactive Genetic Algorithms / Distributed Evolutionary Computation

Share the joy
  •  
  •  
  •  
  •  
  •  
  •  
  •  
  •  
  •  

Distributed Interactive Genetic Algorithms are similar to Interactive Genetic Algorithms with additional features such as lack of a single point of failure, redundancy and potentially anonymous scalable human participation in the process.

Traditionally the Interactive Genetic Algorithm relies on the computer to be in the “innovator” role. The human being typically is in the “selector” role. Often the Evolutionary Computation  is done by a single human sitting in front of a computer who must evaluate the quality of the candidate solutions proposed by the innovator which in this case would be the computer.

An example of a typical Interactive Genetic Algorithm would be the human/computer interaction which generates through Evolutionary Computation the camouflage patterns on military outfits. Additional examples would be to use the Evolutionary Computation process to generate efficient designs for unmanned drones or computer chips.

One of the potential bottlenecks in using Interactive Genetic Algorithms is due to the human who may suffer from fatigue because the human has to ultimately judge the aesthetic quality of each candidate solution.

Distributed Interactive Genetic Algorithms / Distributed Evolutionary Computation would be similar to taking both the humans, the computer processing and putting them both in the cloud. In this descriptive example we could say that Ethereum has the ability to handle the distributed computation necessary and at the same time the humans could be given monetary incentive to participate in the selection process. In that example you would have the makings of a Distributed Interactive Genetic Algorithm running on the Ethereum platform. An example use case could be a distributed content based media retrieval system as has been already accomplished centralized using an Interactive Genetic Algorithm (Patel, Meshram, 2012).

References

Reynolds, C. (2010, July). Using interactive evolution to discover camouflage patterns. In ACM SIGGRAPH 2010 Posters (p. 113). ACM.
Patel, B. V., & Meshram, B. B. (2012). Content based video retrieval systems. arXiv preprint arXiv:1205.1641.

Leave a Reply

Your email address will not be published. Required fields are marked *