Documentation of 'ec.rule.RuleSetConstraints' Java class
RuleSetConstraints
Skip navigation links
ec.rule

Class RuleSetConstraints

  • All Implemented Interfaces:
    Clique, Setup, java.io.Serializable

    public class RuleSetConstraints
extends java.lang.Object
implements Clique
    RuleSetConstraints is an basic class for constraints applicable to rulesets. There are two categories of parameters associated with this class. First, there are parameters which guide the initial number of rules to be created when a ruleset is initialized for the first time, or totally reset. Second, there are parameters which indicate how rulesets are to be mutated under the "default" rule mutation operator.

    First the initialization parameters. You need to specify a distribution from which the system will pick random integer values X. When a ruleset is to be initialized, a random value X is picked from this distribution, and the ruleset will be created with X initial rules. You can specify the distribution in one of two ways. First, you can specify a minimum and maximum number of rules; the system will then pick an X uniformly from between the min and the max. Second, you can specify a full distribution of size probabilities for more control. For example, to specify that the system should make individuals with 0 rules 0.1 of the time, 1 rule 0.2 of the time, and 2 rules 0.7 of the time, you set reset-num-sizes to 3 (for rule sizes up to but not including 3), and then set reset-size.0 to 0.1, reset-size.1 to 0.2, and reset-size.2 to 0.7.

    Next the mutation parameters. The default mutation procedure works as follows. First, every rule in the ruleset is mutated. It is up to the rule to determine by how much or whether it will be mutated (perhaps by flipping a coin) when its mutate function is called. Second, the system repeatedly flips a coin with "p-del" probability of being true, until it comes up false. The number of times it came up true is the number of rules to remove from the ruleset; rules to be removed are chosen at random. Third, the system repeatedly flips a coin with "p-add" probability of being true, until it comes up false. The number of times it came up true is the number of new randomly-generated rules to add to the ruleset; rules are added to the end of the array. Fourth, with "p-rand-order" probability, the order of rules in the ruleset array is randomized; this last item might or might not matter to you depending on whether or not your rule interpreter differs depending on rule order.

    See Also:
    Serialized Form

    • Field Detail

      • minSize

        public int minSize

      • maxSize

        public int maxSize

      • resetMinSize

        public int resetMinSize

      • resetMaxSize

        public int resetMaxSize

      • sizeDistribution

        public double[] sizeDistribution

      • p_add

        public double p_add

      • p_del

        public double p_del

      • P_RAND_ORDER_PROB

        public static final java.lang.String P_RAND_ORDER_PROB
        See Also:
        Constant Field Values

      • p_randorder

        public double p_randorder

      • rulePrototype

        public Rule rulePrototype
        The prototype of the Rule that will be used in the RuleSet (the RuleSet contains only rules with the specified prototype).

      • constraintNumber

        public byte constraintNumber
        The byte value of the constraints -- we can only have 256 of them

      • name

        public java.lang.String name
        The name of the RuleSetConstraints object

    • Constructor Detail

      • RuleSetConstraints

        public RuleSetConstraints()

    • Method Detail

      • pickSize

        public int pickSize(EvolutionState state,
                    int thread)
        Assuming that either resetMinSize and resetMaxSize, or sizeDistribution, is defined, picks a random size from resetMinSize...resetMaxSize inclusive, or randomly from sizeDistribution.

      • numRulesForReset

        public int numRulesForReset(RuleSet ruleset,
                            EvolutionState state,
                            int thread)
        Returns a stochastic value picked to specify the number of rules to generate when calling reset() on this kind of Rule. The default version picks from the min/max or distribution, but you can override this to do whatever kind of thing you like here.

      • toString

        public java.lang.String toString()
        Converting the rule to a string ( the name )
        Overrides:
        toString in class java.lang.Object

      • constraintsFor

        public static RuleSetConstraints constraintsFor(java.lang.String constraintsName,
                                                EvolutionState state)
        You must guarantee that after calling constraintsFor(...) one or several times, you call state.output.exitIfErrors() once.

      • setup

        public void setup(EvolutionState state,
                  Parameter base)
        Description copied from interface: Setup
        Sets up the object by reading it from the parameters stored in state, built off of the parameter base base. If an ancestor implements this method, be sure to call super.setup(state,base); before you do anything else.
        Specified by:
        setup in interface Setup
Skip navigation links

DMelt 3.0 © DataMelt by jWork.ORG

You see the box below because you did not login.