liveBook · Manning

Tournament selection

As explained in Optimization Algorithms: AI techniques for design, planning, and control problems

Tournament selection is a survivor-selection method that randomly groups individuals and retains the best individual from each group.

Tournament selection chooses individuals by randomly forming groups of size k and retaining the highest-fitness individual from each group. It turns local competition into a practical way to select candidates for reproduction or survival. section 7.3.3 section 7.3.5

A tournament-selection pass follows these steps:

  • Randomly choose a group of k individuals from the current population. section 7.3.3
  • Compare group members using their fitness values and identify the best performer. section 7.3.3
  • Repeat the procedure m times when m individuals are needed for reproduction. section 7.3.3
  • Apply genetic operators such as crossover and mutation to the selected winners. section 7.3.3

For survivor selection, tournament groups can draw from both the old population and newly generated offspring. Retaining the best member of each group can preserve diversity more effectively than simply keeping the best existing individuals. section 7.3.5

Ticket-pricing candidates

Use the ranked ticket-pricing candidates as a population. Randomly place several candidates into a tournament, compare their objective-function values, and advance the candidate with the highest value; repeat until enough winners are available for reproduction. section 7.3.3

Figure 7.12 Tournament selection

Use the tournament-selection figure to visualize random grouping, within-group competition, and winner advancement. section 7.3.3

How Optimization Algorithms: AI techniques for design, planning, and control problems covers it

How deep it goes
Explains grouping, fitness comparison, repetition, and winner use, but does not develop implementation code or parameter heuristics.
Examples
Uses ticket-pricing candidates and a tournament-selection figure to illustrate ranking, competition, and winner advancement.
What it assumes
Uses populations, fitness values, genetic operators, and offspring without developing those concepts from first principles.
What you can do after
Apply tournament selection to choose genetic-algorithm reproduction candidates or survivors.
How it connects
Connects genetic algorithms with fitness, reproduction, crossover, mutation, survivor selection, elitist selection, and population diversity.

Also explained in

Compare how 4 books explain tournament selection →

Related concepts

This explainer was generated by Manning's knowledge graph from the text of Optimization Algorithms: AI techniques for design, planning, and control problems. The section links point to the parts of the book it draws on.