Visualising Evolution Competition

Submission Deadline Extended to 20th June 2013

Stop Press: GECCO Competition Prize & Special Issue Information

  • Prizes are available for the best three entries submitted to any of the GECCO competitions this year (see here for more details).
  • The winner of this competition will have the opportunity to submit to a special issue of the journal Genetic Programming and Evolvable Machines (see here for more details).

Goal

This competition is designed to enable participants to exhibit their cutting edge visualisations of evolutionary processes. The competition has two streams: the first involves visualising the evolution of a solution set to an optimisation problem using data generated during the running of an evolutionary algorithm; this data will be provided by the competition organisers. In the second stream, researchers can submit a visualisation of their own data, provided it falls within the remit of evolutionary computing. The goal of the competition is to find visualisations that enhance our understanding of evolutionary processes.

Rules

Entries to the Visualising Evolution Competition should comprise a program and relevant data. If the competition data is not used, competitors should submit data collected from running an evolutionary computing algorithm along with their visualisation program. The materials required for submission are as follows:

  1. The visualisation program and any instructions necessary to install and run it;
  2. Data for the application if the competition data is not used;
  3. A two-page writeup describing the visualisation program, the evolutionary process being visualised, and any insights gained through use of the visualisation.

Data

The competition data is available here. The data consists of the search history of a (1+1)-evolution strategy optimising a four-objective instance of the test problem DTLZ2. The data consists of parameter values and objective values for the parent and child at each generation, as well as information about which parameters were mutated. An additive Gaussian mutation was applied to each parameter with a probability of 1/P (P=13 parameters).

Scoring

The submissions will be scored by a community panel (TBA), which will judge based on the criteria of originality, quality, and relevance:

  • Originality: this criterion is primarily to show that a submission is unique and creative;
  • Quality: a submission scoring high on this criterion would be visually appealing, easy to interact with, and easy to interpret;
  • Relevance: this criterion evaluates the submission on its relevance and applicability to real world evolutionary problems.

Submission

All submissions should be sent to vizgec [at] exeter [dot] ac [dot] uk by June 20th 2013.