04.4 Evolutionary & Swarm-Based AI (GA, ACO, AChem)

Genetic Algorithms

class of evolutionary search and optimization techniques inspired by Darwinian principles of natural selection

Exploration of search space

GAs are stochastic local search methods that balance (1)

Exploitation of promising regions

GAs are stochastic local search methods that balance (2)

Ant Colony Optimization

swarm intelligence algorithm modeled after the foraging behavior of ants

Artificial agents

traverse the problem space and deposit virtual pheromones on paths that represent partial or complete solution

Stigmergic communication

enables distributed and cooperative search for optimal or near-optimal solutions, particularly effective in combinatorial optimization tasks

Agents

ACO, ants searching paths

Pheromones

ACO, memory of good paths

Squad meta

ACO, shortest route discovered

Artificial Chemistry

a computational paradigm inspired by chemical reaction systems

Molecules

symbolic entities are treated as _, that interact according to predefined reaction rules, producing new entities or transforming existing ones

Decentralized computation

abstract chemistries serve as a metaphor for _ _, enabling emergent behavior and adaptive problem-solving in complex environment

Molecules

AChem, solutions

Reactions

AChem, interactions

Emergence

AChem, new stable compounds