Class AdaptiveEvolutionaryAlgorithm<T extends Copyable<T>>

java.lang.Object
org.cicirello.search.evo.AdaptiveEvolutionaryAlgorithm<T>
Type Parameters:
T - The type of object under optimization.
All Implemented Interfaces:
Splittable<TrackableSearch<T>>, Metaheuristic<T>, ReoptimizableMetaheuristic<T>, TrackableSearch<T>

public class AdaptiveEvolutionaryAlgorithm<T extends Copyable<T>> extends Object
This class implements an evolutionary algorithm with adaptive control parameters (i.e., crossover rates and mutation rates that evolve during the search). It follows a generational model, where a population of children are formed by applying genetic operators to members of the parent population, and the children replace the parents in the next generation. It uses the typical generational model using both crossover and mutation, such that each child may be the result of crossover alone, mutation alone, a combination of both crossover and mutation, or a simple copy of a parent.

Rather than specifying crossover and mutation rates, this adaptive evolutionary algorithm evolves these during the search. Each member of the population consists of an encoding of a candidate solution to the problem, along with a crossover rate Ci, a mutation rate Mi, and a parameter σi. During a generation, parents are paired at random. Consider that i and j are parents. One of these is chosen arbitrarily. For example, consider that i was chosen. With probability Ci the crossover operator is applied to the parents, and otherwise it is not. Then, the mutation operator is applied to each member of the population i with probability Mi. Note that this class implements an evolutionary algorithm for the general case, and not strictly bit strings, so the Mi is not a per-bit rate. Rather, it is the probability of a single application of whatever the mutation operator is.

After applying the genetic operators, all of the Ci and Mi are themselves mutated. Specifically, each is mutated with a Gaussian mutation with standard deviation σi. The σi are then also mutated by a Gaussian mutation with standard deviation of 0.01. The Ci and Mi are initialized randomly at the start such that they are each in the interval [0.1, 1.0], and the Gaussian mutation is implemented to ensure that they remain in that interval (e.g., reset to 0.1 if it is ever too low, and to 1.0 if it is ever too high). The σi are initialized randomly in the interval [0.05, 0.15], and constrained to the interval [0.01, 0.2].

This specific form of adaptive control parameters is based on the approach described in the following paper:

Vincent A. Cicirello. Genetic Algorithm Parameter Control: Application to Scheduling with Sequence-Dependent Setups. In Proceedings of the 9th International Conference on Bio-inspired Information and Communications Technologies, pages 136-143. December 2015. [PDF] [BIB] [From the ACM Digital Library]

The crossover, mutation, and selection operators are completely configurable by passing instances of classes that implement the CrossoverOperator, MutationOperator, and SelectionOperator classes to one of the constructors. The EA implemented by this class can also be configured to use elitism, if desired, such that a specified number of the best solutions in the population survive the generation unaltered.

  • Constructor Details

    • AdaptiveEvolutionaryAlgorithm

      public AdaptiveEvolutionaryAlgorithm(int n, MutationOperator<T> mutation, CrossoverOperator<T> crossover, Initializer<T> initializer, FitnessFunction.Double<T> f, SelectionOperator selection, int eliteCount, ProgressTracker<T> tracker)
      Constructs and initializes the evolutionary algorithm. This constructor supports fitness functions with fitnesses of type double, the FitnessFunction.Double interface.
      Parameters:
      n - The population size.
      mutation - The mutation operator.
      crossover - The crossover operator.
      initializer - An initializer for generating random initial population members.
      f - The fitness function.
      selection - The selection operator.
      eliteCount - The number of elite population members. Pass 0 for no elitism. eliteCount must be less than n.
      tracker - A ProgressTracker.
      Throws:
      IllegalArgumentException - if n is less than 1.
      IllegalArgumentException - if eliteCount is greater than or equal to n.
      NullPointerException - if any of mutation, crossover, initializer, f, selection, or tracker are null.
    • AdaptiveEvolutionaryAlgorithm

      public AdaptiveEvolutionaryAlgorithm(int n, MutationOperator<T> mutation, CrossoverOperator<T> crossover, Initializer<T> initializer, FitnessFunction.Integer<T> f, SelectionOperator selection, int eliteCount, ProgressTracker<T> tracker)
      Constructs and initializes the evolutionary algorithm. This constructor supports fitness functions with fitnesses of type int, the FitnessFunction.Integer interface.
      Parameters:
      n - The population size.
      mutation - The mutation operator.
      crossover - The crossover operator.
      initializer - An initializer for generating random initial population members.
      f - The fitness function.
      selection - The selection operator.
      eliteCount - The number of elite population members. Pass 0 for no elitism. eliteCount must be less than n.
      tracker - A ProgressTracker.
      Throws:
      IllegalArgumentException - if n is less than 1.
      IllegalArgumentException - if eliteCount is greater than or equal to n.
      NullPointerException - if any of mutation, crossover, initializer, f, selection, or tracker are null.
    • AdaptiveEvolutionaryAlgorithm

      public AdaptiveEvolutionaryAlgorithm(int n, MutationOperator<T> mutation, CrossoverOperator<T> crossover, Initializer<T> initializer, FitnessFunction.Double<T> f, SelectionOperator selection, ProgressTracker<T> tracker)
      Constructs and initializes the evolutionary algorithm. This constructor supports fitness functions with fitnesses of type double, the FitnessFunction.Double interface.
      Parameters:
      n - The population size.
      mutation - The mutation operator.
      crossover - The crossover operator.
      initializer - An initializer for generating random initial population members.
      f - The fitness function.
      selection - The selection operator.
      tracker - A ProgressTracker.
      Throws:
      IllegalArgumentException - if n is less than 1.
      NullPointerException - if any of mutation, crossover, initializer, f, selection, or tracker are null.
    • AdaptiveEvolutionaryAlgorithm

      public AdaptiveEvolutionaryAlgorithm(int n, MutationOperator<T> mutation, CrossoverOperator<T> crossover, Initializer<T> initializer, FitnessFunction.Integer<T> f, SelectionOperator selection, ProgressTracker<T> tracker)
      Constructs and initializes the evolutionary algorithm. This constructor supports fitness functions with fitnesses of type int, the FitnessFunction.Integer interface.
      Parameters:
      n - The population size.
      mutation - The mutation operator.
      crossover - The crossover operator.
      initializer - An initializer for generating random initial population members.
      f - The fitness function.
      selection - The selection operator.
      tracker - A ProgressTracker.
      Throws:
      IllegalArgumentException - if n is less than 1.
      NullPointerException - if any of mutation, crossover, initializer, f, selection, or tracker are null.
    • AdaptiveEvolutionaryAlgorithm

      public AdaptiveEvolutionaryAlgorithm(int n, MutationOperator<T> mutation, CrossoverOperator<T> crossover, Initializer<T> initializer, FitnessFunction.Double<T> f, SelectionOperator selection, int eliteCount)
      Constructs and initializes the evolutionary algorithm. This constructor supports fitness functions with fitnesses of type double, the FitnessFunction.Double interface.
      Parameters:
      n - The population size.
      mutation - The mutation operator.
      crossover - The crossover operator.
      initializer - An initializer for generating random initial population members.
      f - The fitness function.
      selection - The selection operator.
      eliteCount - The number of elite population members. Pass 0 for no elitism. eliteCount must be less than n.
      Throws:
      IllegalArgumentException - if n is less than 1.
      IllegalArgumentException - if eliteCount is greater than or equal to n.
      NullPointerException - if any of mutation, crossover, initializer, f, or selection are null.
    • AdaptiveEvolutionaryAlgorithm

      public AdaptiveEvolutionaryAlgorithm(int n, MutationOperator<T> mutation, CrossoverOperator<T> crossover, Initializer<T> initializer, FitnessFunction.Integer<T> f, SelectionOperator selection, int eliteCount)
      Constructs and initializes the evolutionary algorithm. This constructor supports fitness functions with fitnesses of type int, the FitnessFunction.Integer interface.
      Parameters:
      n - The population size.
      mutation - The mutation operator.
      crossover - The crossover operator.
      initializer - An initializer for generating random initial population members.
      f - The fitness function.
      selection - The selection operator.
      eliteCount - The number of elite population members. Pass 0 for no elitism. eliteCount must be less than n.
      Throws:
      IllegalArgumentException - if n is less than 1.
      IllegalArgumentException - if eliteCount is greater than or equal to n.
      NullPointerException - if any of mutation, crossover, initializer, f, or selection are null.
    • AdaptiveEvolutionaryAlgorithm

      public AdaptiveEvolutionaryAlgorithm(int n, MutationOperator<T> mutation, CrossoverOperator<T> crossover, Initializer<T> initializer, FitnessFunction.Double<T> f, SelectionOperator selection)
      Constructs and initializes the evolutionary algorithm. This constructor supports fitness functions with fitnesses of type double, the FitnessFunction.Double interface.
      Parameters:
      n - The population size.
      mutation - The mutation operator.
      crossover - The crossover operator.
      initializer - An initializer for generating random initial population members.
      f - The fitness function.
      selection - The selection operator.
      Throws:
      IllegalArgumentException - if n is less than 1.
      NullPointerException - if any of mutation, crossover, initializer, f, or selection are null.
    • AdaptiveEvolutionaryAlgorithm

      public AdaptiveEvolutionaryAlgorithm(int n, MutationOperator<T> mutation, CrossoverOperator<T> crossover, Initializer<T> initializer, FitnessFunction.Integer<T> f, SelectionOperator selection)
      Constructs and initializes the evolutionary algorithm. This constructor supports fitness functions with fitnesses of type int, the FitnessFunction.Integer interface.
      Parameters:
      n - The population size.
      mutation - The mutation operator.
      crossover - The crossover operator.
      initializer - An initializer for generating random initial population members.
      f - The fitness function.
      selection - The selection operator.
      Throws:
      IllegalArgumentException - if n is less than 1.
      NullPointerException - if any of mutation, crossover, initializer, f, or selection are null.
  • Method Details

    • split

      Description copied from interface: Splittable
      Generates a functionally identical copy of this object, for use in multithreaded implementations of search algorithms. The state of the object that is returned may or may not be identical to that of the original. Thus, this is a distinct concept from the functionality of the Copyable interface. Classes that implement this interface must ensure that the object returned performs the same functionality, and that it does not share any state data that would be either unsafe or inefficient for concurrent access by multiple threads. The split method is allowed to simply return the this reference, provided that it is both safe and efficient for multiple threads to share a single copy of the Splittable object. The intention is to provide a multithreaded search with the capability to provide spawned threads with their own distinct search operators. Such multithreaded algorithms can call the split method for each thread it spawns to generate a functionally identical copy of the operator, but with independent state.
      Specified by:
      split in interface Metaheuristic<T extends Copyable<T>>
      Specified by:
      split in interface ReoptimizableMetaheuristic<T extends Copyable<T>>
      Specified by:
      split in interface Splittable<T extends Copyable<T>>
      Returns:
      A functionally identical copy of the object, or a reference to this if it is both safe and efficient for multiple threads to share a single instance of this Splittable object.
    • optimize

      public final SolutionCostPair<T> optimize(int numGenerations)
      Runs the evolutionary algorithm beginning from a randomly generated population. If this method is called multiple times, each call begins at a new randomly generated population.
      Specified by:
      optimize in interface Metaheuristic<T extends Copyable<T>>
      Parameters:
      numGenerations - The number of generations to run.
      Returns:
      The best solution found during this set of generations, which may or may not be the same as the solution contained in the ProgressTracker, which contains the best across all calls to optimize as well as ReoptimizableMetaheuristic.reoptimize(int). Returns null if the run did not execute, such as if the ProgressTracker already contains the theoretical best solution.
    • reoptimize

      public final SolutionCostPair<T> reoptimize(int numGenerations)
      Runs the evolutionary algorithm continuing from the final population from the most recent call to either Metaheuristic.optimize(int) or ReoptimizableMetaheuristic.reoptimize(int), or from a random population if this is the first call to either method.
      Specified by:
      reoptimize in interface ReoptimizableMetaheuristic<T extends Copyable<T>>
      Parameters:
      numGenerations - The number of generations to run.
      Returns:
      The best solution found during this set of generations, which may or may not be the same as the solution contained in the ProgressTracker, which contains the best across all calls to optimize as well as Metaheuristic.optimize(int). Returns null if the run did not execute, such as if the ProgressTracker already contains the theoretical best solution.
    • getProgressTracker

      public final ProgressTracker<T> getProgressTracker()
      Description copied from interface: TrackableSearch
      Gets the ProgressTracker object that is in use for tracking search progress. The object returned by this method contains the best solution found during the search (including across multiple concurrent runs if the search is multithreaded, or across multiple restarts if the run methods were called multiple times), as well as cost of that solution, among other information. See the ProgressTracker documentation for more information about the search data tracked by this object.
      Specified by:
      getProgressTracker in interface TrackableSearch<T extends Copyable<T>>
      Returns:
      the ProgressTracker in use by this metaheuristic.
    • setProgressTracker

      public final void setProgressTracker(ProgressTracker<T> tracker)
      Description copied from interface: TrackableSearch
      Sets the ProgressTracker object that is in use for tracking search progress. Any previously set ProgressTracker is replaced by this one.
      Specified by:
      setProgressTracker in interface TrackableSearch<T extends Copyable<T>>
      Parameters:
      tracker - The new ProgressTracker to set. The tracker must not be null. This method does nothing if tracker is null.
    • getProblem

      public final Problem<T> getProblem()
      Description copied from interface: TrackableSearch
      Gets a reference to the problem that this search is solving.
      Specified by:
      getProblem in interface TrackableSearch<T extends Copyable<T>>
      Returns:
      a reference to the problem.
    • getTotalRunLength

      public long getTotalRunLength()
      Gets the total run length in number of fitness evaluations. This is the total run length across all calls to Metaheuristic.optimize(int) and ReoptimizableMetaheuristic.reoptimize(int). This may differ from what may be expected based on run lengths. For example, the search terminates if it finds the theoretical best solution, and also immediately returns if a prior call found the theoretical best. In such cases, the total run length may be less than the requested run length.
      Specified by:
      getTotalRunLength in interface TrackableSearch<T extends Copyable<T>>
      Returns:
      The total number of generations completed across all calls to Metaheuristic.optimize(int) and ReoptimizableMetaheuristic.reoptimize(int).