- Introduction to EAs
- walking across search space (exploration, exploitation)
- general characteristics of EAs
- Brief history of EAs
- old-day trials
- evolution strategies
- evolutionary programming
- genetic algorithms
- General schema of EAs
- Basic building blocks of EAs
- selection
- sampling methods
- fitness proportional sampling (SSwR, SSwoR, RSSwR,
RSSwoR, DS, RSIS, SUS)
- tournaments
- random sampling (uniform and non-uniform
probability distributions)
- truncating
- fitness remapping
- scaling (windowing, sigma, linear)
- ranking (linear, exponential)
- replacement
- representations (binary, n-ary, float, matrices, trees)
- basic genetic operators
- mutation (bit flipping, boundary, Muehlenbein,
non-uniform)
- crossover (1-point, 2-point, n-point, uniform,
discrete, intermediate, linear)
- Typical (canonical) forms of particular EAs
- genetic algorithm
- evolution strategy
- evolutionary programming
- Why EA work?
- Schema theorem
- Building block hypothesis
- Deception and EAs
- minimal deceptive problem
- coding (Gray code)
- Diversity maintaining
- (genetic load, hyper-mutation, restricted competition,
sharing)
- Special types of EAs
- niching EA
- parallel (crowding, sharing)
- serial
- multi-objective EAs
- aggregation methods
- VEGA
- ranking (feature based, Pareto based)
- tracking changing environment
- hyper-mutation
- redundant coding
- diploid schemes
- polygenic schemes
- constrained search space
- decoders
- repair procedures
- penalty functions
- feasibility preserving operators
- Optimal parameter setting
- No free lunch theorem
- methods for parameter setting
- manual
- meta-evolution
- self-adaptation
- Co-evolution
- (parasitic, niche, constrained search, multi-objective examples)
- Hybridisation with other methods
- EAs + hill-climbing (Lamarck, Baldwin)
- EAs + Fuzzy logic
- Parallel versions of EAs
- global
- coarse grained (migration operator)
- fine grained
- Applications of EAs
- travelling salesman problem
- machine learning
|
