Class IloEAOperatorFactory

This factory provides a set of parameters as well as functions that build operators based on these parameters.

Assuming that you have already created:

• A solution prototype: (prototype).
• An array of decision variables (vars).

Then the following code creates two operators which operate on vars.

 IloEAOperatorFactory factory(env, vars);
 factory.setPrototype(prototype);
 factory.setSearchLimit(IloFailLimit(env, 100));
 IloPoolProc mut = factory.mutate(1.0 / vars.getSize(), "mutate");
 IloPoolProc xo  = factory.uniformXover(0.5, "uXover");
 

This constructor creates a handle object (an instance of the class IloEAOperatorFactory) from a pointer to an implementation object (an instance of the class IloEAOperatorFactoryI).

This constructor creates an operator factory associated with the environment env which will produce operators on the array of integer variables scope.

This constructor creates an array operator factory associated with the environment env which will produce operators on the array of Boolean variables scope.

This member function returns the implementation object of the invoking handle.

This member function returns the probability, for each array cell, of being left unbound by created operators. It returns the probability set at the previous call to IloEAOperatorFactory::setLeaveFactor or to 0.0 if no probability was set.

This operator selects one parent and produces a solution by inheriting its values. Each inherited value is mutated with a given mutation probability mutationProb. name, if provided becomes the name of the newly created operator.

This recombination operator selects two parents and produces a new solution by inheriting values from the second parent for array indices greater than a randomly chosen cut point. Other values are inherited from the first parent. name, if provided becomes the name of the newly created operator.

This operator instantiates variables by choosing random values in their domain. name, if provided becomes the name of the newly created operator.

This operator selects one parent and produces a new solution by leaving some variables unbound. The probability of leaving a variable unbound is given by the leaveFactor parameter of this factory. name, if provided becomes the name of the newly created operator.

This operator selects one parent and produces a new solution by copying the parent values with the exception of randomly chosen value sequences. The given probability breakProbability specifies, for each array site, the likelihood of starting (or ending) a relaxed sequence. name, if provided becomes the name of the newly created operator.

This member function specifies the probability leaveFactor, for each cell, of being left unbound by created operators.

This swap operator selects a parent solution, chooses two random locations of the accessed array, and attempts to exchange their values. Upon failure, it backtracks to choose different exchange locations and fails when all exchange points have been tried. name, if provided becomes the name of the newly created operator.

This translocation operator selects a parent solution and then chooses a random interval of the accessed array and a random insertion point. It then attempts to create a solution in which the sequence delimited by the chosen interval will be moved to the insertion point. Upon failure, it backtracks to choose a different insertion point. When this fails, it backtracks to change the interval to try to move. It fails when all intervals and insertion points have been tried. name, if provided becomes the name of the newly created operator.

This transposition operator selects a parent solution and chooses a random interval of the accessed array. It then attempts to create a solution in which the sequence delimited by the chosen interval will be reversed. Upon failure, it backtracks to choose a different interval and fails when all intervals have been tried. name, if provided becomes the name of the newly created operator.

This recombination operator selects two parents and produces a solution by inheriting values from the second parent for array indices contained between two randomly chosen cut points. Other values are inherited from the first parent. Upon failure, this goal backtracks by trying other cut points. name, if provided becomes the name of the newly created operator.

This recombination operator selects two parents and produces a new solution by inheriting values from the first or second parent based on a crossover rate of 0.5. This goal does not backtrack. name, if provided becomes the name of the newly created operator.

This recombination operator selects two parents and produces a new solution by inheriting values from the first or second parent based on the given crossover rate. This goal does not backtrack. The parameter xoverRate sets the probability of each feature being exchanged. name, if provided becomes the name of the newly created operator.