Tuning the ParEvo parameters

Three types of parameters can be varied:

  1. The parameters available to different participants
  2. The parameters as seen within the evolutionary algorithm
  3. Parameters affecting granularity

1. The parameters available to different participants

Parameters are attributes of the ParEvo process that can be varied. They are of two types

  1. Those participants can vary.
    • Which storyline they decide to add to, in any given iteration
    • The contents of the contributions they make
  2. Those the facilitator can vary
    • The number of participants
      • In a given exercise, and
      • Within each iteration of a given exercise
    • The types of participants
    • The text of the guidance that is given to participants at the start of each iteration
    • If feedback is given to participants on how well they are doing, compared to others, in having their participants build on their contributions
    • The maximum length of the text that participants can submit in each iteration
    • The duration of each iteration
    • The number of iterations
    • The evaluation criteria used at the evaluation stage

2. The parameters as seen within the evolutionary algorithm

Looking at the process as an evolutionary algorithm, changes can be made to each of the three steps in the algorithm:

  1. Variation:
    1. Who the facilitator allows to contribute to any iteration
    2. The content constraints imposed by the Facilitator on the contents of any contribution. These are visible in the text at the top of any Parevo exercise
    3. The maximum length of the contribution participants can make during an iteration
  2. Selection
    1. The number of contributions a participant can make within any iteration. The default is one per iteration.
  3. Retention
    1. The type of existing storylines that can be built on by participants within a given iteration. The default is only those storylines that were built on within the previous iteration.

3. Parameters affecting granularity

The granularity of the storylines will be affected by a combination of three parameters:

  1. The total time span that an exercise is expected to cover
  2. The number of iterations within that timespan
  3. The maximum size of allowable contributions within an iteration