Mars Lander – Genetic Algorithm Simulator

Quick Start

In the viewer: green line = landing zone · red lines = terrain

Initialise
Create a population of random landing command sequences — the algorithm's first guesses.
Evaluate
Simulate each lander and score it: how close to a safe landing did it get?
Select
Pick the highest-scoring landers as parents — better landers have a better chance.
Crossover
Combine parts of two parents' command sequences so offspring inherit the best of each.
Mutate
Randomly tweak a few commands to keep diversity and avoid getting stuck.

Repeat until a lander lands safely — or the generation limit is reached.

Terrain

Each terrain tests the algorithm's adaptability in different ways.

Population size

More landers = broader exploration per generation.

Generations

More generations = more improvement opportunities.

Selection mode

Tournament is balanced; Roulette rewards top performers more aggressively.

Elite rate — 20%

The best landers survive unchanged. Too high: diversity drops. Too low: good solutions are lost.

Crossover rate — 90%

High: offspring blend both parents' strategies. Low: offspring resemble one parent closely.

Mutation rate — 10%

Low: stable but may stagnate. High: explores widely but can disrupt good solutions.

Select a terrain and click Run Algorithm to start.

The algorithm will evolve a population of landers across generations until one lands safely.

Generation — / —