Genetic Algorithms

Biological evolution principles and Darwinian natural selection, Genetic algorithms as population-based stochastic optimization, Search space representation and fitness landscapes, Schema theorem and building block hypothesis, Convergence properties and implicit parallelism, Comparison with gradient-based optimization methods, No Free Lunch theorem implications for EAs.

Encoding schemes (binary, real-valued, permutation, tree-based), Fitness function design and scaling (rank, linear, exponential), Selection methods (roulette wheel, tournament, rank, elitism), Crossover operators (single-point, uniform, arithmetic, blend), Mutation strategies (bit-flip, Gaussian, swap, scramble), Reproduction and generational replacement policies.

Adaptive parameter control (self-adaptive mutation rates, dynamic crossover probabilities), Niching methods (fitness sharing, crowding, island models), Multi-objective optimization (Pareto dominance, NSGA-II, SPEA2), Constraint handling (penalty functions, repair algorithms, decoder-based), Hybridization with local search (memetic algorithms), Parallel GA architectures.

Tree-based representation for GP, Primitive sets and strongly-typed GP, Crossover and mutation for tree structures (subtree exchange, hoist), Bloat control mechanisms (Pareto front approximation, operator equal fitness), Automatic programming and symbolic regression, Koza's GP parameters and ramped half-and-half initialization, Multi-expression programming (MEP).

Combinatorial optimization (traveling salesman problem, scheduling, knapsack), Function optimization (Rastrigin, Ackley, multimodal deceptive functions), Machine learning applications (feature selection, neural architecture search, hyperparameter optimization), Engineering design optimization (truss structures, aerodynamic shapes), Parameter tuning for deep learning and robotics controllers.