Tutorial

Although this tutorial doesn’t make reference directly to the complete API of the framework, we think it is the place to start to understand the principles of DEAP. The core of the architecture is based on the creator and the Toolbox. Their usage to create types is shown in the first part of this tutorial. Then, a next step is taken in the direction of building generic algorithms by showing how to use the different tools present in the framework. Finally, we conclude on how those algorithms can be made parallel with minimal changes to the overall code (generally, adding a single line of code will do the job).

• 1. Where to Start?
  • 1.1. Types
  • 1.2. Initialization
  • 1.3. Operators
  • 1.4. Algorithms
• 2. Creating Types
  • 2.1. Fitness
  • 2.2. Individual
  • 2.3. Population
• 3. Next Step Toward Evolution
  • 3.1. A First Individual
  • 3.2. Evaluation
  • 3.3. Mutation
  • 3.4. Crossover
  • 3.5. Selection
  • 3.6. Using the Toolbox
  • 3.7. Variations
  • 3.8. Algorithms
• 4. Genetic Programming
  • 4.1. Loosely Typed GP
  • 4.2. Strongly Typed GP
  • 4.3. Ephemeral Constants
  • 4.4. Generation of Tree Individuals
  • 4.5. Evaluation of Trees
  • 4.6. Tree Size Limit and Bloat Control
  • 4.7. How to Evolve Programs
• 5. Speed up the evolution
  • 5.1. The comparison problem : Sorting Networks
  • 5.2. Normal run
  • 5.3. Using PyPy
  • 5.4. Using the C++ version of the NSGA II algorithm
  • 5.5. Using an home-made C++ evaluation function
  • 5.6. Combining C++ version of NSGA II and evaluation function
  • 5.7. Speedups summary
  • 5.8. Distribution
• 6. Using Multiple Processors
  • 6.1. Scalable Concurent Operations in Python (SCOOP)
  • 6.2. Distributed Task Manager
  • 6.3. Multiprocessing Module
• 7. Benchmarking Against the Best (BBOB)
  • 7.1. Preparing the Algorithm
  • 7.2. Running the Benchmark