Using real data sets to simulate evolution within complex environments<br />1<br />Bruce Edmonds, Centre for Policy Modell...
Upcoming SlideShare
Loading in …5
×

Using Data Sets to Simulate Evolution within Complex Environments (Poster)

417 views

Published on

An agent-based model of evolution and adaption is presented with complex genomes, sexual and asexual reproduction. However, unlike other models, this has a purposefully complex environment. Individuals are spread over a 2D space upon which is projected a complex data set derived from observations of the real world. Energy extraction for the purposes of life and reproduction is achieved by the match between the genome and the data in its locality. In this model different genomes develop to exploit different parts of the space, with some genomes acquiring a greater generality than others. This model can be used to explore the trade-offs inherent between diversity, specialization and inter-breeding among individuals.

Published in: Education, Technology
0 Comments
0 Likes
Statistics
Notes
  • Be the first to comment

  • Be the first to like this

No Downloads
Views
Total views
417
On SlideShare
0
From Embeds
0
Number of Embeds
1
Actions
Shares
0
Downloads
3
Comments
0
Likes
0
Embeds 0
No embeds

No notes for slide

Using Data Sets to Simulate Evolution within Complex Environments (Poster)

  1. 1. Using real data sets to simulate evolution within complex environments<br />1<br />Bruce Edmonds, Centre for Policy Modelling, Manchester Metropolitan University<br />3.7<br />1.1<br />ca<br />Target Question: Does the complexity of the environment significantly affect evolutionary processes?<br /><ul><li>(here “complexity” means that there are exploitable patterns in the environment but these are difficult to discover)
  2. 2. Adding randomness to the environment (or the fitness function) is not satisfactory and the NK model of fitness adjusts the difficulty in a uniform manner (second order uniformity)</li></ul>sex<br />slope<br />restecg+fbs+1<br />0.8<br />fbs/oldpeak<br />Main Ideas:<br /><ul><li> Evolutionary data-mining is where ideas from biological evolution are applied to data-mining – finding patterns in data
  3. 3. Data sets exist for the purpose of testing different ML algorithms that have patterns in them, albeit difficult to discover
  4. 4. Reversing this... I am suggesting the use of complex data sets as a test bed to investigate how the complexity of the environment might affect evolution</li></ul>The Data Set Environment:<br /><ul><li>Find a rich data set (preferably one derived from a naturally complex system) with many independent variables
  5. 5. The gene of an individual is an arbitrary arithmetic expression stored as a tree (or similar technique)
  6. 6. Resource in the model is modelled by distributing to individuals predicting the outcome variable of local data better than its competitors
  7. 7. The gene are mutated and crossed as the simulation progresses
  8. 8. Individuals are selected for/against depending on their total success in predicting </li></ul>(1) Evaluation Phase of Model:<br />(2) Selection Phase of Model:<br />Data Space<br />Data points are spread over the space according to two of their component values<br />Data Space<br />15.5<br />12.5<br />N times:<br />probabilistically select winner(s)on fitness<br />probabilistically select a loser on lack of fitness<br />kill loser<br />Either<br />propagate winner locally with possible mutation<br />mate with another local based on fitness<br />For each data point (or a random subset of them) evaluate (a random selection of) near individuals to determine the share of fitness each receive (depending on predictive success)<br />Sum of fitness determines which breed and die<br />3.2<br />17.6<br />23.7<br />Individuals each with genes composed of an arithmetic expression to predict HD based on the other 13 variables<br />Start of Simulation (HD Data)<br />12.3<br />8.6<br />Data points from set distributed over space dependent on 2 variables<br />9.0<br />8.1<br />chol (x) &thalach (y)<br />An example simulation run using the HD data set:<br />Individuals each with gene which is an arithmetic expression, e.g.:<br />Simulation – 25 ticks<br />Simulation – 50 ticks<br />Simulation – 300 ticks (100 w/o variation)<br />Illustrative Results:<br />Using Reduced Cleveland Heart Disease Data Set, 20 runs with each setting, 1000 individuals, 1000 iterations,Locality parameter 0.1 (radius), Comparison of OriginalvsErsatz Data Sets (same distribution in each dimension but random), Fixed normal noise (0, 0.1) added to both data sets<br />Fitness<br />Gene Complexity/Depth<br />Original <br />Ersatz <br />Whenmighty the complexity of the environment effect evolutionary processes?<br />How might the complexity of the environment effect evolutionary processes?<br />Will models with a simple environment tell us about evolution in the wild?<br /><ul><li>When and about what aspects will models with simple environments be sufficient?
  9. 9. In what ways might evolution differ when in complex environments?</li></ul>What kind of complexity might we need?<br />How might one measure this complexity in the wild (if this is even possible)?<br />Questions:<br />That might be explored using this approach<br />

×