Solutions that evolve.
NovaGenetica is a curated, growing home for everything worth knowing about Genetic Algorithms and Genetic Programming — plain-language primers, the foundational books and papers, the software you can actually run, and real projects that put evolution to work. Whether you're new to the idea or going deep, start here.
Pick a path through the field.
Evolutionary computation borrows one idea from biology — selection, variation, inheritance — and turns it into a general-purpose way to search for good solutions. Here's the map.
Genetic Algorithms
Evolve a population of candidate solutions encoded as "chromosomes." Selection, crossover, and mutation push each generation toward better answers. The workhorse of evolutionary optimization.
Read the primer →Genetic Programming
Don't just tune numbers — evolve whole programs. GP represents candidate solutions as trees of code that grow, recombine, and mutate until they compute what you need. Evolution that writes its own logic.
Read the primer →The Library
A sourced, growing collection: the foundational books, survey papers and tutorials, software libraries you can install today, courses, and communities. Each entry says what it is and why it's worth your time.
Browse resources →Robot Project: SEED-Nav
A real autonomous robot that navigates using a genetic algorithm — evolving its own navigation parameters from experience, with memory stored as virtual DNA. See how the theory becomes a moving machine.
See the project →One idea, an enormous range of uses.
You don't need a perfect model of a problem to evolve a good solution to it — you just need a way to score candidates and a way to vary them. That's why evolutionary methods turn up everywhere.
Optimization
Scheduling, routing, layout, parameter tuning, and design problems where the search space is huge and the math is ugly. GAs find strong answers without needing gradients or a tidy formula.
Machine learning
Evolving features, neural-network weights and structures (neuroevolution), rule sets, and symbolic models that stay readable — a complement to gradient-based learning.
Robotics & control
Evolving controllers, gaits, and navigation policies on real and simulated hardware — including the SEED-Nav robot featured on this site.
How evolution searches.
Every genetic algorithm and genetic program runs the same basic cycle. Once you see it, you see it everywhere.
-
Represent. Encode each candidate solution as a genome — a string of numbers, a bit vector, or a tree of program code.
-
Evaluate. Score every candidate with a fitness function: how good is this solution at the job you care about?
-
Select. Give fitter candidates a better chance of becoming parents — the pressure that drives improvement.
-
Vary. Recombine parents (crossover) and introduce small random changes (mutation) to produce a new generation — then repeat until it's good enough.
Built something with evolution? Share it.
NovaGenetica is a community resource. If you've written a paper, built a library, taught a course, or evolved a project in genetic algorithms or genetic programming, we'd love to feature it in the library — with full credit and a link back to you. Beginners welcome.
Share your workCurated by Darin R. Molnar, PhD
NovaGenetica is maintained by Darin R. Molnar, PhD — a technologist and researcher who builds with evolutionary methods, including the SEED-Nav navigation robot featured here. Consulting work runs through Shoreline Consulting.
More about Darin & Shoreline