Bird Migration Visualization
Evolutionary Algorithms in birds’ adaptation to migration
  • Darwin's Theory of Evolution- Individuals with traits that enable them to adapt to their environments will
    • survive
    • have more offspring who inherit those traits
  • Evolutionary Algorithms
    • create optimization procedures to solve problems
    • require a large amount of trials to gather, select and mutate the produced data to achieve the desired goals
    • have four elements: Population, source of variation, reproductive fitness, selection
    • Components :
      Genetic Algorithms– invented by John Holland in the 1960’s
      Evolution Strategies– invented by Ingo Rechenberg in the 1960's
  • Steps :
    1. Initialize a random population
      Each bird is defined simply by three variables: starting position, flying direction and speed
    2. Calculate the fitness of the population
      In the game it is measured by if the birds reach the migration destination
    3. Repeat until you reach the goal
      In the game it is the bird's survival rate.

      While stopping criterion is not satisfied, the following steps is processed :

      • 3.1. Select parents
        Select the data from the birds have high fitness score
      • 3.2. Perform the crossover to produce offsprings
        Try different combinations of the variables’ values from the selected data
      • 3.3. Perform mutation to increase the variety of the population
        Change certain values in the offspring. Each offspring inherits traits from the parents, which are starting position, flying direction and speed. Randomly pick one of the traits and replace it with an arbitrary value.
      • 3.4. Calculate the survival rate
        Set the next input to be : crossover number counts + mutation number counts + clone counts (Evolutionary Algorithms allow some children to be exact copy of parents)