Genetic Algorithms: Concepts and Applications

Study Notes

A genetic algorithm (GA) is a computer program that tries to find a good solution to a problem by evolving a population of candidate solutions.
The GA maintains a population of solutions and makes it evolve by iteratively applying a set of stochastic operators.
The GA has been subject of many (early) studies and still often used as a benchmark for novel GAs.
The GA shows many shortcomings, e.g. it is too restrictive in its representation of solutions, mutation and crossovers only applicable for bit-string and integer representations, and selection mechanism sensitive for converging populations with close fitness values.
We will use Tournament Selection to choose a solution and place it in the mating pool. Two other solutions will be picked and another solution in the mating pool will be filled up with the better solution.
Order-1 crossover is explorative and mutation is exploitative.
Mating individualsto generate pop_size offspring is a steady-state replacement process.
Each individualsurvives for exactly one generation and the entire set of parents is replaced by the offspring.
The elitism option is used to keep one or more of the best solutions discovered so far and copy them to the next generation.
Schema theory seeks to give a theoretical justification for the efficacy of the field of genetic algorithms.
What is a schema:
a template for new gene arrangements  {0,1,*} where * is a don't care.
Schema is favorable traits in a solution, where a favorable schema is called an above average schema.