amlalejini/aagos

Presented as a late-breaking poster at ALife 2018. Variable-length NK landscape used for investigating the effect of environmental change on gene length even in the presence of high mutation rates.

Aagos

• Auto-Adaptive Genetic Organization System

Motivation

• In static systems, high mutation rates create a pressure for gene overlap and smaller genomes
• However, more modular, separated genetic architectures have been shown to be more evolvable in previous work
• Furthermore, in evolutionary biology, we see a general trend of larger essential genomes in more complex organisms
• Therefore, what environmental pressures have selected for modular, evolvable genetic architectures in biological systems?

System Model

• NK-Inspired landscape
• Variable length genomes

Results

• High mutation rates in a static landscape do indeed lead to smaller minimal genomes
• Moderate environmental change leads to larger minimal genomes, however high environmental change leads to meltdown
• Even with a high mutation rate, in the presence of moderate environmental change, minimal genome size is significantly larger than in static environments

Experiment Replication

To run these experiments for yourself, simply download the Empirical library and this repository and enter the commands:

make

./Aagos -[parameters]

Parameters we used:

• GENE_MOVE_PROB = 0.003
• BIT_FLIP_PROB = 0, .00001, .0001, .001, .003, .01, .03, .1
• BIT_INS_PROB = 0.001
• BIT_DEL_PROB = 0.001
• CHANGE_RATE = 0, 1, 2, 5, 10, 20, 50, 100, 200, 500, 1000, 5000

Changeable Parameters

Environmental Parameters

• CHANGE_RATE, default 0, How many changes to fitness tables each generation
• POP_SIZE, default 1000, How many organisms should be in the population
• MAX_GENS, default 50000, How many generations should the runs go for
• SEED, default 0, Random number seed (0 for based on time)
• ELITE_COUNT, default 0, How many organisms should be selected via elite selection
• TOURNAMENT_SIZE, default 2, How many organisms should be chosen for each tournament

Genomic Structure

• NUM_BITS, default 128, Starting number of bits in each organism
• NUM_GENES, default 16, Number of genes in each organism
• GENE_SIZE, default 8, Size of each gene in each organism
• MAX_SIZE, default 1024, maxiumum size of a genome
• MIN_SIZE, default 8, minimum size of a genome

Mutations

• GENE_MOVE_PROB, default 0.01, Probability of each gene moving each generation
• BIT_FLIP_PROB, default 0.01, Probability of each bit toggling
• BIT_INS_PROB, default 0.01, Probability of a single bit being inserted
• BIT_DEL_PROB, default 0.01, Probability of a single bit being removed

Output

• PRINT_INTERVAL, default 1000, How many updates between prints?