Detailed summary of Evolution Computing Lecture notes + additional information

🧑🏼‍🏫
Summary Slides
Chapter 0 - Evolutionary Problem Solving
The EC metaphor
Chapter 1 - Problems to be Solved
Black Box Model
Optimisation
Modelling
Simulation
Search Problems
Problem versus Problem-Solvers
Optimisation versus Constraint Satisfaction
NP Problems
Key Notions
Classes
Difference between Classes
Chapter 2 - The Origins
Motivation for EC
Darwinian Evolution
Survival of the Fittest
Diversity drives Change
Summary
Adaptive Landscape Metaphor
Genes and the Genome
Genetics
Homo Sapiens
Reproductive Cells
Crossing-over during Meiosis
Mutation
Development after Fertilisation
Genetic Code
Transcription, Translation
Important Points
Chapter 3: What is an EA?
Common Model of Evolutionary Processes
Two Pillars of Evolution
What is an EA?



Summary Slides 1

, Main EA Components
Representation
Evaluation / Fitness Function
Population
Selection
Survivor Selection
Variation Operators
Mutation
Recombination
Initialisation/Termination
Different Types of EAs
Solving the 8-queens problem
Problem
Representation
Fitness Evaluation
Mutation
Recombination
Parent Selection
Survivor Selection (Replacement)
EA Specification Tablea
Typical EA behaviour
Stages
Working of an EA demo
Progression of Fitness
Are long runs beneficial?
Is it worth expending effort on smart Initialisation?
EA in Context
EA as Problem Solvers
Goldberg View, 1989
EAs and Domain Knowledge
Michalewicz View, 1996
EC and Global Optimisation
EC and Neighbourhood Search
General Scheme of EAs
Version 1 versus Version 2
Important Points
Chapter 4: Representation, Mutation, Recombination
Role of Representation and Variation Operators
Two Sides of Representation
Representation (dis)continuity
Strong Causality Principle, Continuity, Locality
Good Representations



Summary Slides 2

, Binary Representation
Bit-Flip Mutation
1-point crossover
Alternative Crossover
n-point Crossover
Uniform Crossover
Crossover OR Mutation?
Integer Representation
Crossover
Mutation
Real-Valued or Floating-Point Representation
Mapping Real Values on Bit Strings
Uniform Mutation
Non-Uniform Mutation
Self-Adaptive Mutation
Uncorrelated Mutation with one σ 
Uncorrelated Mutation with n σ ’s
Correlated Mutations
Recombination
Single Arithmetic Crossover At Random
Single Arithmetic Crossover Mix Values
Whole Arithmetic Crossover
Blend Crossover
Overview different possible offspring
Multi-Parent Recombination
Multi-Parent Recombination: Type 1
Multi-Parent Recombination: Type 2
Permutation Representation
TSP Example
Representation
Mutation
Swap Mutation
Insert Mutation
Scramble Mutation
Inversion Mutation
Crossover
Order 1 Crossover
Partially Mapped Crossover (PMX)
Cycle Crossover
Edge Recombination
Tree Representation




Summary Slides 3

, Mutation
Recombination
Important Points
Chapter 5: Fitness, Selection, and Population Management
Population Management
Two different population management models exist:
Generation Gap
Fitness Based Competition
Fitness-Proportionate Selection (FPS)
FPS and Function Transposition
Parent Selection: Scaling
Windowing
Sigma Scaling
Rank-Based Selection
Linear Ranking
Exponential Ranking
Sampling Algorithms
Tournament Selection
Parent Selection: Uniform
Survivor Selection
Fitness-Based Survivor Selection
Elitism
Delete Worst
Round-Robin Tournament
(μ, λ)-selection (“comma strategy”)
(μ + λ)-selection (“plus strategy”)
Selection Pressure
Multimodality
Genetic Drift
Approaches for Preserving Diversity
Explicit Approaches for Preserving Diversity
Fitness Sharing
Fitness Sharing Example
Crowding
Crowding versus Fitness Sharing
Implicit Approaches for Preserving Diversity
Automatic Speciation
Island Model Parallel EAs
Cellular EAs
Different Spaces
Important Points




Summary Slides 4