Evolving Open Complexity

TL;DR

This paper analyzes the information dynamics of large evolved programs, revealing that many genetic changes are silent and proposing an open architecture to facilitate effective evolution in complex programs.

Contribution

It introduces the concept of evolving large, complex programs with an open architecture to improve mutation impact measurement and evolutionary progress.

Findings

Large programs exhibit silent genetic changes due to entropy loss.

Most crossover sites are distant from the root in large trees.

Open architectures with mutation sites near the environment enhance evolution.

Abstract

Information theoretic analysis of large evolved programs produced by running genetic programming for up to a million generations has shown even functions as smooth and well behaved as floating point addition and multiplication loose entropy and consequently are robust and fail to propagate disruption to their outputs. This means, while dependent upon fitness tests, many genetic changes deep within trees are silent. For evolution to proceed at reasonable rate it must be possible to measure the impact of most code changes, yet in large trees most crossover sites are distant from the root node. We suggest to evolve very large very complex programs, it will be necessary to adopt an open architecture where most mutation sites are within 10 to 100 levels of the organism's environment.