Forced evolution in silico by artificial transposons and their genetic operators: The ant navigation problem

Area of Science:

• Computational intelligence
• Evolutionary computation
• Bio-inspired algorithms

Background:

• Evolutionary computation uses Darwinian principles for optimization.
• Biological macroevolution, particularly genomic parasites like transposons, offers inspiration for new algorithms.
• The co-evolution of populations with genomic parasites is theorized to enhance evolutionary search efficiency.

Purpose of the Study:

• To develop algorithms modeling genomic parasites, specifically artificial transposons.
• To integrate these artificial transposons into genetic algorithms (GA).
• To test their efficacy on the Artificial Ant Problem benchmark.

Main Methods:

• Introduced artificial transposons as distinct code fragments within the GA's host (artificial ant).
• Developed transposon-specific operators: mutation and reproduction (multiplication within host populations).
• Analyzed population dynamics of transposons during ant evolution simulations.

Main Results:

• Artificial transposons were observed to propagate and influence the selection of navigation program blocks.
• The presence of transposons significantly increased the speed of the evolutionary search.
• Transposons demonstrated adaptive behavior in response to the evolutionary problem.

Conclusions:

• Artificial transposons can effectively act as intelligent mutators in evolutionary algorithms.
• They adapt in response to evolutionary challenges during co-evolution with their hosts.
• This work represents a foundational step in simulating genomic parasites for enhanced computational evolution.